From jaksonaquino at yahoo.com.br Sat Feb 5 23:26:48 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sat, 5 Feb 2005 19:26:48 -0300 (ART) Subject: es_ES.po Message-ID: <20050205222649.7147.qmail@web51905.mail.yahoo.com> Hello all! I added the following translations to es_ES.po. but they should be checked by someone who speaks Spanish and knows statistcs, that is, probably Carleo himself (if he still is subscribing this mail list). msgid "G (check value for Chi^2-Test) = %f\n" msgstr "G (valor de prueba para contraste ji^2) = %f\n" msgid "Hypothesis H0: Correlation according to probit-model exists\n" msgstr "Hip?tesis H0: Existe correlaci?n seg?n el modelo probit\n" msgid "set title 'STATIST: Cummulative distribtion'\n" msgstr "set title 'STATIST: Distribuci?n cumulativa'" msgid "set title 'STATIST: Polynomial Regression of Order %i'\n" msgstr "set title 'STATIST: Regression Polinomial de Orden %i'\n" msgid " Compiled with DEBUG.\n" msgstr "Compilado con DEBUG.\n" msgid " predicted" msgstr " predicto" Is the following translation correct, or is it necessary to translate the word "rank"? msgid "Matrix of SPEARMAN'S Rank-Correlation-Coefficients\n" msgstr "Matriz de coeficientes de correlaci?n de SPEARMANN\n" Best regards, Jakson _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From jaksonaquino at yahoo.com.br Sun Feb 6 22:45:04 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sun, 6 Feb 2005 18:45:04 -0300 (ART) Subject: lang_defs.h Message-ID: <20050206214504.96312.qmail@web51901.mail.yahoo.com> Hello All, On May 21, 2001, Andreas Beyer wrote: > I got a message from someone who would like o write a Spanish > translation of StatistX. I don't think it is wise to expand the > translation-via-macros method to an infinite number of languages. I agree with Andreas Beyer. It seems that lang_defs.h was created as a way of internationalizing statist. However, it's not practical to have messages written directly in the source code in more than two languages. Thus, I think that we don't need the file lang_defs.h anymore. I'm submitting a patch where, with some deletions and with no addition, the program is compiling as before, but without using the file lang_defs.h. To be sure that no bug was introduced, I succesfully recompiled and ran statist 5 times: Linux: English, German, and gettext. DOS : English, and German. DLL are files used by dos and windows, right? While reading info pages for gettext, I found the following passage which implies that gettext can be used under DOS: "To convert a PO file to a `.dll' file, the `msgfmt' program can be used with the option `--csharp'. The result will be a `.dll' file containing a subclass of `GettextResourceSet', which itself is a subclass of `ResourceSet'. To convert a `.dll' file containing a `GettextResourceSet' subclass back to a PO file, the `msgunfmt' program can be used with the option `--csharp'." I had made an attempting of creating a .dll file, but I got the following error: [jakson at localhost po]$ msgfmt --csharp -l pt_BR -d /temp --verbose pt_BR.po msgfmt: C# compiler not found, try installing pnet msgfmt: compilation of C# class failed, please try --verbose 353 translated messages, 18 fuzzy translations. [jakson at localhost po]$ Is anyone interested in internationalizing Statist for DOS? I will not become responsible for such a task, because it's not easy to me, and because I'm already satisfied with Linux. Best regards, Jakson __________________________________________________ Converse com seus amigos em tempo real com o Yahoo! Messenger http://br.download.yahoo.com/messenger/ -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: statist_lang_defs_patch Url: http://www.intevation.de/pipermail/statist-list/attachments/20050206/451ba3e3/statist_lang_defs_patch.txt From jaksonaquino at yahoo.com.br Sun Feb 6 22:46:14 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sun, 6 Feb 2005 18:46:14 -0300 (ART) Subject: missing values Message-ID: <20050206214614.10998.qmail@web51906.mail.yahoo.com> Hello All! I didn't look at the source code carefuly, but from the "behavior" of Statist it seems to me that we have a problem. Suppose that a research team applied a questionnaire to 600 people, asking them 50 questions. Some question wouldn't apply to every one, and many people wouldn't answer many questions. Now, I have a database with 50 columns and 600 rows, and many missing values. Every time I want to do some correlation between variables I will have to do: 1) Open with Statist a version of the database with no 'M' inside. The missing values would be indicated by some valid value. 2) Save the columns that I want to correlate in a ASCII file. 3) Quit Statist 4) Use an external program to recode the new and smaller database to turn into 'M' all values that, in fact, are missing. 5) Load the new small recoded database with the option -delrow and, then, run the analysis. While still becoming familiar with the database, I should run many analyzes, mainly regression analyzes or multiple linear correlations, with different combinations of variables. It's easy to imagine how boring this work would be. Statist always deletes missing values while reading a database. The option -delrow mean only that the entire row, instead of just one data point, must be deleted. That's the wrong approach because the rows can be broken without the option -delrow, and valid values can be deleted with the option -delrow. Data in the same row are supposed to be a single case, and they can't be split in different rows. Thus, rows with missing values should not be deleted. They should be ignored during calculations, when necessary, not while loading a data file. And the problem can be more than a mere inconvenient; it can be a bug if we don't follow the five steps described above, and, unluckily, have one of the two circumstances described below: 1) A data base is opened without the option -delrow, the columns used in a calculation have missing values, and, for coincidence, they have the same number of missing values, but not in the same rows. The bug happens because Statist only tests whether columns have the same number of data points before proceeding. In this case, the correlation for these columns would proceed and Statist would report some wrong result. 2) The option -delrow is used, but the analysis don't uses all columns, and the columns don't have their missing values at exactly the same rows, consequently some rows that should be used in the analysis were deleted while opening the data base. Again, Statist would proceed with the analysis and report some result. We need some way of keeping Statist aware of missing values, instead of simply deleting them and testing whether the columns have the same number of data points. Perhaps we can define the biggest possible double as missing value! [I don't know what the biggest double value is.] That is, whenever Statist finds 'M' in the database it puts 9.99999e+99 in its memory, and whenever it finds 9.99999e+99 in its memory it assumes that it is a missing value. While writing columns as ASCII files, missing would come back to the usual 'M'. And, just for precaution, if it finds 9.99999e+99 while reading an ASCII file, it will reports a error. Of course this will slow down the computations, since an if(xx[i] == 9.99999e+99) will be necessary when reading each single data point from memory. And, the worst, it would be necessary to fix all functions! Any better solution? Anyway, we need to write a documentation in English and include this information in the documentation. Am I wrong, and there is no problem with the way Statist deal with missing values? Best regards, Jakson __________________________________________________ Converse com seus amigos em tempo real com o Yahoo! Messenger http://br.download.yahoo.com/messenger/ From jaksonaquino at yahoo.com.br Sat Feb 5 23:25:10 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sat, 5 Feb 2005 19:25:10 -0300 (ART) Subject: gettext Message-ID: <20050205222510.12991.qmail@web51902.mail.yahoo.com> Hello All! I'm a social scientist and I plan to teach quantitative methods of research when I finish my doctorate. Of course I will need a software. I surveyed the Internet looking for a program that was easy to use, without any necessity of remembering commands, since we, social scientists, in general, only use statistics for some days in an entire year. We easily forget commands. It also would be very nice a program running in Portuguese (since few Brazilian undergraduate students can speak English), and that was open source. Statist seems to be the best option, and I would be glad if I could contribute to its development. As my first contribution, I wrote some translations. I have just subscribed the mail list, but I read some messages posted in the last years. In one of them (Fri, Jun 11, 2004), Bernhard Reiter wrote: > statist in CVS is prepared for gettext support. ... > However I did not create the necessary infrastructures > for the message files yet, thus there is no release. Since the po directory is not in the cvs yet, I created one with Portuguese, Spanish and German translations. The German translations were extracted from the source code with the gettext option -k_:1,2. The same option was used to extract the translations made by Carleo some time ago and that are in the cvs (or in the mail archive, I don't remember). I know that there is an Italian translation. How can I get it? Spanish and Portuguese are very similar, and thanks to Carleo's Spanish translations, I could translate many entries in my pt_BR.po that were difficult to me. However, the two languages don't share all words, and I still need that some Brazilian statistician reviews my pt_BR.po. I believe that in two weeks I'll be able to get my file revised. The translations are in IS0-8859-1, and the words are with their accents. For people with difficult to display accentuated words in their computers, we can easily create an option like "make noaccents" to automatically create new .po files with all accents removed. The opposite process, putting the accents back in the words, would be a very hard manual work. Could someone put the accents in the de_DE.po file? The strings were without accents in the source code. The po directory has a Makefile. Typing "make po" extracts the strings from the source files and updates (msgmerge) the three .po files (pt_BR, es_ES, and de_DE). Typing "make mo" compiles the .po files and creates the .mo files, and typing "make install" copies the .mo files to /usr/share/locale/xx_XX/LC_MESSAGES/. Typing "make" with no options prints a short help text. On Feb 20, 2002, Bernhard Reiter wrote: > probably also would be appropriate. You should get the credit of course > and people otherwise might contact me and assume that I speak Spanish. :) Thus, I put Carleo's full name and email in the es_ES.po file. However, the de_DE.po file is with "no" translator, because I don't know who will be responsible for it. I was getting "?" instead of accentuated letters when statist was running in Portuguese. I just replaced "LC_MESSAGES" with "LC_ALL" and now the accents are correctly displayed. Even r^2 is being nicely printed as r? (r + 0xb2) and mu is now ? (0xb5). While translating the messages, I found some possible problems in the English original ones. I, then, made small changes in the source code, and I'm submitting these changes in the statist_gettext_patch: menue.c: 82: "Tests" must be translated in languages other than German. Thus, I replaced out_d( " 3 = Tests\n") with out_d( _(" 3 = Tests\n", " 3 = Tests\n") 269: I replaced "Main Menu", with "Main menu" (because "Main menu" was used many times in other files, and "Main Menu" only once). 623: The string in this line was in German only, but I found the translation in another place and copied it: "Bitte Anzahl der zu korrelierenden Spalten eingeben,\n" "Erste angewaehlte Spalte wird als y-Wert genommen: " "Please enter number of columns to be correlated\n" "(first selected column will be taken as y-value): " Note that the original English message was a little confusing and that I changed it, adopting the same solution that Carleo's applied to his Spanish translation. I put the verb in the future because after this message, we have to enter the number of variables and not the column's label. This change was applied in three places in menu.c. However, I did nothing with the German text. 276: I replaced " $ = Chi^2-fourfould-test\n" with " 4 = Chi^2-fourfould-test\n" procs.c 1139: "since frequences <= 5 ocure!" => "since there are frequencies <= 5!" data.c: 204: I replaced "column" with "line". (the %i refers to the line in the file, not to the column in the database) I had also used Aspell to find misspelled words in statist.pot and I corrected them in the source code: caclulated ==> calculated literatur ==> literature fourfould => fourfold Modul => Module numnber => number differen => different enties => entries Discret => Discrete normaly => normally MISCELANEOUS => MISCELLANEOUS miscelaneous => miscellaneous indecated => indicated meaningfull => meaningful Cummulative => Cumulative distribtion => distribution Probitanalyse => Probit analysis Probitanalysis => Probit analysis Lillefors => Lilliefors Lillifors => Lilliefors Liliefors => Lilliefors SPEARMANN => SPEARMAN quantile => quartile get_quantile => get_quartile Note: The last one (get_quantile) is the name of a function, not a message. If the changes are applied, the .po files will need update, since they are still reflecting the old sources. I already made most of the necessary update: since there is no real change, I mainly deleted the commentary #, fuzzy. However, there are two fuzzy translations in both es_ES.po and de_DE.po. I'm sending three files as attachment: po.tar.gz po_after_changes.tar.gz statist_gettext_patch The po.tar.gz, has the po directory before I changed some messages in the sources. The po_after_changes.tar.gz has the po directory after I made the changes. The statist_gettext_patch is the patch to all changes I made. I created it with diff -Naur. This is the first time I use diff. Please, tell me if I have to change something to make the use of my patches easier. Best regards, Jakson http://www.geocities.com/jakson-aquino/ P.S.: The rule "memory_handling.o:" was missing in the Makefile. I added the rule, although Statist was compiling without it. _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis -------------- next part -------------- A non-text attachment was scrubbed... Name: po.tar.gz Type: application/x-gzip-compressed Size: 46513 bytes Desc: po.tar.gz Url : http://www.intevation.de/pipermail/statist-list/attachments/20050205/b693aeb9/po.tar.gz -------------- next part -------------- A non-text attachment was scrubbed... Name: po_after_changes.tar.gz Type: application/x-gzip-compressed Size: 46527 bytes Desc: po_after_changes.tar.gz Url : http://www.intevation.de/pipermail/statist-list/attachments/20050205/b693aeb9/po_after_changes.tar.gz -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: statist_gettext_patch Url: http://www.intevation.de/pipermail/statist-list/attachments/20050205/b693aeb9/statist_gettext_patch.txt From bernhard at intevation.de Mon Feb 7 10:52:44 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 7 Feb 2005 10:52:44 +0100 Subject: lang_defs.h In-Reply-To: <20050206214504.96312.qmail@web51901.mail.yahoo.com> References: <20050206214504.96312.qmail@web51901.mail.yahoo.com> Message-ID: <20050207095244.GE30059@intevation.de> Hi Jakson, On Sun, Feb 06, 2005 at 06:45:04PM -0300, Jakson Aquino wrote: > On May 21, 2001, Andreas Beyer wrote: > > translation of StatistX. I don't think it is wise to expand the > > translation-via-macros method to an infinite number of languages. > > I agree with Andreas Beyer. You probably also saw my answer in 2001: http://intevation.de/pipermail/statist-list/2001/000009.html This is why we improved the gettext mechanisms since then. > It seems that lang_defs.h > was created as a way of internationalizing statist. > However, it's not practical to have messages written > directly in the source code in more than two > languages. Thus, I think that we don't need the file > lang_defs.h anymore. I'm submitting a patch where, > with some deletions and with no addition, the program > is compiling as before, but without using the file > lang_defs.h. The idea behind this was, to keep statist as simple and small as possible. Now that we have two languages in the source code, I think we should just keep them there and use gettext for the additional ones. The ideal solution would be that someone makes a special version of gettext so that we can extract the German translation directly from the source. ;) I'll look at the patch, it is probably fine to just have two possibilities in the source code and simplify it. > DLL are files used by dos and windows, right? AFAIK yes. > While > reading info pages for gettext, I found the following > passage which implies that gettext can be used under > DOS: > > "To convert a PO file to a `.dll' file, the > `msgfmt' program can be used with the option > `--csharp'. The result will be a `.dll' > file containing a subclass of > `GettextResourceSet', which itself is a > subclass of `ResourceSet'. To convert a > `.dll' file containing a `GettextResourceSet' > subclass back to a PO file, the `msgunfmt' > program can be used with the option `--csharp'." > > I had made an attempting of creating a .dll file, but > I got the following error: > > [jakson at localhost po]$ msgfmt --csharp -l pt_BR -d > /temp --verbose pt_BR.po > msgfmt: C# compiler not found, try installing pnet > msgfmt: compilation of C# class failed, please try > --verbose > 353 translated messages, 18 fuzzy translations. > [jakson at localhost po]$ > > Is anyone interested in internationalizing Statist for > DOS? I will not become responsible for such a task, > because it's not easy to me, and because I'm already > satisfied with Linux. Let us wait until somebody jumps up for DOS, if possible I do not want to break the possibilities that statist has under DOS, but I have not used them myself since a very long while. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050207/2d623cda/attachment.bin From bernhard at intevation.de Mon Feb 7 11:01:07 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 7 Feb 2005 11:01:07 +0100 Subject: Italian translation (Re: gettext) In-Reply-To: <20050205222510.12991.qmail@web51902.mail.yahoo.com> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> Message-ID: <20050207100107.GG30059@intevation.de> On Sat, Feb 05, 2005 at 07:25:10PM -0300, Jakson Aquino wrote: > I knowthat there is an Italian translation. > How can I get it? italian_translation.patch Italien language support by Nicola Intini Note: add -p5 as option to patch when you are in the statist-1.0.0 dir http://www.usf.uni-osnabrueck.de/~breiter/tools/statist/v1.0.0/contrib/ Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050207/d8e55e5c/attachment.bin From bernhard at intevation.de Mon Feb 7 11:08:34 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 7 Feb 2005 11:08:34 +0100 Subject: missing values In-Reply-To: <20050206214614.10998.qmail@web51906.mail.yahoo.com> References: <20050206214614.10998.qmail@web51906.mail.yahoo.com> Message-ID: <20050207100834.GH30059@intevation.de> Hi Jakson, On Sun, Feb 06, 2005 at 06:46:14PM -0300, Jakson Aquino wrote: > I didn't look at the source code carefuly, but from > the "behavior" of Statist it seems to me that we have > a problem. I agree that the behaviour you describe below is problematic and we need to test it more thoroughly and find a solution. > Suppose that a research team applied a questionnaire > to 600 people, asking them 50 questions. Some question > wouldn't apply to every one, and many people wouldn't > answer many questions. Now, I have a database with 50 > columns and 600 rows, and many missing values. Every > time I want to do some correlation between variables I > will have to do: > > 1) Open with Statist a version of the database with no > 'M' inside. The missing values would be indicated by > some valid value. > > 2) Save the columns that I want to correlate in a > ASCII file. > > 3) Quit Statist > > 4) Use an external program to recode the new and > smaller database to turn into 'M' all values that, in > fact, are missing. > > 5) Load the new small recoded database with the option > -delrow and, then, run the analysis. My idea was that recoding the long file with 'M' on the missing values for the full database should be a one step operation and then statist can do all the various analysis on it. If this is not the case, this is indeed bad. Could you produce a few example files so that the error can be made obvious. We can use that as test cases when we fix the bug. > While still becoming familiar with the database, I > should run many analyzes, mainly regression analyzes > or multiple linear correlations, with different > combinations of variables. It's easy to imagine how > boring this work would be. > We need some way of keeping Statist aware of missing > values, instead of simply deleting them and testing > whether the columns have the same number of data > points. Yes. This could be done before the calculations. > Perhaps we can define the biggest possible > double as missing value! [I don't know what the > biggest double value is.] There will be a standard define for this. > That is, whenever Statist > finds 'M' in the database it puts 9.99999e+99 in its > memory, and whenever it finds 9.99999e+99 in its > memory it assumes that it is a missing value. While > writing columns as ASCII files, missing would come > back to the usual 'M'. And, just for precaution, if it > finds 9.99999e+99 while reading an ASCII file, it will > reports a error. Of course this will slow down the > computations, since an if(xx[i] == 9.99999e+99) will > be necessary when reading each single data point from > memory. And, the worst, it would be necessary to fix > all functions! Any better solution? Putting a special value in is not the niced solution, however I would need to look at the code to think about others. > Anyway, we need to write a documentation in English > and include this information in the documentation. An English documentation is really needed, that is true. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050207/f7882f3b/attachment.bin From bernhard at intevation.de Mon Feb 7 11:13:09 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 7 Feb 2005 11:13:09 +0100 Subject: gettext In-Reply-To: <20050205222510.12991.qmail@web51902.mail.yahoo.com> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> Message-ID: <20050207101309.GI30059@intevation.de> On Sat, Feb 05, 2005 at 07:25:10PM -0300, Jakson Aquino wrote: > However, the de_DE.po file is with "no" > translator, because I don't know who will be > responsible for it. You can put in the list or me as interims translator, because of course I do speak German. ;) The German text actually is the original one done by Dirk Melcher and Andreas Beyer made the English translation later. > The statist_gettext_patch is the patch to all changes > I made. I created it with diff -Naur. This is the > first time I use diff. Please, tell me if I have to > change something to make the use of my patches easier. Diff usage looks fine to me. It sometimes takes a while until I have time to visit changes. At Intevation we want to set up a new platform for projects like this based on Gforge, then contributions will even be easier. > P.S.: The rule "memory_handling.o:" was missing in the > Makefile. I added the rule, although Statist was > compiling without it. Makefiles have build in rules for c files, but this denies the dependencies, so it better shall be in. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050207/5ce6a4ab/attachment.bin From jaksonaquino at yahoo.com.br Tue Feb 8 00:29:19 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 7 Feb 2005 20:29:19 -0300 (ART) Subject: missing values In-Reply-To: <20050207100834.GH30059@intevation.de> Message-ID: <20050207232919.2786.qmail@web51903.mail.yahoo.com> >The ideal solution would be that someone makes a special version >of gettext so that we can extract the German translation directly >from the source. ;) Gettext has an option to extract messages from ( , ). I used it to build the de_DE.po and es_ES.po. And I learned it from the statist Makefile, where we have: xgettext --keyword="_:1" I guessed that -k_:1,2 would extract both English and German, and it did work! German became msgid_plural. I had just to replace msgid_plural by msgstr. >You can put in the list or me as interims translator, >because of course I do speak German. ;) I'll have to update the potfiles because I found about 10 messages that were only in German (I used something like grep out_r\(\" to find them). They don't need translation from German to English, but need in other languages. I also made a rough translation of a few commentaries that still were in German in the sources. Perhaps we'll have more people reading and commenting the source code if it is entirely in English. I don't know Statist code very well yet, but it seemed to me that it would be no problem if Statist saved the columns in ASCII files with their names. So, a third thing that I did was to add 4 lines to menue.c. Now, when I open a file containing columns saved by my version of Statist the columns are already with their names. I'll submit a new patch with these changes together with the revised .po files, including (I hope) the Italian one. >It sometimes takes a while until I have time to visit changes. >At Intevation we want to set up a new platform for projects like >this based on Gforge, then contributions will even be easier. Sometimes I also take a few days to answer emails because I'm without Internet in my house. I'm using either cybercafes or the computers of my uni. >Makefiles have build in rules for c files, >but this denies the dependencies, so it better shall be in. Statist only compiled under DOS after I added the rule to memory_handling.o. My idea was to show Statist to some friends under DOS and tell them that they could see the program running in Portuguese under Linux. None of my friends has ever used Linux. I'll prepare a small data base to make it easy to see the problem with missing values. The first thing that I did after compiling Statist for the first time was to run a multiple linear correlation with 4 variables. I used a database that comes with SPSS, and run a linear regression in SPSS with the same 4 variables. Each column had originally 1500 rows, but they dropped to about 925 after -delrow did its work. In SPSS, the "N" was also 925. The results (coefficients, etc) were exactly the same, but I had to follow the steps described in my previous email to run the analysis in Statist. I also opened, but did no test, with a database that has 17 columns and 350000 rows. The for variables were extracted from PNAD (Brazilian National Household Survey). Best regards, Jakson _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From jaksonaquino at yahoo.com.br Tue Feb 8 00:30:48 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 7 Feb 2005 20:30:48 -0300 (ART) Subject: es_ES.po In-Reply-To: Message-ID: <20050207233048.90442.qmail@web51907.mail.yahoo.com> Muchas gracias for revising the translations. I think that soon it will be in a po directory in Statist cvs. Saludos cordiales, Jakson --- Carleos Artime escreveu: > Jakson Aquino writes: > > > Hello all! > > Hello! > > > I added the following translations to es_ES.po. > but > > they should be checked by someone who speaks > Spanish > > and knows statistcs, that is, probably Carleo > himself > > (if he still is subscribing this mail list). > > Yes, I'm still here, but haven't been using statist > for a long time. > > > msgid "set title 'STATIST: Cummulative > distribtion'\n" > > > > msgstr "set title 'STATIST: Distribuci?n > cumulativa'" > > In Spanish I have seen "Distribucion acumulada", and > once > "Distribucion acumulativa" (the most usual here is > "Funcion de > distribucion", but I am not completely sure that we > are speaking > of the same thing). > > [Please pardon that I do no use accented letters; > these days I hesitate > between ISO-8859-15 and UTF-8, so most times I do > not write accents.] > > > msgid "set title 'STATIST: Polynomial Regression > of > > Order %i'\n" > > msgstr "set title 'STATIST: Regression Polinomial > de > > Orden %i'\n" > > `Regression' should be with one `s' (regresion) and > with accented `o' > (regresi?n). `Polinomial' with lower case initial. > > > msgid " predicted" > > msgstr " predicto" > > " predicho". > > > Is the following translation correct, or is it > > necessary to translate the word "rank"? > > > > msgid "Matrix of SPEARMAN'S > > Rank-Correlation-Coefficients\n" > > msgstr "Matriz de coeficientes de correlaci?n de > > SPEARMANN\n" > > I have seen "coeficiente de correlacion de rangos", > but your version is > right (well, English and Spanish "Spearmen" with > same number of final `n'). > > Muito obrigado :-) > > > > -- > > _________________________________________________________________________ > > Carlos Enrique Carleos Artime > , GNU.org , > Dep-to de Statistiko kaj Plejbonigiko, > / \ > kaj Matematika Didaktiko > ((__-^^-,-^^-__)) > Universitato de Oviedo > `-_---' `---_-' > > `--|o` 'o|--' > Retposhto: carleos at uniovi.es > \ ` / > Telefono: +34 985 181 904 > ): :( > Fakso: +34 985 181 902 > :o_o: > Adreso: Escuela de Marina Civil, 33203 Hispanio > "-" > > __________________________________________________ Converse com seus amigos em tempo real com o Yahoo! Messenger http://br.download.yahoo.com/messenger/ From jaksonaquino at yahoo.com.br Tue Feb 8 20:46:58 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 8 Feb 2005 16:46:58 -0300 (ART) Subject: missing values In-Reply-To: <20050207100834.GH30059@intevation.de> Message-ID: <20050208194659.9963.qmail@web51903.mail.yahoo.com> Hi Bernhard! >Could you produce a few example files so that >the error can be made obvious. We can use that as test cases >when we fix the bug. I'm sending the example files as attachment. The file results_gss explains what was done and presents the results. The file gss20x6_file_info has information about the very small database used to make the test. >Putting a special value in is not the niced solution, >however I would need to look at the code to think about >others. Yes, if the special value was present in the database as a valid value this would be a problem. Perhaps a option like --diffsysmis=number (define different system missing value, other than 9.99999e+99) could solve the problem. PSPP, which is cloning spss, seems to use a list of user defined missing values to each colum: [jakson at localhost src]$ grep MISSING *.h -n exprP.h:187: OP_STR_MIS, /* MISSING(strvar). */ var.h:282: MISSING_NONE, /* No user-missing values. */ var.h:283: MISSING_1, /* One user-missing value. */ var.h:284: MISSING_2, /* Two user-missing values. */ var.h:285: MISSING_3, /* Three user-missing values. */ var.h:286: MISSING_RANGE, /* [a,b]. */ var.h:287: MISSING_LOW, /* (-inf,a]. */ var.h:288: MISSING_HIGH, /* (a,+inf]. */ var.h:289: MISSING_RANGE_1, /* [a,b], c. */ var.h:290: MISSING_LOW_1, /* (-inf,a], b. */ var.h:291: MISSING_HIGH_1, /* (a,+inf), b. */ var.h:292: MISSING_COUNT PSPP also seems to have a special value reserved to sysmis. Look at the following lines: var.h, line 40 and ss.: /* Special values. */ #define SYSMIS (-DBL_MAX) #define LOWEST second_lowest_value #define HIGHEST DBL_MAX magic.h, line 29 and ss.: fndef SECOND_LOWEST_VALUE /* "Second lowest" value for a flt64; that is, (-FLT64_MAX) + epsilon. */ double second_lowest_value; #endif PSPP code is too big and complex to me, and I just used grep to find the word "missing". But it seems that they are using the biggest possible double negative number as sysmis, and a lot of other user defined missing values. Perhaps using a very big (and negative) number was not so dangerous. Anyway, I don't know if this is a good solution. Anyway, I don't think it is necessary to complicate the code creating a complex list of user defined missing values. I prefer the simpler approach of Statist of simply putting a 'M' in the database. In social research we have three main sources of abudant missing values: (1) people don't know how to answer a question; (2) they do know how to, but are unwilling to answer, and (3) the question don't apply. In some analysis, all three cases are better considered missing values, but in others, often using the same data base, "I don't know" and "I prefer don't answer this question" must be distinctly counted and analysed. But this would happens only rarely, and it would not be that difficult to write a program to automatically recode an original database and create a new one with the 'M's correctly replacing values that have to be recoded as missing. Best, Jakson _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis -------------- next part -------------- A non-text attachment was scrubbed... Name: missing_problem.tar.gz Type: application/x-gzip-compressed Size: 3484 bytes Desc: missing_problem.tar.gz Url : http://www.intevation.de/pipermail/statist-list/attachments/20050208/55b59284/missing_problem.tar.gz From jaksonaquino at yahoo.com.br Tue Feb 8 20:49:30 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 8 Feb 2005 16:49:30 -0300 (ART) Subject: gettext In-Reply-To: <20050207101309.GI30059@intevation.de> Message-ID: <20050208194930.25567.qmail@web51901.mail.yahoo.com> Hello All! I finished some tasks related with the full internationalization of Statist. I made the new patches translanting the comments that still were in German, and updated the new po directory, now including the Italian translation. To avoid sending emails with attachments biger than 40 kb, I put the new files on the web at the following address: htpp://www.geocities.com/jakson-aquino/statist.html To create and update the .po files I always followed some steps: 1) Used xgettext to extracted the messages and their translations from the source code. (he exception was the Portguese translation, made directly in the .po file). 2) Typed "make po" to update the strings. 3) Looked at each translation marked as "fuzzy" by msgmerge. When possible, the problem was corrected consulting the original patch (es and it) or source code (de). 4) Looked at each msgid that was not translated, and again looked the source to find the translation. 5) Typed "make mo" to see if there were syntax problems in the .po files. 6) Ran statist in different languages to have fun! I removed the comentary "fuzzy" when I could correct the translation. But, some few translation still are marked as fuzzy. Best, Jakson __________________________________________________ Converse com seus amigos em tempo real com o Yahoo! Messenger http://br.download.yahoo.com/messenger/ From beyer at imb-jena.de Wed Feb 9 09:39:31 2005 From: beyer at imb-jena.de (Andreas Beyer) Date: Wed, 09 Feb 2005 09:39:31 +0100 Subject: missing values In-Reply-To: <20050208194659.9963.qmail@web51903.mail.yahoo.com> References: <20050208194659.9963.qmail@web51903.mail.yahoo.com> Message-ID: <4209CC43.2000407@imb-jena.de> Hi, instead of using something like -99 as a representation for missing values we should use something like infinity. The GNU C-lib knows the constant: float NAN; which is "not-a-number". Unfortunately this constant is only defined on GNU systems. In addition, NAN is unequal to itself. Hence, the following is true: float x = NAN; assert(x != x); Instead we might use the maximum double, i.e. something like the following: #include float.h // DBL_MAX static double _no_value = DBL_MAX; /* Obtaining the internal representation for a missing value. */ double missing_value() { return _no_value; } /* Check if val is a missing value. */ int is_missing_value(double val) { return val == _no_value; } I prefer encapsulating the missing-value check in a function, because that leaves room for future extensions that might become more complex. Andreas Jakson Aquino schrieb: > Hi Bernhard! > > >>Could you produce a few example files so that >>the error can be made obvious. We can use that as > > test cases > >>when we fix the bug. > > > I'm sending the example files as attachment. The file > results_gss explains what was done and presents the > results. The file gss20x6_file_info has information > about the very small database used to make the test. > > > >>Putting a special value in is not the niced solution, >>however I would need to look at the code to think > > about > >>others. > > > Yes, if the special value was present in the database > as a valid value this would be a problem. Perhaps a > option like --diffsysmis=number (define different > system missing value, other than 9.99999e+99) could > solve the problem. PSPP, which is cloning spss, seems > to use a list of user defined missing values to each > colum: > > [jakson at localhost src]$ grep MISSING *.h -n > exprP.h:187: OP_STR_MIS, /* > MISSING(strvar). */ > var.h:282: MISSING_NONE, /* No > user-missing values. */ > var.h:283: MISSING_1, /* One > user-missing value. */ > var.h:284: MISSING_2, /* Two > user-missing values. */ > var.h:285: MISSING_3, /* > Three user-missing values. */ > var.h:286: MISSING_RANGE, /* [a,b]. */ > var.h:287: MISSING_LOW, /* (-inf,a]. > */ > var.h:288: MISSING_HIGH, /* (a,+inf]. > */ > var.h:289: MISSING_RANGE_1, /* [a,b], c. > */ > var.h:290: MISSING_LOW_1, /* (-inf,a], > b. */ > var.h:291: MISSING_HIGH_1, /* (a,+inf), > b. */ > var.h:292: MISSING_COUNT > > PSPP also seems to have a special value reserved to > sysmis. Look at the following lines: > > var.h, line 40 and ss.: > /* Special values. */ > #define SYSMIS (-DBL_MAX) > #define LOWEST second_lowest_value > #define HIGHEST DBL_MAX > > magic.h, line 29 and ss.: > fndef SECOND_LOWEST_VALUE > /* "Second lowest" value for a flt64; that is, > (-FLT64_MAX) + epsilon. */ > double second_lowest_value; > #endif > > PSPP code is too big and complex to me, and I just > used grep to find the word "missing". But it seems > that they are using the biggest possible double > negative number as sysmis, and a lot of other user > defined missing values. Perhaps using a very big (and > negative) number was not so dangerous. Anyway, I don't > know if this is a good solution. > > Anyway, I don't think it is necessary to complicate > the code creating a complex list of user defined > missing values. I prefer the simpler approach of > Statist of simply putting a 'M' in the database. In > social research we have three main sources of abudant > missing values: (1) people don't know how to answer a > question; (2) they do know how to, but are unwilling > to answer, and (3) the question don't apply. In some > analysis, all three cases are better considered > missing values, but in others, often using the same > data base, "I don't know" and "I prefer don't answer > this question" must be distinctly counted and > analysed. But this would happens only rarely, and it > would not be that difficult to write a program to > automatically recode an original database and create a > new one with the 'M's correctly replacing values that > have to be recoded as missing. > > Best, > > Jakson > > > > > > > _______________________________________________________ > Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis > > > ------------------------------------------------------------------------ > > _______________________________________________ > Statist-list mailing list > Statist-list at intevation.de > https://intevation.de/mailman/listinfo/statist-list From jaksonaquino at yahoo.com.br Wed Feb 9 13:29:21 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Wed, 9 Feb 2005 09:29:21 -0300 (ART) Subject: missing values In-Reply-To: <4209CC43.2000407@imb-jena.de> Message-ID: <20050209122921.2141.qmail@web51904.mail.yahoo.com> Hi! Before reading the email written by Andreas Beyer, I had written something similar. I also wrote some suggestions of code, but I don't know Statist source code sufficiently well to have any certainty that what I wrote would work. If I'm not wrong, we can fix the problem of missing values writing few lines of code. I'm still thinking that we can define one specific double number as system missing value. I - DEFINING SYSMIS First we would have to give to Statist the ability to keep missing values. We could borrow one line from pspp and put in statist.h: #define SYSMIS (-DBL_MAX) I wrote the following program to see the exact value of -DBL_MAX: #include #include #define SYSMIS (-DBL_MAX) int main(){ printf("\n%e\n\n", SYSMIS); } The program was compiled wit g++, and the output was: -1.797693e+308 The complete number is: -17976931348623157081452742373170435679807056752584499 659891747680315726078002853876058955863276687817154045 895351438246423432132688946418276846754670353751698604 991057655128207624549009038932894407586850845513394230 458323690322294816580855933212334827479782620414472316 8738177180919299881250404026184124858368.000000 Certainly the probability of finding this specific number in a database is very small. We can tell in the documentation that if this number appears in the database it is the responsibility of the user to define a different SYSMIS with the option -sysmis=number, or something similar. II - WRITING SYSMIS IN TEMPFILES Now that we have the variable SYSMIS, we have to read columns correctly. Perhaps the following replacement will be enough: data.c, line 261 and ff: if ((token[0] == NODATA) && (strlen(token)==1)) { colread ++; } else if (sscanf(token, "%lf", &test)==1) { FWRITE(&test, sizeof(REAL), 1, ttempfile[actcol]); nn[actcol] ++; colread ++; } Should be changed to something like: if ((token[0] == NODATA) && (strlen(token)==1)) { FWRITE(&SYSMIS, sizeof(REAL), 1, ttempfile[actcol]); nn[actcol] ++; colread ++; } else if (sscanf(token, "%lf", &test)==1) { FWRITE(&test, sizeof(REAL), 1, ttempfile[actcol]); nn[actcol] ++; colread ++; } III - READING SYSMIS FROM TEMPFILES At first, I thought that it would be necessary to rewrite all functions because of the missing values problem. But now I have some hope that this will not be necessary. If I'm guessing correctly, a code similar to the following can solve the problem: After the following code (data.c : 482 and ff.): for (k=0; k escreveu: > Hi, > > instead of using something like -99 as a > representation for missing > values we should use something like infinity. The > GNU C-lib knows the > constant: > > float NAN; > > which is "not-a-number". Unfortunately this constant > is only defined on > GNU systems. In addition, NAN is unequal to itself. > Hence, the following > is true: > > float x = NAN; > assert(x != x); > > Instead we might use the maximum double, i.e. > something like the following: > > #include float.h // DBL_MAX > > static double _no_value = DBL_MAX; > > /* Obtaining the internal representation for a > missing value. */ > double missing_value() > { > return _no_value; > } > > /* Check if val is a missing value. */ > int is_missing_value(double val) > { > return val == _no_value; > } _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From jaksonaquino at yahoo.com.br Thu Feb 10 13:37:47 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Thu, 10 Feb 2005 09:37:47 -0300 (ART) Subject: locale.h and gnuplot Message-ID: <20050210123747.76023.qmail@web51901.mail.yahoo.com> Hello All! When trying to see a boxplot I got the following gnuplot error: Spalte fuer Variable 1: TVHOURS Erstelle gnuplot-Graphik ... 0 vermutliche Ausreisser gefunden --- Bitte mit ENTER fortfahren ... --- gnuplot> plot [0,8:5,2][9:13] '/home/jakson/stat_gpl.dat' with lines ^ line 0: ':' or keyword 'to' expected The problem is that, in many languages (among them: German, French, Spanish, Italian, and Portuguese), the decimal separator is not a "."; it's a ",". However, the culprit is . With locale.h, the printf function becomes very smart. It's good to see the results of calculations printed with the correct decimal separator, but, while writing a file to gnuplot, locale.h is a problem. Another --far less important-- side effect of locale.h is that if we save columns in an ascii file running Statist in a language other than English, it will not be possible to open the file running Statist in English. We would have, for example, 1,30000e+01 instead of 1.30000e+01. Since this side effect isn't properly a bug of Statist, I think that it will be enough to document it. And the solution is very simple: use some tool to replace all "," with ".". Is there any easy way of fixing the problem with gnuplot? Best, Jakson _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From bernhard at intevation.de Thu Feb 10 13:45:24 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Thu, 10 Feb 2005 13:45:24 +0100 Subject: locale.h and gnuplot In-Reply-To: <20050210123747.76023.qmail@web51901.mail.yahoo.com> References: <20050210123747.76023.qmail@web51901.mail.yahoo.com> Message-ID: <20050210124524.GA18784@intevation.de> On Thu, Feb 10, 2005 at 09:37:47AM -0300, Jakson Aquino wrote: > When trying to see a boxplot I got the following > gnuplot error: > > Spalte fuer Variable 1: TVHOURS > > Erstelle gnuplot-Graphik ... > > > 0 vermutliche Ausreisser gefunden > > --- Bitte mit ENTER fortfahren ... --- > gnuplot> plot [0,8:5,2][9:13] > '/home/jakson/stat_gpl.dat' with lines > ^ > line 0: ':' or keyword 'to' expected > > > The problem is that, in many languages (among them: > German, French, Spanish, Italian, and Portuguese), the > decimal separator is not a "."; it's a ",". However, > the culprit is . With locale.h, the printf > function becomes very smart. It's good to see the > results of calculations printed with the correct > decimal separator, but, while writing a file to > gnuplot, locale.h is a problem. > > Another --far less important-- side effect of locale.h > is that if we save columns in an ascii file running > Statist in a language other than English, it will not > be possible to open the file running Statist in > English. We would have, for example, 1,30000e+01 > instead of 1.30000e+01. Since this side effect isn't > properly a bug of Statist, I think that it will be > enough to document it. And the solution is very > simple: use some tool to replace all "," with ".". > > Is there any easy way of fixing the problem with > gnuplot? We probably need to ignore the locale settings for files and gnuplot. This is possible, I only have to look up how. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050210/099416eb/attachment.bin From beyer at imb-jena.de Thu Feb 10 15:55:53 2005 From: beyer at imb-jena.de (Andreas Beyer) Date: Thu, 10 Feb 2005 15:55:53 +0100 Subject: locale.h and gnuplot In-Reply-To: <20050210124524.GA18784@intevation.de> References: <20050210123747.76023.qmail@web51901.mail.yahoo.com> <20050210124524.GA18784@intevation.de> Message-ID: <420B75F9.5040700@imb-jena.de> I think I had the same problem with StatistX. If I remember correctly, my solution was to set the locale to the default C-locale before passing anything to gnuplot. Afterwards I set it back to the previous setting. Andreas Bernhard Reiter schrieb: > On Thu, Feb 10, 2005 at 09:37:47AM -0300, Jakson Aquino wrote: > >>When trying to see a boxplot I got the following >>gnuplot error: >> >>Spalte fuer Variable 1: TVHOURS >> >>Erstelle gnuplot-Graphik ... >> >> >>0 vermutliche Ausreisser gefunden >> >> --- Bitte mit ENTER fortfahren ... --- >>gnuplot> plot [0,8:5,2][9:13] >>'/home/jakson/stat_gpl.dat' with lines >> ^ >> line 0: ':' or keyword 'to' expected >> >> >>The problem is that, in many languages (among them: >>German, French, Spanish, Italian, and Portuguese), the >>decimal separator is not a "."; it's a ",". However, >>the culprit is . With locale.h, the printf >>function becomes very smart. It's good to see the >>results of calculations printed with the correct >>decimal separator, but, while writing a file to >>gnuplot, locale.h is a problem. >> >>Another --far less important-- side effect of locale.h >>is that if we save columns in an ascii file running >>Statist in a language other than English, it will not >>be possible to open the file running Statist in >>English. We would have, for example, 1,30000e+01 >>instead of 1.30000e+01. Since this side effect isn't >>properly a bug of Statist, I think that it will be >>enough to document it. And the solution is very >>simple: use some tool to replace all "," with ".". >> >>Is there any easy way of fixing the problem with >>gnuplot? > > > We probably need to ignore > the locale settings for files and gnuplot. > This is possible, I only have to look up how. > > > > ------------------------------------------------------------------------ > > _______________________________________________ > Statist-list mailing list > Statist-list at intevation.de > https://intevation.de/mailman/listinfo/statist-list -- -------------------------------------- Dr. Andreas Beyer - beyer at imb-jena.de http://www.imb-jena.de/~beyer -------------------------------------- From jaksonaquino at yahoo.com.br Thu Feb 10 22:13:35 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Thu, 10 Feb 2005 18:13:35 -0300 (ART) Subject: missing values In-Reply-To: <20050209122921.2141.qmail@web51904.mail.yahoo.com> Message-ID: <20050210211335.22289.qmail@web51906.mail.yahoo.com> Hello All! Did anyone worked on the problem of missing values? I was curious to know whether the code that I wrote in the other email would work, and, then, I tested it. With some adaptations, it seems that it's working. However, I tested it only with and . I opened the file that I sent as example (gss20x6_M.dat), and Statist correctly deleted rows with missing values during many calculations. Latter, I also tested new code with the database that have 17 columns and 350000 rows. It worked well also. I declared REAL SYSMIS as (-DBL_MAX + 1) because Statist is already using the values (-DBL_MAX) and (DBL_MAX). The changes that I made are below. I can send them as a patch against version 1.0.1, but perhaps it's better to wait for the solution of the problem with gnuplot. ================================================= statist.h ------------------------------------------------- extern REAL SYSMIS; /* 'M's are put in tempfiles with this value */ extern int tn; /* total number of rows in data base */ ================================================= ================================================= statist.c ------------------------------------------------- REAL SYSMIS = (-DBL_MAX) + 1; int tn; ================================================= ================================================= data.c ------------------------------------------------- in function readsourcefile(), lines 262 and ff. if ((token[0] == NODATA) && (strlen(token)==1)) { FWRITE(&SYSMIS, sizeof(REAL), 1, ttempfile[actcol]); nn[actcol] ++; colread ++; } else if (sscanf(token, "%lf", &test)==1) { FWRITE(&test, sizeof(REAL), 1, ttempfile[actcol]); nn[actcol] ++; colread ++; } . . . tn = lread - first_line; ------------------------------------------------- void alloc_cols(int n_alloc) { int k; for (k=0; k Message-ID: <20050210211525.22890.qmail@web51906.mail.yahoo.com> Hi, I'll look at StatistX source code. Best, Jakson --- Andreas Beyer escreveu: > > I think I had the same problem with StatistX. If I > remember correctly, > my solution was to set the locale to the default > C-locale before passing > anything to gnuplot. Afterwards I set it back to the > previous setting. > > Andreas > > Bernhard Reiter schrieb: > > On Thu, Feb 10, 2005 at 09:37:47AM -0300, Jakson > Aquino wrote: > > > >>When trying to see a boxplot I got the following > >>gnuplot error: > >> > >>Spalte fuer Variable 1: TVHOURS > >> > >>Erstelle gnuplot-Graphik ... > >> > >> > >>0 vermutliche Ausreisser gefunden > >> > >> --- Bitte mit ENTER fortfahren ... --- > >>gnuplot> plot [0,8:5,2][9:13] > >>'/home/jakson/stat_gpl.dat' with lines > >> ^ > >> line 0: ':' or keyword 'to' expected > >> > >> > >>The problem is that, in many languages (among > them: > >>German, French, Spanish, Italian, and Portuguese), > the > >>decimal separator is not a "."; it's a ",". > However, > >>the culprit is . With locale.h, the > printf > >>function becomes very smart. It's good to see the > >>results of calculations printed with the correct > >>decimal separator, but, while writing a file to > >>gnuplot, locale.h is a problem. > >> > >>Another --far less important-- side effect of > locale.h > >>is that if we save columns in an ascii file > running > >>Statist in a language other than English, it will > not > >>be possible to open the file running Statist in > >>English. We would have, for example, 1,30000e+01 > >>instead of 1.30000e+01. Since this side effect > isn't > >>properly a bug of Statist, I think that it will be > >>enough to document it. And the solution is very > >>simple: use some tool to replace all "," with ".". > >> > >>Is there any easy way of fixing the problem with > >>gnuplot? > > > > > > We probably need to ignore > > the locale settings for files and gnuplot. > > This is possible, I only have to look up how. > > > > > > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Statist-list mailing list > > Statist-list at intevation.de > > > https://intevation.de/mailman/listinfo/statist-list > > -- > -------------------------------------- > Dr. Andreas Beyer - beyer at imb-jena.de > http://www.imb-jena.de/~beyer > -------------------------------------- > > _______________________________________________ > Statist-list mailing list > Statist-list at intevation.de > https://intevation.de/mailman/listinfo/statist-list > _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From jaksonaquino at yahoo.com.br Fri Feb 18 22:46:59 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Fri, 18 Feb 2005 18:46:59 -0300 (ART) Subject: locale.h and gnuplot In-Reply-To: <420B75F9.5040700@imb-jena.de> Message-ID: <20050218214659.59351.qmail@web51902.mail.yahoo.com> Hello All! StatistX macro solved the problem with gnuplot. I always called the macros inside the condition #ifndef NO_GETTEXT. I deleted the following lines from the macro, but only because they seems to be written in Objective C and I don't know how to translate them. if (SX_saved_locale == NULL) { \ [NSException raise: NSMallocException \ format: @"set locale: Not enough memory."]; \ } \ I'm sending the patch against the cvs version of Statist with all changes that I made since February 4. I'm not a professional programmer, and I'm not the owner of Statist, thus, although my version of Statist is running properly, you might find it better to make some changes to improve the code or to make it more consistent with Statist "philosophy". The initial idea of simply deleting rows with missing values before calculations would be good enough for social scientists, but other people might have different problems. Now there are three ways of calling the function alloc_cols() (and, thus, getcols(), which calls alloc_cols()): - TRUE: delete rows with missing values. - FALSE: delete missing values, but not the entire row if we have more than one row. - 3: let the column as it is. That is, let it with its missing values. I used the following rule to decide whether to use getcols(i, TRUE, FALSE or 3): - If in the original source code getcols was followed by "if ((!empty) && (equal_rows(2)))", I used TRUE. - I also used TRUE if only one column was being read from tempfile (In this case the result is the same as FALSE). - If getcols was followed only by "if (!empty)", I used FALSE - for and almost the entire menu DATA MANIPULATION, I used "3" (don't delete missing values). In order to "Read another file" with column labels, I turned the variable n_alias into global, and this created a bug. Every time Statist was automatically creating a new column (like the "resi" after a regression, or out_ after deleting outliers etc...) the number of columns (ncol) was updated, but not the number of column labels (n_alias). Then, if we tried to read a new file without deleting the old data, bug! However, I soved the problem adding a lot of n_alias++ to many functions. Another problem, already present in the cvs version of statist, happened when I tried to open a file with 4 columns (two with 10 rows and two with 12 rows). The file was created by Statist with the option "save columns in ASCII file". But Statist couldn't understand " . " as a valid number. I fixed this problem. Now, Statist writes 'M's to complete the smaller rows. I ran practically all Statist menu entries, and everything seems to be OK. The only segmentation fault that I got was due to an error in my pt_BR.po. Sometimes I also ran Statist simultaneously in Portuguese and English, to compare results and graphics. One doubt that I have is whether defining missing value as (-DBL_MAX) can cause problems in functions that already use the value REAL_MIN, defined in statist.h, and that also is (-DBL_MAX). My attempt of defining missing values as (-DBL_MAX + 1) didn't really changed the value of -DBL_MAX. I wrote a draft of an future English documentation. It's incomplete, and the English is not very good, but at least someone who can read English will understand it. I used a translator to get an idea of what Dirk Melcher wrote in the German documentation. I have a lot of work for the next days. Thus, this is my last contribution for the next days (or weeks). If anyone has any doubt about any of the changes I made, please tell me. Some of the changes were explained in previous emails. Best regards, Jakson --- Andreas Beyer escreveu: > > I think I had the same problem with StatistX. If I > remember correctly, > my solution was to set the locale to the default > C-locale before passing > anything to gnuplot. Afterwards I set it back to the > previous setting. > > Andreas > > Bernhard Reiter schrieb: > > On Thu, Feb 10, 2005 at 09:37:47AM -0300, Jakson > Aquino wrote: > > > >>When trying to see a boxplot I got the following > >>gnuplot error: > >> > >>Spalte fuer Variable 1: TVHOURS > >> > >>Erstelle gnuplot-Graphik ... > >> > >> > >>0 vermutliche Ausreisser gefunden > >> > >> --- Bitte mit ENTER fortfahren ... --- > >>gnuplot> plot [0,8:5,2][9:13] > >>'/home/jakson/stat_gpl.dat' with lines > >> ^ > >> line 0: ':' or keyword 'to' expected > >> > >> > >>The problem is that, in many languages (among > them: > >>German, French, Spanish, Italian, and Portuguese), > the > >>decimal separator is not a "."; it's a ",". > However, > >>the culprit is . With locale.h, the > printf > >>function becomes very smart. It's good to see the > >>results of calculations printed with the correct > >>decimal separator, but, while writing a file to > >>gnuplot, locale.h is a problem. > >> > >>Another --far less important-- side effect of > locale.h > >>is that if we save columns in an ascii file > running > >>Statist in a language other than English, it will > not > >>be possible to open the file running Statist in > >>English. We would have, for example, 1,30000e+01 > >>instead of 1.30000e+01. Since this side effect > isn't > >>properly a bug of Statist, I think that it will be > >>enough to document it. And the solution is very > >>simple: use some tool to replace all "," with ".". > >> > >>Is there any easy way of fixing the problem with > >>gnuplot? > > > > > > We probably need to ignore > > the locale settings for files and gnuplot. > > This is possible, I only have to look up how. > > > > > > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Statist-list mailing list > > Statist-list at intevation.de > > > https://intevation.de/mailman/listinfo/statist-list > > -- > -------------------------------------- > Dr. Andreas Beyer - beyer at imb-jena.de > http://www.imb-jena.de/~beyer > -------------------------------------- > __________________________________________________ Converse com seus amigos em tempo real com o Yahoo! Messenger http://br.download.yahoo.com/messenger/ -------------- next part -------------- A non-text attachment was scrubbed... Name: statist_patch_0213.zip Type: application/x-zip-compressed Size: 20961 bytes Desc: statist_patch_0213.zip Url : http://www.intevation.de/pipermail/statist-list/attachments/20050218/7b67a825/statist_patch_0213.zip From bernhard at intevation.de Sun Feb 20 14:13:34 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sun, 20 Feb 2005 14:13:34 +0100 Subject: names of tests and get_quantile (was: gettext) In-Reply-To: <20050205222510.12991.qmail@web51902.mail.yahoo.com> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> Message-ID: <20050220131334.GD15226@intevation.de> Am 5. Feb 2005 um 19:25:10 schrieb Jakson Aquino: > While translating the messages, I found some possible > problems in the English original ones. > I had also used Aspell to find misspelled words in > statist.pot and I corrected them in the source code: The following two are names fo special tests, so I had to look them up and found that the new version is better in both cases. > Lillefors => Lilliefors > SPEARMANN => SPEARMAN I have to check the following change more thorougly as "quantile" is a statistical term and is correct. E.g. see: http://mathworld.wolfram.com/Quantile.html The word quantile has two meanings in probability. Specific elements x in the range of a variate X are called quantiles, and denoted x (Evans et al. 2000, p. 5). The kth n-tile is that value of x, say , which corresponds to a cumulative frequency of (Kenney and Keeping 1962). If n = 4, the quantity is called a quartile, and if n = 100, it is called a percentile. > get_quantile => get_quartile > > Note: The last one (get_quantile) is the name of a > function, not a message. I have to check later what that function does anywhere, it does not seem to get called. Maybe it can be cleaned out... -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050220/0c791ef1/attachment.bin From bernhard at intevation.de Sun Feb 20 14:39:52 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sun, 20 Feb 2005 14:39:52 +0100 Subject: gettext In-Reply-To: <20050205222510.12991.qmail@web51902.mail.yahoo.com> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> Message-ID: <20050220133952.GE15226@intevation.de> Am 5. Feb 2005 um 19:25:10 schrieb Jakson Aquino: Still on the list of things to be checked in is: > I was getting "?" instead of accentuated letters when > statist was running in Portuguese. I just replaced > "LC_MESSAGES" with "LC_ALL" and now the accents are > correctly displayed. Even r^2 is being nicely printed > as r? (r + 0xb2) and mu is now ? (0xb5). I want to do one commit with the solution to the gnuplot and data files formats. > While translating the messages, I found some possible > problems in the English > original ones. I, then, made small changes in the > source code, and I'm > submitting these changes in the statist_gettext_patch: In CVS now, except the get_quantile() change; I have found some places where I could make some minor improvements (I hope). Thanks again for the patches, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050220/00a40126/attachment.bin From bernhard at intevation.de Sun Feb 20 14:54:59 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sun, 20 Feb 2005 14:54:59 +0100 Subject: gettext In-Reply-To: <20050205222510.12991.qmail@web51902.mail.yahoo.com> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> Message-ID: <20050220135459.GF15226@intevation.de> Am 5. Feb 2005 um 19:25:10 schrieb Jakson Aquino: > The po directory has a Makefile. Typing "make po" > extracts the strings from the source files and updates > (msgmerge) the three .po files (pt_BR, es_ES, and > de_DE). Which version of msgmerge is needed at least? msgmerge (GNU gettext) 0.10.40 does not have the -U flag, and msgmerge (GNU gettext-tools) 0.13.1 has it. > Typing "make mo" compiles the .po files and > creates the .mo files, and typing "make install" > copies the .mo files to > /usr/share/locale/xx_XX/LC_MESSAGES/. Typing "make" > with no options prints a short help text. Another possible improvement would be to use one place to configure the installation prefix "/usr/share/locale" and document it. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050220/70b90876/attachment.bin From bernhard at intevation.de Sun Feb 20 15:03:33 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sun, 20 Feb 2005 15:03:33 +0100 Subject: gettext In-Reply-To: <20050205222510.12991.qmail@web51902.mail.yahoo.com> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> Message-ID: <20050220140333.GG15226@intevation.de> Am 5. Feb 2005 um 19:25:10 schrieb Jakson Aquino: > Since the po directory is not in the cvs yet, I > created one with Portuguese, Spanish and German > translations. I checked in your improved po files and the directory. They will be out of date a bit, because of some minor changes I did. This is untested by me, as on the machine I did not have a very mondern gettext and there is no mechanism to check this without installing it as root. Jakson: If you like, I can give you CVS writte access, this probably will make it easier for you to update the translations. Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050220/74630659/attachment.bin From beyer at imb-jena.de Mon Feb 21 09:40:40 2005 From: beyer at imb-jena.de (Andreas Beyer) Date: Mon, 21 Feb 2005 09:40:40 +0100 Subject: locale.h and gnuplot In-Reply-To: <20050218214659.59351.qmail@web51902.mail.yahoo.com> References: <20050218214659.59351.qmail@web51902.mail.yahoo.com> Message-ID: <42199E88.2070008@imb-jena.de> Jakson Aquino schrieb: > StatistX macro solved the problem with gnuplot. I > always called the macros inside the condition #ifndef > NO_GETTEXT. I deleted the following lines from the > macro, but only because they seems to be written in > Objective C and I don't know how to translate them. > > if (SX_saved_locale == NULL) { \ > [NSException raise: NSMallocException \ > format: @"set locale: Not enough memory."]; \ > } \ > When you store the current locale information, the program could potentially run out of memory. You have to check that, which StatistX does by testing the condition SX_saved_locale == NULL If this condition is true, the program has no memory left (or something else went wrong). statist should perform the same test at the same place and handle the problem in some way. E.g. the following is a possible way to deal with it: if (saved_locale == NULL) { /* Run out of memory. Fatal, stop the program. */ out_err(FAT, ERR_FILE, ERR_LINE, _("set locale: Not enough memory.")); } Look at "statist.h" for details about out_err(). If possible, one could also use mymalloc() from memory_handling.h. Andreas From bernhard at intevation.de Mon Feb 21 22:50:04 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 21 Feb 2005 22:50:04 +0100 Subject: LC_NUMERC? (was: locale.h and gnuplot) In-Reply-To: <420B75F9.5040700@imb-jena.de> References: <20050210123747.76023.qmail@web51901.mail.yahoo.com> <20050210124524.GA18784@intevation.de> <420B75F9.5040700@imb-jena.de> Message-ID: <20050221215004.GA28609@intevation.de> Am 10. Feb 2005 um 15:55:53 schrieb Andreas Beyer: > > I think I had the same problem with StatistX. If I remember correctly, > my solution was to set the locale to the default C-locale before passing > anything to gnuplot. Afterwards I set it back to the previous setting. Another way to solve this is to only set the LC_CTYPE and LC_MESSAGES locales and keep all printf and scanf functions to use the "C" locale instead. (Because this is a small change, I just checked it into CVS.) The question is: What do we want? Gnuplot command: needs the "C" locale for sure. Data files: I would hope to keep them interoperable internationally, so I tend to prefer "C" locale here, too. Number display during interactive use: When the datafile is international I also think it is reasonable to keep the numbers printed during the interactive sessions international. There is another point in that people that are used to the old statist might be surprised by the new behaviour if we also changed LC_NUMERC. Jakson, Andreas: What do you think? Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050221/db40b7ea/attachment.bin From bernhard at intevation.de Mon Feb 21 23:31:31 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 21 Feb 2005 23:31:31 +0100 Subject: CVS: Makefile improvements Message-ID: <20050221223131.GA29670@intevation.de> I have made a couple of improvements to the build system in CVS. Now it is a lot easier for me to test the translations and I saw statist in Portugese and Spanish: Very nice! -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050221/457ef26a/attachment.bin From beyer at imb-jena.de Tue Feb 22 09:14:32 2005 From: beyer at imb-jena.de (Andreas Beyer) Date: Tue, 22 Feb 2005 09:14:32 +0100 Subject: LC_NUMERC? In-Reply-To: <20050221215004.GA28609@intevation.de> References: <20050210123747.76023.qmail@web51901.mail.yahoo.com> <20050210124524.GA18784@intevation.de> <420B75F9.5040700@imb-jena.de> <20050221215004.GA28609@intevation.de> Message-ID: <421AE9E8.9060200@imb-jena.de> I guess the American standard (== C-locale) is virtually an international standard. So, we might go for the C-locale in case of LC_NUMERIC. In any case this should only be the default! The makefiles should have an option to set the locale explicitely. Thus, in order to communicate with gnuplot we would still have to set the C-locale explicitely (and set it back afterwards). The makefile should allow for something like # set this to 0 to use your local number representation in data files USE_C_LOCALE 1 [...] gcc [...] $(USE_C_LOCALE) There could be people relying on the fact that statist uses their own locale when reading data. E.g. MSExcel uses the country specific settings. If someone exports data from there, he/she will not necessarily get files that are compatible with the C-locale. Anyway it seems better not to change the entire locale when sending data to gnuplot. Otherwise also (error-)messages would be printed in English during this phase. That would be a petty after all these translation efforts. ;-) Thus, we should definitely only change LC_NUMERIC. Andreas Bernhard Reiter schrieb: > Am 10. Feb 2005 um 15:55:53 schrieb Andreas Beyer: > >>I think I had the same problem with StatistX. If I remember correctly, >>my solution was to set the locale to the default C-locale before passing >>anything to gnuplot. Afterwards I set it back to the previous setting. > > > Another way to solve this is to only set the LC_CTYPE and LC_MESSAGES > locales and keep all printf and scanf functions to use the "C" locale instead. > (Because this is a small change, I just checked it into CVS.) > > The question is: What do we want? > > Gnuplot command: needs the "C" locale for sure. > > Data files: I would hope to keep them interoperable internationally, > so I tend to prefer "C" locale here, too. > > Number display during interactive use: When the datafile is international > I also think it is reasonable to keep the numbers printed during the > interactive sessions international. > > There is another point in that people that are used to the old statist > might be surprised by the new behaviour if we also changed LC_NUMERC. > > Jakson, Andreas: What do you think? > > Bernhard > > > ------------------------------------------------------------------------ > > _______________________________________________ > Statist-list mailing list > Statist-list at intevation.de > https://intevation.de/mailman/listinfo/statist-list From bernhard at intevation.de Tue Feb 22 10:19:22 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 22 Feb 2005 10:19:22 +0100 Subject: LC_NUMERC? In-Reply-To: <421AE9E8.9060200@imb-jena.de> References: <20050210123747.76023.qmail@web51901.mail.yahoo.com> <20050210124524.GA18784@intevation.de> <420B75F9.5040700@imb-jena.de> <20050221215004.GA28609@intevation.de> <421AE9E8.9060200@imb-jena.de> Message-ID: <20050222091922.GE7178@intevation.de> Am 22. Feb 2005 um 09:14:32 schrieb Andreas Beyer: > I guess the American standard (== C-locale) is virtually an > international standard. So, we might go for the C-locale in case of > LC_NUMERIC. In any case this should only be the default! This is the current situation in CVS, after I read up on that locale documentation. > The makefiles > should have an option to set the locale explicitely. Even better we could make it a command line option, then people can decide each time they call statist. Or do you think that would be worse? > Thus, in order to > communicate with gnuplot we would still have to set the C-locale > explicitely (and set it back afterwards). Yes and LC_NUMERIC should be enough for this. Anybody wants to do the patch? :) Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050222/0a37fc1d/attachment.bin From jaksonaquino at yahoo.com.br Tue Feb 22 15:21:02 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 22 Feb 2005 11:21:02 -0300 (ART) Subject: names of tests and get_quantile (was: gettext) In-Reply-To: <20050220131334.GD15226@intevation.de> Message-ID: <20050222142102.86771.qmail@web51909.mail.yahoo.com> Hi Bernhard, I didn't realized it. But, in fact, "grep quantile *.c" returns only one occurrence of "get_quantile". And, what get_quantile() does can also be done with percentiles(), which is called by "Miscellaneous | Percentiles". Certainly get_quantile() can be cleaned out! Jakson --- Bernhard Reiter escreveu: > I have to check later what that function does > anywhere, > it does not seem to get called. Maybe it can be > cleaned out... _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From jaksonaquino at yahoo.com.br Tue Feb 22 15:31:13 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 22 Feb 2005 11:31:13 -0300 (ART) Subject: LC_NUMERC? In-Reply-To: <20050222091922.GE7178@intevation.de> Message-ID: <20050222143113.96543.qmail@web51902.mail.yahoo.com> Hello All, I tested the new cvs version of Statist, and setting LC_TYPE and LC_MESSAGES, the accents are being correctly displayed. But I still prefer to see the decimal separator being a comma when the output is in Portuguese. My suggestion is to set the locale to "C" before reading data files and writing columns, and re-set it back to saved_locale after these operations. Only plot.c needs a lot of set and reset, but I already did it in the patch of last week. If this solution is accepted, I think that we have only to set & reset locale in two other places: readsourcefile() (data.c) and menue.c ("Data management | Export columns as ASCII-data"). The Statist documentation should warn the user that the data file uses "dots" as decimal separators, but that the results are internationalized. We could also add a menu option like "Data management | Export columns as ASCII-data using current locale". The documentation would explain what this means. The last week patch regarding the problem with missing values was made against the old cvs version of Statist. I don't have experience with diff, but since the cvs was modified, and because you made corrections in some of my changes, I guess that the patch is no more applicable. If this is the case, I can prepare the patch again. That is, I will - run diff to compare the current cvs version with my version of statist. - manually do in my version of statist the corrections that you did in cvs. - run diff again Jakson Best, Jakson --- Bernhard Reiter escreveu: > Am 22. Feb 2005 um 09:14:32 schrieb Andreas Beyer: > > I guess the American standard (== C-locale) is > virtually an > > international standard. So, we might go for the > C-locale in case of > > LC_NUMERIC. In any case this should only be the > default! > > This is the current situation in CVS, > after I read up on that locale documentation. > > > The makefiles > > should have an option to set the locale > explicitely. > > Even better we could make it a command line option, > then people can decide each time they call statist. > Or do you think that would be worse? > > > Thus, in order to > > communicate with gnuplot we would still have to > set the C-locale > > explicitely (and set it back afterwards). > > Yes and LC_NUMERIC should be enough for this. > Anybody wants to do the patch? :) > > Bernhard > > ATTACHMENT part 1.2 application/pgp-signature > _______________________________________________ > Statist-list mailing list > Statist-list at intevation.de > https://intevation.de/mailman/listinfo/statist-list _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From jaksonaquino at yahoo.com.br Tue Feb 22 15:33:57 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 22 Feb 2005 11:33:57 -0300 (ART) Subject: locale.h and gnuplot In-Reply-To: <42199E88.2070008@imb-jena.de> Message-ID: <20050222143357.97600.qmail@web51902.mail.yahoo.com> Hi Andreas, Thank you for completing the code. Next time I send a patch, I can make the changes. I think that the code below now is correct, right? #ifndef NO_GETTEXT char * SX_saved_locale, * SX_old_locale; #define SX_SET_LOCALE \ SX_old_locale = setlocale (LC_NUMERIC, NULL); \ SX_saved_locale = mymalloc(strlen(SX_old_locale)+1); \ strcpy(SX_saved_locale, SX_old_locale); \ setlocale (LC_NUMERIC, "C"); #define SX_RESET_LOCALE \ setlocale (LC_NUMERIC, SX_saved_locale); \ free(SX_saved_locale); SX_saved_locale = NULL; #endif I think that we can put the above lines in statist.h. Best regards, Jakson --- Andreas Beyer escreveu: > When you store the current locale information, the > program could > potentially run out of memory. You have to check > that, which StatistX > does by testing the condition > SX_saved_locale == NULL > If this condition is true, the program has no memory > left (or something > else went wrong). statist should perform the same > test at the same place > and handle the problem in some way. E.g. the > following is a possible way > to deal with it: > > if (saved_locale == NULL) { > /* Run out of memory. Fatal, stop the program. > */ > out_err(FAT, ERR_FILE, ERR_LINE, > _("set locale: Not enough memory.")); > } > > > Look at "statist.h" for details about out_err(). If > possible, one could > also use mymalloc() from memory_handling.h. > > Andreas _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From jaksonaquino at yahoo.com.br Tue Feb 22 16:00:13 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 22 Feb 2005 12:00:13 -0300 (ART) Subject: CVS: Makefile improvements In-Reply-To: <20050221223131.GA29670@intevation.de> Message-ID: <20050222150013.82301.qmail@web51901.mail.yahoo.com> Hi Bernhard, >Jakson: If you like, I can give you CVS writte access, this probably >will make it easier for you to update the translations. Yes, I do. I can update the .po files before the next release of Statist. >Which version of msgmerge is needed at least? >msgmerge (GNU gettext) 0.10.40 >does not have the -U flag, >and msgmerge (GNU gettext-tools) 0.13.1 has it. I'm sorry, but I don't know. I use Fedora Core 2, which comes with msgmerge (GNU gettext-tools) 0.14.1. I wrote the Makefile for the po directory after reading some tutorials on gettext and the gettext info pages. Looking at the po directory of some programs, I could see that they have a very complicated Makefile (probably written automatically by some tool). The po directories also come with .gmo files. Regarding the .gmo files, the gettext info page says: "Files ending with `.gmo' are really MO files, when it is known that these files use the GNU format" Since some people might not have the gettext tools, I can change the Makefile to produce the object files with the extension .gmo, and, install them as .mo. While shipping Statist, we could include the already compiled .gmo files. >Another possible improvement would be to use one place >to configure the installation prefix "/usr/share/locale" >and document it. Well, you already solved this with the Makefile.inc. The ideal is that the user can choose the PREFIX and everything else becomes automatically set. Perhaps we still can make some small changes in the Makefiles: Makefile.inc ------------------------------------------------------ #if you have root privileges choose one of the following prefixes: #PREFIX = /usr/local PREFIX = /usr #if you can't become root, uncomment the following line: #PREFIX = $$HOME ------------------------------------------------------ src/Makefile ------------------------------------------------------ INSTALLDIR = $(PREFIX)/bin LOCALEDIR = $(PREFIX)/share/locale DOCDIR = $(PREFIX)/share/doc/statist EXTRACFLAGS = -DLOCALEDIR='"$(LOCALE)"' >> PROBLEM HERE! << install: $(MAIN) install -d $(PREFIX)/bin install -s statist $(PREFIX)/bin/ (cd ../doc; make PREFIX=$(PREFIX) install) (cd ../po; make PREFIX=$(LOCALE) install) uninstall: rm -f $(PREFIX)/bin/statist (cd ../doc; make PREFIX=$(PREFIX) uninstall) (cd ../po; make PREFIX=$(LOCALE) uninstall) ------------------------------------------------------ po/Makefile: ------------------------------------------------------ #PREFIX = $(PREFIX)/share/locale POSTFIX = /LC_MESSAGES/ POSTFIXMO = $(POSTFIX)/statist.mo install: install -d $(PREFIX)/pt_BR$(POSTFIX) install -d $(PREFIX)/de$(POSTFIX) install -d $(PREFIX)/es$(POSTFIX) install -d $(PREFIX)/it$(POSTFIX) cp pt_BR.gmo $(PREFIX)/pt_BR$(POSTFIXMO) cp de_DE.gmo $(PREFIX)/de$(POSTFIXMO) cp es_ES.gmo $(PREFIX)/es$(POSTFIXMO) cp it_IT.gmo $(PREFIX)/it$(POSTFIXMO) uninstall: rm -f $(PREFIX)/pt_BR$(POSTFIXMO) rm -f $(PREFIX)/de$(POSTFIXMO) rm -f $(PREFIX)/es$(POSTFIXMO) rm -f $(PREFIX)/it$(POSTFIXMO) ------------------------------------------------------ The only problem that I got with the above lines was that EXTRACFLAGS = -DLOCALEDIR='"$(LOCALE)"' didn't work. That's, EXTRACFLAGS didn't become "/usr/share/locale". If the user isn't using the default "/usr" prefix, it'll have to set both PREFIX and EXTRACFLAGS. Best, Jakson --- Bernhard Reiter escreveu: > I have made a couple of improvements to the build > system in CVS. > Now it is a lot easier for me to test the > translations > and I saw statist in Portugese and Spanish: Very > nice! > _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From bernhard at intevation.de Tue Feb 22 16:25:17 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 22 Feb 2005 16:25:17 +0100 Subject: Write access (was: CVS: Makefile improvements) In-Reply-To: <20050222150013.82301.qmail@web51901.mail.yahoo.com> References: <20050221223131.GA29670@intevation.de> <20050222150013.82301.qmail@web51901.mail.yahoo.com> Message-ID: <20050222152517.GI20915@intevation.de> Am 22. Feb 2005 um 12:00:13 schrieb Jakson Aquino: > >Jakson: If you like, I can give you CVS writte > access, this probably > >will make it easier for you to update the > translations. > > Yes, I do. Send me your ssh1 key in personal email. The access is similiar to http://freegis.org/grass/howto_grass-sshcvs.en.html Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050222/3d4d1cba/attachment.bin From bernhard at intevation.de Tue Feb 22 17:19:59 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 22 Feb 2005 17:19:59 +0100 Subject: CVS: Makefile improvements In-Reply-To: <20050222150013.82301.qmail@web51901.mail.yahoo.com> References: <20050221223131.GA29670@intevation.de> <20050222150013.82301.qmail@web51901.mail.yahoo.com> Message-ID: <20050222161959.GJ20915@intevation.de> Hi Jakson, Am 22. Feb 2005 um 12:00:13 schrieb Jakson Aquino: > Since some people might not have the gettext tools, I > can change the Makefile to produce the object files > with the extension .gmo, and, install them as .mo. > While shipping Statist, we could include the already > compiled .gmo files. Yes, we would make to create the .gmo or .mo files before packing the tarball. A "dist" target would be perfect for this. > >Another possible improvement would be to use one > place > >to configure the installation prefix > "/usr/share/locale" > >and document it. > > Well, you already solved this with the Makefile.inc. > The ideal is that the user can choose the PREFIX and > everything else becomes automatically set. Yes, I solved this after writing the email. We might want to think about calling the po/Makefile from the src/Makefile to unify this further. There is no real need to decend to two directories. > Perhaps we > still can make some small changes in the Makefiles: > > > Makefile.inc > ------------------------------------------------------ > #if you have root privileges choose one of the > following prefixes: > #PREFIX = /usr/local > PREFIX = /usr > > #if you can't become root, uncomment the following > line: > #PREFIX = $$HOME I think we should better explain this "root" non-root issues in the documentation, as those instructions are generic and will be known to many administrators. > src/Makefile > ------------------------------------------------------ > INSTALLDIR = $(PREFIX)/bin > LOCALEDIR = $(PREFIX)/share/locale > DOCDIR = $(PREFIX)/share/doc/statist > > EXTRACFLAGS = -DLOCALEDIR='"$(LOCALE)"' >> PROBLEM > HERE! << Did you mean EXTRACFLAGS = -DLOCALEDIR='"$(LOCALEDIR)"' ? > The only problem that I got with the above lines was > that EXTRACFLAGS = -DLOCALEDIR='"$(LOCALE)"' didn't > work. That's, EXTRACFLAGS didn't become > "/usr/share/locale". Just from what I saw: Where did you define LOCALE? ;) Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050222/7cee1bdf/attachment.bin From bernhard at intevation.de Tue Feb 22 17:32:05 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 22 Feb 2005 17:32:05 +0100 Subject: LC_NUMERC? In-Reply-To: <20050222143113.96543.qmail@web51902.mail.yahoo.com> References: <20050222091922.GE7178@intevation.de> <20050222143113.96543.qmail@web51902.mail.yahoo.com> Message-ID: <20050222163205.GA27768@intevation.de> Am 22. Feb 2005 um 11:31:13 schrieb Jakson Aquino: > I tested the new cvs version of Statist, and setting > LC_TYPE and LC_MESSAGES, the accents are being > correctly displayed. But I still prefer to see the > decimal separator being a comma when the output is in > Portuguese. What do you think of a command line option? > The Statist documentation should warn the user that > the data file uses "dots" as decimal separators, but > that the results are internationalized. Yes, in any case: It needs to be documented. > We could also add a menu option like "Data management > | Export columns as ASCII-data using current locale". > The documentation would explain what this means. Good idea. > The last week patch regarding the problem with missing > values was made against the old cvs version of > Statist. I don't have experience with diff, but since > the cvs was modified, and because you made corrections > in some of my changes, I guess that the patch is no > more applicable. It should be doable for me to apply the patches. Patch uses a fuzzy algorithm and still works when other parts of the file have been changed. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050222/ec352ff3/attachment.bin From jaksonaquino at yahoo.com.br Tue Feb 22 20:56:02 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 22 Feb 2005 16:56:02 -0300 (ART) Subject: LC_NUMERC? In-Reply-To: <20050222163205.GA27768@intevation.de> Message-ID: <20050222195602.25451.qmail@web51908.mail.yahoo.com> Hi Bernhard, --- Bernhard Reiter escreveu: > Am 22. Feb 2005 um 11:31:13 schrieb Jakson Aquino: > What do you think of a command line option? > The user that I imagine is some student of social science that is migrating from Windows. He will use Statist only some times in an year, and will have problems memorizing commands. I think that it's a good idea to have the options in command line, but I also think that the same options can also be in some menu option like "Main menu | Options". The options could be toogled on/off from the menu. My preference is that, as its default behavior, Statist saves database in "C" locale, and shows results and saves logfiles with LC_NUMERIC set to current locale. The user would interact with a program fully running in its language, but databases in "C" locale has some adavantages: portability among people in different countries, and compatibility with gnuplot. People who knows gnuplot commands can save columns as ASCII file and do something with them using gnuplot without any need of changing the standard behavior of Statist. > It should be doable for me to apply the patches. > Patch uses a fuzzy algorithm and still works when > other parts > of the file have been changed. Very good! It would be boring to repeat the same work again :) Best, Jakson __________________________________________________ Converse com seus amigos em tempo real com o Yahoo! Messenger http://br.download.yahoo.com/messenger/ From bernhard at intevation.de Tue Feb 22 21:00:50 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 22 Feb 2005 21:00:50 +0100 Subject: names of tests and get_quantile (was: gettext) In-Reply-To: <20050222142102.86771.qmail@web51909.mail.yahoo.com> References: <20050220131334.GD15226@intevation.de> <20050222142102.86771.qmail@web51909.mail.yahoo.com> Message-ID: <20050222200050.GJ27768@intevation.de> Am 22. Feb 2005 um 11:21:02 schrieb Jakson Aquino: > I didn't realized it. But, in fact, "grep quantile > *.c" returns only one occurrence of "get_quantile". > And, what get_quantile() does can also be done with > percentiles(), which is called by "Miscellaneous | > Percentiles". > > Certainly get_quantile() can be cleaned out! I checked back to 0.11.b010 and even it was not used. So removed now in CVS. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050222/b80960d6/attachment.bin From bernhard at intevation.de Wed Feb 23 19:30:00 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 23 Feb 2005 19:30:00 +0100 Subject: lang_defs.h In-Reply-To: <20050207095244.GE30059@intevation.de> References: <20050206214504.96312.qmail@web51901.mail.yahoo.com> <20050207095244.GE30059@intevation.de> Message-ID: <20050223183000.GB15659@intevation.de> Am 7. Feb 2005 um 10:52:44 schrieb Bernhard Reiter: > On Sun, Feb 06, 2005 at 06:45:04PM -0300, Jakson Aquino wrote: > > However, it's not practical to have messages written > > directly in the source code in more than two > > languages. Thus, I think that we don't need the file > > lang_defs.h anymore. I'm submitting a patch where, > > with some deletions and with no addition, the program > > is compiling as before, but without using the file > > lang_defs.h. > I'll look at the patch, > it is probably fine to just have two possibilities > in the source code and simplify it. Hi Jakson, the patch is in CVS now. Note that you have missed the DEBUG section in statist.c which did not directly compile when debugging is turned on with you patch. ;) Thanks again for contributing of course. There are still more patches from you in the queue, I know. :) Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050223/5500d8a1/attachment.bin From bernhard at intevation.de Wed Feb 23 19:30:42 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 23 Feb 2005 19:30:42 +0100 Subject: Italian translation (Re: gettext) In-Reply-To: <20050207100107.GG30059@intevation.de> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> <20050207100107.GG30059@intevation.de> Message-ID: <20050223183042.GC15659@intevation.de> Hi Jakson, any news on the Italian translation? E.g. are you planning to add the corresponding po file? Bernhard Am 7. Feb 2005 um 11:01:07 schrieb Bernhard Reiter: > On Sat, Feb 05, 2005 at 07:25:10PM -0300, Jakson Aquino wrote: > > I knowthat there is an Italian translation. > > How can I get it? > > italian_translation.patch > Italien language support by Nicola Intini > Note: add -p5 as option to patch when you are in the statist-1.0.0 dir > > http://www.usf.uni-osnabrueck.de/~breiter/tools/statist/v1.0.0/contrib/ -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050223/b48d1802/attachment.bin From jaksonaquino at yahoo.com.br Thu Feb 24 13:33:01 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Thu, 24 Feb 2005 09:33:01 -0300 (ART) Subject: Italian translation (Re: gettext) In-Reply-To: <20050223183042.GC15659@intevation.de> Message-ID: <20050224123301.23209.qmail@web51909.mail.yahoo.com> Hi Bernhard, I don't have the Italian translation here in this computer, but some days ago I put it in: http://www.geocities.com/jakson-aquino/statist.html Almost all patches are there too. Perhaps I forgot to tell this in an email :) Best, Jakson --- Bernhard Reiter escreveu: > Hi Jakson, > > any news on the Italian translation? > E.g. are you planning to add the corresponding po > file? > > Bernhard > _______________________________________________________ Yahoo! Acesso Gr?tis - Instale o discador do Yahoo! agora. http://br.acesso.yahoo.com/ - Internet r?pida e gr?tis From bernhard at intevation.de Thu Feb 24 17:05:17 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Thu, 24 Feb 2005 17:05:17 +0100 Subject: Italian translation In-Reply-To: <20050223183042.GC15659@intevation.de> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> <20050207100107.GG30059@intevation.de> <20050223183042.GC15659@intevation.de> Message-ID: <20050224160517.GA23009@intevation.de> Am 23. Feb 2005 um 19:30:42 schrieb Bernhard Reiter: > any news on the Italian translation? Jakson had put it on http://www.geocities.com/jakson-aquino/statist.html I just added it to CVS (and improved the Makefile slightly while doing so.) -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050224/1519c318/attachment.bin From nicola at intini.org Mon Feb 28 00:30:17 2005 From: nicola at intini.org (nicola intini) Date: Mon, 28 Feb 2005 00:30:17 +0100 Subject: Italian translation In-Reply-To: <20050224160517.GA23009@intevation.de> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> <20050223183042.GC15659@intevation.de> <20050224160517.GA23009@intevation.de> Message-ID: <200502280030.29349.nicola@intini.org> Hi statist developers do you need some help for the italian translation? I see that the big work has been done (putting all in gettext). I was not able to do it. If you need some help in translating or completing... I will take a look. Ciao Nicola Alle 17:05, gioved? 24 febbraio 2005, Bernhard Reiter ha scritto: > Am 23. Feb 2005 um 19:30:42 schrieb Bernhard Reiter: > > any news on the Italian translation? > > Jakson had put it on > http://www.geocities.com/jakson-aquino/statist.html > > I just added it to CVS (and improved the Makefile slightly while doing so.) -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20050228/cd321bf9/attachment.bin From bernhard at intevation.de Mon Feb 28 10:15:19 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 28 Feb 2005 10:15:19 +0100 Subject: Italian translation In-Reply-To: <200502280030.29349.nicola@intini.org> References: <20050205222510.12991.qmail@web51902.mail.yahoo.com> <20050223183042.GC15659@intevation.de> <20050224160517.GA23009@intevation.de> <200502280030.29349.nicola@intini.org> Message-ID: <20050228091519.GA506@intevation.de> Hi Nicola, Am 28. Feb 2005 um 00:30:17 schrieb nicola intini: > do you need some help for the italian translation? > I see that the big work has been done (putting all in gettext). Yes, thanks to Jakson! > If you need some help in translating or completing... > I will take a look. I think that I first need to * integrate Jakson's proposal for handling of LC_NUMERIC * integrate Jakson's changes for missing value handling * try to autogenerate the German .po from the source and then we are ready for a new release of statist and would welcome updates of each translation. Of course you can start earlier, if you wish. ;) Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050228/2169f4f8/attachment.bin From jaksonaquino at yahoo.com.br Mon Mar 14 23:50:30 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 14 Mar 2005 19:50:30 -0300 (ART) Subject: patches Message-ID: <20050314225031.47317.qmail@web51902.mail.yahoo.com> Hi Bernhard, Did you have time to check the patch regarding the missing values? Did you find serious problems in them or in any other patch? Best, Jakson __________________________________________________ Converse com seus amigos em tempo real com o Yahoo! Messenger http://br.download.yahoo.com/messenger/ From jaksonaquino at yahoo.com.br Tue Mar 15 15:33:09 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 15 Mar 2005 11:33:09 -0300 (ART) Subject: documentation Message-ID: <20050315143309.20685.qmail@web51904.mail.yahoo.com> Hello All, I'm sending as attachment a formatted version of the English documentation draft that I have sent some weeks ago to this list. The file is in latex format, but there is also a translation into html made with tth. The draft needs be revised. Please, feel free to correct or add anything, including grammar corrections (my English is far from perfect). The draft is also incomplete: some sections of the German documentation have no correspondent in the English one, and others can become more comprehensive. Best, Jakson Yahoo! Mail - Com 250MB de espa?o. Abra sua conta! http://mail.yahoo.com.br/ -------------- next part -------------- A non-text attachment was scrubbed... Name: statist_manual.zip Type: application/x-zip-compressed Size: 14509 bytes Desc: statist_manual.zip Url : http://www.intevation.de/pipermail/statist-list/attachments/20050315/5ac288de/statist_manual.zip From bernhard at intevation.de Wed Mar 16 10:25:49 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 16 Mar 2005 10:25:49 +0100 Subject: patches In-Reply-To: <20050314225031.47317.qmail@web51902.mail.yahoo.com> References: <20050314225031.47317.qmail@web51902.mail.yahoo.com> Message-ID: <20050316092549.GC26428@intevation.de> Hi Jakson, Am 14. Mar 2005 um 19:50:30 schrieb Jakson Aquino: > Did you have time to check the patch regarding the > missing values? Did you find serious problems in them > or in any other patch? unfortunately I did not yet find the time to do it. (It is CeBIT time here in Germany.) I hope for next week. Sorry for taking so long, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050316/90e85337/attachment.bin From bernhard at intevation.de Tue Mar 22 18:11:30 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 22 Mar 2005 18:11:30 +0100 Subject: documentation In-Reply-To: <20050315143309.20685.qmail@web51904.mail.yahoo.com> References: <20050315143309.20685.qmail@web51904.mail.yahoo.com> Message-ID: <20050322171130.GH9921@intevation.de> Hi Jakson, Am 15. Mar 2005 um 11:33:09 schrieb Jakson Aquino: > I'm sending as attachment a formatted version of the > English documentation draft that I have sent some > weeks ago to this list. just managed to check this into CVS today. Did some renaming and used "Free Software" instead of open source, because I personally prefer that term. ;) > The file is in latex format, > but there is also a translation into html made with > tth. I found that tth is non-free, we could juse latex2html instead, but testing revealed that pdflatex works on this. So I have put this in the readme. There is no question who is the author of this manual right now: It's you. If someone else makes significant contributions, we can add them as author. Thanks again for this manual, it is a good start, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050322/759ea0cf/attachment.bin From jaksonaquino at yahoo.com.br Wed Jun 22 17:00:41 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Wed, 22 Jun 2005 12:00:41 -0300 Subject: patches In-Reply-To: <20050316092549.GC26428@intevation.de> References: <20050314225031.47317.qmail@web51902.mail.yahoo.com> <20050316092549.GC26428@intevation.de> Message-ID: <200506221200.41105.jaksonaquino@yahoo.com.br> Hi Bernhard! How are you? Did you have time to evaluate my patches? Did anyone else test the patches? I'd like to propose the addition of some new features to Statist that will be useful for social scientists, but first I'd like to know if there are problems with my previous patches (or if there is a better solution to the missing values problem). Best, Jakson Em Quarta 16 Mar?o 2005 06:25, Bernhard Reiter escreveu: > Hi Jakson, > > Am 14. Mar 2005 um 19:50:30 schrieb Jakson Aquino: > > Did you have time to check the patch regarding the > > missing values? Did you find serious problems in them > > or in any other patch? > > unfortunately I did not yet find the time to do it. > (It is CeBIT time here in Germany.) > I hope for next week. > > Sorry for taking so long, > Bernhard _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Fri Jun 24 12:51:24 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Fri, 24 Jun 2005 07:51:24 -0300 Subject: recode, get sample, and get columns Message-ID: <20050624105124.GA1933@localhost.localdomain> Hello All! I developed some tools for Statist while trying to use the PNAD database. PNAD is a national survey made by the Brazilian government almost every year. The resulting fixed-width database has more than 300 variables and 360000 cases, and it probably is the most popular database among Brazilian social scientists. I created a program to extract variables from a fixed width database (xcols.c), another to extract a sample of the rows of a Statist data file (xsample.c), and a third one to recode a Statist data file (recode.c). I also wrote the first draft documentation for the three programs. Although many researchers in the natural sciences might never need any program like these three, I believe that they are undispensable for a social scientist. If the new code from recode.c, xsample.c and xcols.c is added to Statist, we can give to Statist users two options: (1) invoke Statist with options -r, -s, or -x, and (2) use the menu to choose "recode", "create a sample" or "extract columns". xcols.c and xsample.c are very small, and if they are added to Statist, they can become functions of data.c. recode.c is bigger and more complex, and, probably, it would be better to keep it as a separate file. If the code is accepted as part of Statist, at least some small adaptations are necessary, like using out_r, mymalloc, FREAD, and FWRITE, instead of printf, malloc, fread and fwrite. Comments and bug reports are welcome! Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: statist_tools.tar.gz Type: application/x-gzip Size: 6981 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050624/1754e3f9/statist_tools.tar.gz From jaksonaquino at yahoo.com.br Sat Jul 2 10:42:41 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sat, 2 Jul 2005 05:42:41 -0300 Subject: translating German comments Message-ID: <200507020542.55840.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello all! I'm trying to translate some German comments (not the messages) that are in the source code to make it easier for people who don't speak German (including myself) to understand the code. Please, could anyone revise the following translations? The lines below were extracted from a patch generated with diff, but you can correct them directly in a reply to this email, and I'll apply the changes. Thank you! Jakson menue.c @@ -55,12 +55,13 @@ - - Prueft, ob Spalten unterschiedliche Laenge besitzen + Test whether columns have different lengths ==================================================================== procs.c @@ -45,8 +45,8 @@ - - berechnet partielle korrelation; maximal 3 feste korrelationen + calculate partial correlation; at most 3 constant correlations @@ -777,7 +779,7 @@ - - Test auf gleiche Effektwahrscheinlichkeiten bzw. gleiche Dosen + Test on same effect-probability as well as same dosis @@ -1612,7 +1614,7 @@ - - kritische Werte fuer 95% Konfidenzinterval des Medians aus NEAVE: + critical values for 95% confidence interval of the Medians (see NEAVE): @@ -1784,8 +1786,8 @@ - - kritische roh-Werte fuer signifikanzniveaus 1%, 2%, 5% und 10% fuer n<=30 - - indizes (zweiseitig!), gilt erst ab n >= 5! (Aus SACHS, S. 230) + critical roh-Values for significance levels 1%, 2%, 5% and 10% for n <=30 + indexes (bilateral!), is valid only for n >= 5! (in SACHS, p. 230) @@ -1883,7 +1885,7 @@ - - t-Test auf Differenz zweier Mittelwerte aus zwei Stichproben + t-Test of difference between two median values from two data sets @@ -1940,8 +1942,7 @@ - - t-Test auf Differenz zweier Mittelwerte aus zwei paarweise erhobenen Stichproben + t-Test of difference between two median values from two paired data sets ==================================================================== statist.h @@ -64,23 +64,22 @@ - - #define MPOLY 20 /* maximaler Grad des Polynoms + #define MPOLY 20 /* Maximum degree of polynomials - -#define MLINE 255 /* Max. Zeichen fuer Standard-Zeilenvariable +#define MLINE 255 /* Maximum number of chars for standard line variable -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFCxlOMg2jPD2DL3GURAvzNAJ4gOsLYF3cHgewVXZNuP3OYmFpEfQCdEl3H /+3sAfnDMGm/MxnFg3/ftwY= =hN8G -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Mon Jul 4 11:25:49 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 4 Jul 2005 11:25:49 +0200 Subject: translating German comments In-Reply-To: <200507020542.55840.jaksonaquino@yahoo.com.br> References: <200507020542.55840.jaksonaquino@yahoo.com.br> Message-ID: <20050704092549.GK21670@intevation.de> Hi Jakson, not reading the code around it, but the comments, here are some remarks. Note that the optimal result would be that the code is understood before the comment is translated as the comment might as well be not 100 acurate. ;-) On Sat, Jul 02, 2005 at 05:42:41AM -0300, Jakson Aquino wrote: > menue.c > @@ -55,12 +55,13 @@ > - Prueft, ob Spalten unterschiedliche Laenge besitzen > + Test whether columns have different lengths Literally translated it would be "Tests", but I do not think this matters. ok. > ==================================================================== > > procs.c > @@ -45,8 +45,8 @@ > - berechnet partielle korrelation; maximal 3 feste > korrelationen > + calculate partial correlation; at most 3 constant > correlations ok. > @@ -777,7 +779,7 @@ > - Test auf gleiche Effektwahrscheinlichkeiten bzw. > gleiche Dosen > + Test on same effect-probability as well as same dosis ok. > > @@ -1612,7 +1614,7 @@ > - kritische Werte fuer 95% Konfidenzinterval des > Medians aus NEAVE: > + critical values for 95% confidence interval of the > Medians (see NEAVE): ok. > @@ -1784,8 +1786,8 @@ > - kritische roh-Werte fuer signifikanzniveaus 1%, 2%, > 5% und 10% fuer n<=30 > - indizes (zweiseitig!), gilt erst ab n >= 5! (Aus > SACHS, S. 230) > + critical roh-Values for significance levels 1%, 2%, > 5% and 10% for n <=30 "roh" means "raw" > + indexes (bilateral!), is valid only for n >= 5! (in > SACHS, p. 230) ok. > > @@ -1883,7 +1885,7 @@ > - t-Test auf Differenz zweier Mittelwerte aus zwei > Stichproben > + t-Test of difference between two median values from > two data sets I believe that median != average, and "Mittelwert" mean "average". "Mittelwerte" is plural. "Stichproben" means "samples", though "data set" is not wrong of course. > @@ -1940,8 +1942,7 @@ > - t-Test auf Differenz zweier Mittelwerte aus zwei > paarweise erhobenen > Stichproben > + t-Test of difference between two median values from > two paired data sets See remarks to the last translation. > ==================================================================== > > statist.h > @@ -64,23 +64,22 @@ > - #define MPOLY 20 /* maximaler Grad des Polynoms > + #define MPOLY 20 /* Maximum degree of polynomials "des polynoms" is singular possesive, but I think this does not matter. > -#define MLINE 255 /* Max. Zeichen fuer > Standard-Zeilenvariable > +#define MLINE 255 /* Maximum number of chars for > standard line variable ok. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050704/3274e21d/attachment.bin From jaksonaquino at yahoo.com.br Mon Jul 4 15:57:45 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 4 Jul 2005 10:57:45 -0300 Subject: translating German comments In-Reply-To: <20050704092549.GK21670@intevation.de> References: <200507020542.55840.jaksonaquino@yahoo.com.br> <20050704092549.GK21670@intevation.de> Message-ID: <200507041057.57380.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello All! Bernhard: thanks for the corrections! Michael Gebhardt also revised the translations and he too suggested that "data set" was replaced with "sample". Should I make the replacement also in the English messages that are shown to users? Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFCyUBig2jPD2DL3GURArbiAJ0Qi+sPB1IZPbta47x5Z19Hfvf7lwCgqhgG cHx2tkL5tMKq9oGSI1REcK0= =ea5Q -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Mon Jul 4 18:22:26 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 4 Jul 2005 18:22:26 +0200 Subject: translating German comments In-Reply-To: <200507041057.57380.jaksonaquino@yahoo.com.br> References: <200507020542.55840.jaksonaquino@yahoo.com.br> <20050704092549.GK21670@intevation.de> <200507041057.57380.jaksonaquino@yahoo.com.br> Message-ID: <20050704162226.GA2343@intevation.de> Hi Jakson, Am 4. Jul 2005 um 10:57:45 schrieb Jakson Aquino: > Bernhard: thanks for the corrections! > > Michael Gebhardt also revised the translations and he > too suggested that "data set" was replaced with > "sample". Should I make the replacement also in the > English messages that are shown to users? Yes, why not, if you also feel that it is appropriate. :) Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050704/1722eae4/attachment.bin From jaksonaquino at yahoo.com.br Wed Jul 13 05:23:25 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Wed, 13 Jul 2005 00:23:25 -0300 Subject: news in the cvs Message-ID: <200507130023.27317.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello All, Bernhard give me write access to cvs, and I made some changes in the Statist sources and documentation. The po files were also updated. The changes are registered in the file CHANGES, and, the majority of them already were discussed in this list. In the new code that I wrote, the messages that should be in German are in English. Regarding the documentation, I replaced the boxplot-en.png with boxplot-en.fig because it's easier to translate to other languages (I'll translate the manual into Portuguese). But now, pdflatex isn't including the picture in the final pdf file. Thus, I used dvipdfm, which is part of teTex package. I put the pdf files into the cvs because I think that some users might not have teTex and xfig installed, and these packages are necessary to compile the pdf files. I also added a Makefile to compile the pdf files. Comments, suggestions, bug reports etc... are welcome! Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC1Iktg2jPD2DL3GURAsZrAJ4yGcE773lPQX21RNXkkUHgCIhdmwCggwxD W9yEN6uU54X9iTwh9dFz2AY= =OQgQ -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Wed Jul 13 12:11:31 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 13 Jul 2005 12:11:31 +0200 Subject: news in the cvs In-Reply-To: <200507130023.27317.jaksonaquino@yahoo.com.br> References: <200507130023.27317.jaksonaquino@yahoo.com.br> Message-ID: <20050713101131.GA16049@intevation.de> Hi Jaskson, Am 13. Jul 2005 um 00:23:25 schrieb Jakson Aquino: > Bernhard give me write access to cvs, and I made > some changes in the Statist sources and > documentation. The po files were also updated. very nice. > In the new code that I wrote, the messages that > should be in German are in English. Can you post an overview, then maybe sombody can help with the translation. > Regarding the documentation, I replaced the > boxplot-en.png with boxplot-en.fig Personally I prefer Skencil (www.skencil.org) over xfig these days. Skencil can read .fig files. > because it's > easier to translate to other languages (I'll > translate the manual into Portuguese). But now, > pdflatex isn't including the picture in the > final pdf file. As far as I know you need to have an .eps and then use epstopdf to make a figure go best with pdflatex. Or use .png for images. > I put the pdf files into the cvs because I think > that some users might not have teTex and xfig > installed, and these packages are necessary to > compile the pdf files. I also added a Makefile > to compile the pdf files. Usually I refrain from putting non-sources into the CVS, as they sometimes get out of sync. Especially if developers do not have the necessary tools. (Probalby only transfig.) This is different from a tarball where a pdf should be included. We could also directly offer a pdf manual for download somewhere. Thanks for the great work you are putting into statist! Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050713/55f90ba1/attachment.bin From jaksonaquino at yahoo.com.br Fri Jul 15 06:20:13 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Fri, 15 Jul 2005 01:20:13 -0300 Subject: news in the cvs In-Reply-To: <20050713101131.GA16049@intevation.de> References: <200507130023.27317.jaksonaquino@yahoo.com.br> <20050713101131.GA16049@intevation.de> Message-ID: <200507150120.16009.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, > Can you post an overview, then maybe sombody can > help with the translation. Basically, new messages were added at the end of data.c and in the file recode.c. They are related with the new features: extract columns from fixed width data file, extract sample of rows and recode. I didn't change the original messages. But I also added to gettext two or three strings that were only in German (because they were the same in both languages), and in two places I replaced "data set" with "sample". The translator have only to seek for the word "fuzzy" in the po file and also look for untranslated messages. A tool like kbabel can be helpfull because it can show only unstranlated and fuzzy message, and it allows the user to do a spell check in the translated messages only. Emacs has a po mode, but I don't know what its feature are. But perhaps we should plan a new release of Statist before translanting the new strings. The translations should be the last step before the release. I already made the Portuguese translation of the strings, but I'm aware that the English strings still might be changed. > > Regarding the documentation, I replaced the > > boxplot-en.png with boxplot-en.fig > > Personally I prefer Skencil (www.skencil.org) over > xfig these days. Skencil can read .fig files. I didn't know Skencil. I'll try it. > As far as I know you need to have an .eps and then > use epstopdf to make a figure go best with pdflatex. > Or use .png for images. I didn't have the idea of using epstopdf. I'll test it. The nice thing about dvipdfm is that it creates hyperlinks in the pdf document, and after the creation of the dvi file with latex, it's very fast the conversion into pdf. > Usually I refrain from putting non-sources into the > CVS, as they sometimes get out of sync. > Especially if developers do not have the necessary > tools. (Probalby only transfig.) > This is different from a tarball where a pdf should > be included. We could also directly offer a pdf > manual for download somewhere. We can remove the pdfs, but where will we maintain the pdf files for download? Perhaps we should maintain the pdf files in the cvs only untill the release of the next tarball, and, then, remove them from the cvs... Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC1zl9g2jPD2DL3GURAr9jAJ48Jl3N5sIa5fYHqgjok2tFXm92cQCfSNGc hpPpinXWdbmuvUgs8fc87rw= =V5Gw -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Fri Jul 15 10:06:10 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 15 Jul 2005 10:06:10 +0200 Subject: news in the cvs In-Reply-To: <200507150120.16009.jaksonaquino@yahoo.com.br> References: <200507130023.27317.jaksonaquino@yahoo.com.br> <20050713101131.GA16049@intevation.de> <200507150120.16009.jaksonaquino@yahoo.com.br> Message-ID: <20050715080610.GE9273@intevation.de> Hi Jakson, Am 15. Jul 2005 um 01:20:13 schrieb Jakson Aquino: > But perhaps we should > plan a new release of Statist before translanting the > new strings. The translations should be the last step > before the release. yes, this is a good strategy. As for .po tools, there is also www.poedit.org. > I didn't have the idea of using epstopdf. I'll test it. > The nice thing about dvipdfm is that it creates > hyperlinks in the pdf document, and after the creation > of the dvi file with latex, it's very fast the > conversion into pdf. I think pdflatex will also create hyperlinks, if you used the right LaTeX package. > We can remove the pdfs, but where will we maintain the > pdf files for download? As mentioned before, I will offer a new infrastructure for statist soon that will have place for stuff to put online. Currently I can put it in the same place as statist releases. > Perhaps we should maintain the > pdf files in the cvs only untill the release of the > next tarball, and, then, remove them from the cvs... I'd rather not have them in CVS. Shall I place the pdf somewhere for downloading? Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050715/8d7537f2/attachment.bin From jaksonaquino at yahoo.com.br Fri Jul 15 17:02:45 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Fri, 15 Jul 2005 12:02:45 -0300 Subject: news in the cvs In-Reply-To: <20050715080610.GE9273@intevation.de> References: <200507130023.27317.jaksonaquino@yahoo.com.br> <200507150120.16009.jaksonaquino@yahoo.com.br> <20050715080610.GE9273@intevation.de> Message-ID: <200507151202.47866.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Sex 15 Jul 2005 05:06, Bernhard Reiter escreveu: > I think pdflatex will also create hyperlinks, > if you used the right LaTeX package. I have to solve some dependency problems to install Skencil, before trying to re-writte the doc/Makefile. But, if you prefer, you can change the Makefile and the manual-en.tex as necessary now. > I'd rather not have them in CVS. > Shall I place the pdf somewhere for downloading? Sure! You can remove the pdf files from the CVS and put them in another place. The new infrastructure for Statist is very promising! Will it have space for forums in laguages other than English? Some users might want to contribute comments, bug reports, or usage tips, but not be confortable writting in English. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC19AVg2jPD2DL3GURAudrAJ9yVwXWO92HBR3/NAZSxp1lDc8EyACgjNgT +F6Nif5ADVmIXt4Po1caqvI= =Nl85 -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Mon Jul 18 15:12:09 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 18 Jul 2005 15:12:09 +0200 Subject: news in the cvs In-Reply-To: <200507151202.47866.jaksonaquino@yahoo.com.br> References: <200507130023.27317.jaksonaquino@yahoo.com.br> <200507150120.16009.jaksonaquino@yahoo.com.br> <20050715080610.GE9273@intevation.de> <200507151202.47866.jaksonaquino@yahoo.com.br> Message-ID: <20050718131209.GC28053@intevation.de> Hi Jakson, Am 15. Jul 2005 um 12:02:45 schrieb Jakson Aquino: > Em Sex 15 Jul 2005 05:06, Bernhard Reiter escreveu: > I have to solve some dependency problems to install > Skencil, before trying to re-writte the doc/Makefile. > But, if you prefer, you can change the Makefile and > the manual-en.tex as necessary now. No problem, take your time. If you have Skencil problems, ask on their list, I am also reading there from time to time. > > I'd rather not have them in CVS. > > Shall I place the pdf somewhere for downloading? > > Sure! You can remove the pdf files from the CVS and put > them in another place. Will do so, but might take me a few days. > The new infrastructure for > Statist is very promising! Will it have space for > forums in laguages other than English? Some users > might want to contribute comments, bug reports, or > usage tips, but not be confortable writting in > English. It will be a regular Gforge, I'll have to check if there is space for other language, but probably we can arrange something. We just need a responsible person for each language that we do not speak. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050718/7b5b0665/attachment.bin From jaksonaquino at yahoo.com.br Fri Jul 22 12:09:29 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Fri, 22 Jul 2005 07:09:29 -0300 Subject: release schedule Message-ID: <200507220709.32995.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello All, I'm presenting here a proposal of release schedule. I added new features to Statist, and the new version could be 1.1.0. However, I plan to add some new features on next months, and, I think we could release the next version as 1.0.2, waiting for the new features to release the (still unstable) version 1.1.0. I propose the following schedule: 01 Aug - Deadline for tests, bug fixes, string changes, and documentation writting. Strings are ready for translation. 08 Aug - Translation deadline. 15 Aug - Release of Statist 1.0.2. Because some users might not have all the necessary tools, in addition to what we have in the cvs, I think that the Statist tarball should have the .gmo files in the po directory and the pdf files in the doc directory. I plan to translate the Statist manual into Portuguese, and I'll do that after 01 Aug. Is the proposed schedule OK? Did I miss any step or are the dates inadequate? Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFC4MXag2jPD2DL3GURAqK5AKC4C4nBiyZTZCMXwd+m9ZGg4HqY5ACeJvs3 N5IR9athPZ8YFGdxd4jD7cM= =wsCk -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Fri Jul 22 18:31:09 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 22 Jul 2005 18:31:09 +0200 Subject: release schedule In-Reply-To: <200507220709.32995.jaksonaquino@yahoo.com.br> References: <200507220709.32995.jaksonaquino@yahoo.com.br> Message-ID: <20050722163109.GF23164@intevation.de> Hi Jakson, Am 22. Jul 2005 um 07:09:29 schrieb Jakson Aquino: > I'm presenting here a proposal of release schedule. > > I added new features to Statist, and the new version > could be 1.1.0. I wonder if statist should adopt a licensing schema like Linux, with 1.odd. be development and 1.even. be stable releases. > However, I plan to add some new > features on next months, and, I think we could release > the next version as 1.0.2, waiting for the new > features to release the (still unstable) version > 1.1.0. To have internationalisation in place would warrant a 1.2.0 release, too as this is a bigger step. > I propose the following schedule: > > 01 Aug - Deadline for tests, bug fixes, string changes, > and documentation writting. Strings are > ready for translation. > 08 Aug - Translation deadline. > 15 Aug - Release of Statist 1.0.2. Fine with me, I will not have much time to help, so how could I be against the schedule? :) > Because some users might not have all the necessary > tools, in addition to what we have in the cvs, I think > that the Statist tarball should have the .gmo files in > the po directory and the pdf files in the doc directory. Sounds good. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050722/45349c54/attachment.bin From jaksonaquino at yahoo.com.br Sat Jul 23 16:39:27 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sat, 23 Jul 2005 11:39:27 -0300 Subject: release schedule In-Reply-To: <20050722163109.GF23164@intevation.de> References: <200507220709.32995.jaksonaquino@yahoo.com.br> <20050722163109.GF23164@intevation.de> Message-ID: <200507231139.30281.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Sex 22 Jul 2005 13:31, Bernhard Reiter escreveu: > I wonder if statist should adopt a licensing schema > like Linux, with 1.odd. be development and 1.even. > be stable releases. > > To have internationalisation in place > would warrant a 1.2.0 release, too > as this is a bigger step. What version number do you think is more adequate, 1.1.0 or 1.2.0? I tested many items of Statist menu, and they don't seem to have problems. However I don't know what kind of data would be necessary to run some analyses (like probit), and thus, I can't guarantee that the graphics are really working with gettext in these functions. On the other hand, I believe that we have no bugs in the statistical results, and this is the most important thing to consider Statist stable. I'll be happy with any version number :-) Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFC4lagg2jPD2DL3GURAqK5AKCRsTHTZpDJRtcJ7jhWA1JHmbvlQwCeOofO eKl2eaoqh+dkJ9qBY9ALpMA= =iGvQ -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Mon Jul 25 12:29:12 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 25 Jul 2005 12:29:12 +0200 Subject: release schedule In-Reply-To: <200507231139.30281.jaksonaquino@yahoo.com.br> References: <200507220709.32995.jaksonaquino@yahoo.com.br> <20050722163109.GF23164@intevation.de> <200507231139.30281.jaksonaquino@yahoo.com.br> Message-ID: <20050725102912.GC16999@intevation.de> Am 23. Jul 2005 um 11:39:27 schrieb Jakson Aquino: > Em Sex 22 Jul 2005 13:31, Bernhard Reiter escreveu: > > I wonder if statist should adopt a licensing schema > > like Linux, with 1.odd. be development and 1.even. > > be stable releases. > > > > To have internationalisation in place > > would warrant a 1.2.0 release, too > > as this is a bigger step. > > What version number do you think is more adequate, > 1.1.0 or 1.2.0? Hi Jakson, 1.2.0 for the step to internationalisation and 1.3.x for your development items. > I tested many items of Statist menu, and they > don't seem to have problems. However I don't know > what kind of data would be necessary to run > some analyses (like probit), and thus, I can't > guarantee that the graphics are really working with > gettext in these functions. On the other hand, I > believe that we have no bugs in the statistical > results, and this is the most important thing to > consider Statist stable. I'll be happy with any > version number :-) I am merely voicing ideas here. You are in the driving seat. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050725/6a974ca3/attachment.bin From jaksonaquino at yahoo.com.br Mon Aug 1 03:35:45 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sun, 31 Jul 2005 22:35:45 -0300 Subject: translations Message-ID: <200507312235.48325.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello All! I'll not do any change in the Statist strings during the next weeks. So, the po files are ready for translation. I have just updated the po directory. Now, the strings are sorted by source file. There are less than 10 fuzzy messages and less than 30 new messages in the po files. Poedit, and Kbabel are helpful applications. Do we have volunteer translators? The past translations (Spanish and Italian) were made respectively by Carleos Artime and Nicola Intini. Could you translate the new strings? The new messages have to be translated into German in two places: the po file and the source code. Currently, they are in English in the source code. Who will be our German volunteer? Bernhard? Regarding the version number, since 1.1.0 is too low a number, I think we can jump to 1.3.0. I prefer to use an even number only when more users have used the current version of Statist for some time. Today I compiled Statist under DOS/Windows with gettext support using MinGW (included in Dev-C++) and CygWin. With CygWin Statist works like in Linux, with no limitations. With MinGW the gnuplot has to be opened separately, but it works fine with Statist. After plotting a new graphic we have just to open stat_gpl.com again. One problem with both compilations was that the accents aren't being correctly printed in the screen. The accents are correctly printed only in the graphics generated by Statist compiled in with Dev-C++/MinGW. During the next week, I'll make the necessary changes in the src/Makefile and the README file to make it easier for people to compile Statist under DOS/Windows. Best regards, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC7Xxxg2jPD2DL3GURAthzAJwLFGAN1nLVx5nmj9hBVKk1hWWCMgCdErPt D4ssunM1wduG59l/a8T43bE= =xeHA -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Mon Aug 1 14:45:38 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 1 Aug 2005 14:45:38 +0200 Subject: translations In-Reply-To: <200507312235.48325.jaksonaquino@yahoo.com.br> References: <200507312235.48325.jaksonaquino@yahoo.com.br> Message-ID: <20050801124538.GD23267@intevation.de> Hi Jakson, Am 31. Jul 2005 um 22:35:45 schrieb Jakson Aquino: > I'll not do any change in the Statist strings > during the next weeks. So, the po files are > ready for translation. nice! > Do we have volunteer translators? > > The past translations (Spanish and Italian) were > made respectively by Carleos Artime and Nicola > Intini. Could you translate the new strings? > > The new messages have to be translated into > German in two places: the po file and the source > code. Currently, they are in English in the > source code. Who will be our German volunteer? > Bernhard? Currently I cannot promise anything. So I would prefer if somebody else would step in. Doing the translations is fairly easy. > Regarding the version number, since 1.1.0 is too > low a number, I think we can jump to 1.3.0. Fine with me. > During the next week, I'll > make the necessary changes in the src/Makefile > and the README file to make it easier for people > to compile Statist under DOS/Windows. Good. One idea of not making gettext a necessity were very old systems (286) or embedded systems with smaller c-libraries, like dietlibc that might not have gettext. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050801/0a5fa22f/attachment.bin From jaksonaquino at yahoo.com.br Thu Aug 4 02:59:09 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Wed, 3 Aug 2005 21:59:09 -0300 Subject: Portugues translations Message-ID: <200508032159.11315.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello all! Many undergraduate Brazilian students don't know English very well. So, I'll make a translation of the English part of the README file. As usual in source code written in Portuguese, the file will be named LEIAME. Its first line will have an English message saying that it's a translation of README. I think the file can stay at the same directory of README or would it better to put it together with the contributions for Statist? One problem with translations is that they can become outdated. If this happens, we can move it to the contributions place or delete it. The manual (which is rather a tutorial) is already translated into Portuguese. I'll revise the translation on the next days and put it on the cvs. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFC8Whdg2jPD2DL3GURArFIAJ4wo/pXgwJn+B0vfF7NXx/hsFG7xACdGQbW w0lOgj0Z3p8MC93SjD5B9TU= =b4AT -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Thu Aug 4 03:14:07 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Wed, 3 Aug 2005 22:14:07 -0300 Subject: translations In-Reply-To: <20050801124538.GD23267@intevation.de> References: <200507312235.48325.jaksonaquino@yahoo.com.br> <20050801124538.GD23267@intevation.de> Message-ID: <200508032214.09136.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello! Gettext support is working on DOS with ibintl3.dll which is one of the GnuWin32 packages. The compilation was done with MinGW gcc that comes with Dev-C++. On Windows I couldn't set the LOCALEDIR in the Makefile. The LOCALEDIR is always defined as integer, and not as string as happens on Linux. The solution that I found was to use a fix place for catalogs: %WINDIR%/locale. However, I'm not sure that the environment variable %WINDIR% is defined in all versions of Windows. I didn't find on the Internet an "official" solution to the problem of accents on DOS. However, the accents are correctly displayed (on Windows 98) if we use iconv to convert the charset encoding of the po files to cp850 but keep the information "CHARSET=iso8859-1" in the po file header. This works for the messages displayed in the DOS window. The 12 messages sent to gnuplot have to be converted back to iso-8859-1 in the po files. I couldn't solve the accents problem in CygWin. A letter ?, for example, is being displayed as "a. I have two sed scripts to strip the accents from letters. One for German (?, for example, becomes ue) and other for the other languages (? becomes u). I can put the scripts on the po directory. Or would it be better to add them to another place, like a web page with contributions for Statist? > One idea of not making gettext a necessity were > very old systems (286) or embedded systems with smaller c-libraries, > like dietlibc that might not have gettext. I used a pre-compiled binary of libintl, but certainly it would be possible to compile the library in small C systems. Indeed, libintl is small. However, libiconv isn't and perhaps using libintl as it is currently implemented would turn small systems into big ones. But I believe that it would not be difficult to create a special version of libintl without using libiconv, and using only the 26 English letters, like the Statist language macros (but I don't intend to try creating this libintl). With Western European languages it would work reasonably well and the systems would stay small. It's very easy to strip. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFC8Wvfg2jPD2DL3GURAsV6AJ9qK/72gbtVd74ikSu1wzlRjmxdBgCgghk7 Jr2ksfo7IdLk79m0uM+Ws0Q= =n7hi -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Thu Aug 4 10:11:18 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Thu, 4 Aug 2005 05:11:18 -0300 Subject: Gnuplot on Windows Message-ID: <200508040511.29667.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello! Now it's possible to use pipe with gnuplot on Windows through the program pgnuplot.exe which send signals to gnuplot.exe. However, I don't know how to detect when the pipe is broken. I put a warning on the manual about that. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC8c2ug2jPD2DL3GURAvnQAKDHErjzXH88bkXk30WE1X1XYJ/tNgCfS8ne jdyt/RMAD49ylBTlVjwTPKY= =O1tm -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Sun Aug 7 11:57:42 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sun, 7 Aug 2005 11:57:42 +0200 Subject: translations In-Reply-To: <200508032214.09136.jaksonaquino@yahoo.com.br> References: <200507312235.48325.jaksonaquino@yahoo.com.br> <20050801124538.GD23267@intevation.de> <200508032214.09136.jaksonaquino@yahoo.com.br> Message-ID: <20050807095742.GB21900@intevation.de> Hi Jakson, Am 3. Aug 2005 um 22:14:07 schrieb Jakson Aquino: > I couldn't solve the accents problem in CygWin. A letter > ?, for example, is being displayed as "a. I have two sed > scripts to strip the accents from letters. One for German > (?, for example, becomes ue) and other for the other > languages (? becomes u). I can put the scripts on the po > directory. Or would it be better to add them to another > place, like a web page with contributions for Statist? I would put them in CVS. Maybe in a different place, but not in "contrib" as they are done by yourself which makes them official. ;) > > One idea of not making gettext a necessity were > > very old systems (286) or embedded systems with smaller c-libraries, > > like dietlibc that might not have gettext. > > I used a pre-compiled binary of libintl, but certainly it > would be possible to compile the library in small C > systems. Indeed, libintl is small. However, libiconv isn't > and perhaps using libintl as it is currently implemented > would turn small systems into big ones. But I believe that > it would not be difficult to create a special version of > libintl without using libiconv, and using only the 26 > English letters, like the Statist language macros (but I > don't intend to try creating this libintl). With > Western European languages it would work reasonably well > and the systems would stay small. It's very easy to strip. This is just a general idea of mine, I have never though much more about the requirements of smaller systems except that I was very happy years ago, that statist ran on 286 systems. We did not have the gettext then, so I do not know if this is actually a concern for real users. So do not go to length to make this possible, unless you aim at a special target group. Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050807/abfe66f4/attachment.bin From bernhard at intevation.de Sun Aug 7 11:59:00 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sun, 7 Aug 2005 11:59:00 +0200 Subject: Portugues translations In-Reply-To: <200508032159.11315.jaksonaquino@yahoo.com.br> References: <200508032159.11315.jaksonaquino@yahoo.com.br> Message-ID: <20050807095900.GC21900@intevation.de> Hi Jakson, Am 3. Aug 2005 um 21:59:09 schrieb Jakson Aquino: > Many undergraduate Brazilian students don't know English > very well. So, I'll make a translation of the English part > of the README file. As usual in source code written in > Portuguese, the file will be named LEIAME. Its first line > will have an English message saying that it's a translation > of README. I think the file can stay at the same directory of > README or would it better to put it together with the > contributions for Statist? let is stay in the directory. > One problem with translations is > that they can become outdated. If this happens, we can move it > to the contributions place or delete it. Make the note in the translation mention that this is a translation of README revision x.y and that the original might be newer. > The manual (which is rather a tutorial) is already > translated into Portuguese. I'll revise the translation on > the next days and put it on the cvs. Cool. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050807/b6fc564e/attachment.bin From jaksonaquino at yahoo.com.br Mon Aug 8 19:11:27 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 8 Aug 2005 14:11:27 -0300 Subject: German, Italian and Spanish Message-ID: <200508081411.29810.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hello! Still missing are the translations of some German, Italian, and Spanish messages. All other files, including the documentation (English and Portuguese) are ready for the release of version 1.3.0. The release of Statist was scheduled for today, but it was postponed by a few days while we finish the update of the web page and the preparation of the tarballs. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC95I/g2jPD2DL3GURAgjrAJ97opbozqIi/8HRGaPRHB3runPBKQCfbW54 DjGmKXTt+pm4AUi7zeSmN7k= =N7M1 -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Mon Aug 8 19:15:15 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 8 Aug 2005 14:15:15 -0300 Subject: Release announcement Message-ID: <200508081415.17862.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 STATIST - 1.3.0 Date: 08 Aug 2005 =============== Statist 1.3.0. will be released in a few days. Statist is a small and portable statistics program written in C. It is terminal-based, but can utilise Gnuplot for plotting purposes. It can be run in scripts, and big datasets are handled reasonably well on small machines. It is also easy to use due to its interactive menu, being a very handy tool also for those who do not use statistics every day. THE DOCUMENTATION ================= The statist user manual is available in both English and Portuguese. The German documentation is partially outdated, but it contains information about the statistical functions and the way the program works that are not covered by the new user manual. REQUIREMENTS ============ Statist is very portable. It can be compiled in pratically any system with any C compiler. However, Statist needs Gnuplot to create graphics, and the system must have libintl and libiconv to compile the program with support for languages other than English and German. Some GNU core utils -- gawk and sort -- are useful programs to prepare databases for use with Statist. DISTRIBUTION ============ Statist source code and pre-compiled binaries can be found at: http://www.usf.uos.de/~breiter/tools/statist/index.en.html NEW FEATURES IN VERSION 1.3.0 ============================= - Gettext support, with message catalogs for German, Italian, Portuguese, and Spanish. - English and Portuguese documentation. - The use of the option -delrow was replaced with a new system. Now, the rows are deleted during analyzes (when necessary), and we are using the smallest double number as missing value. It was necessary a lot of data manipulation to do some analyzes with the old system. - Added new features: (1) extract columns from fixed width data file; (2) extract a sample of rows from a file; and (3) recode a Statist data file. - Small adaptions in the code to keep it compiling on DOS/Windows. Gettext support is possible if compiled with MinGW or CygWin. The use of pipes is possible with pgnuplot, which is part of new versions of gnuplot. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC95Mjg2jPD2DL3GURAnOvAJsHlgJ/gMAqhlA8Z9ZxQhwkg2lF5ACgx1bl ecvCP22dcR/pk+PdWMcRGHo= =WWiD -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Thu Aug 11 16:43:55 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Thu, 11 Aug 2005 16:43:55 +0200 Subject: German translation (was: translations) In-Reply-To: <20050801124538.GD23267@intevation.de> References: <200507312235.48325.jaksonaquino@yahoo.com.br> <20050801124538.GD23267@intevation.de> Message-ID: <20050811144355.GB29291@intevation.de> Am 1. Aug 2005 um 14:45:38 schrieb Bernhard Reiter: > Am 31. Jul 2005 um 22:35:45 schrieb Jakson Aquino: > > The new messages have to be translated into > > German in two places: the po file and the source > > code. Currently, they are in English in the > > source code. Who will be our German volunteer? > > Bernhard? > Currently I cannot promise anything. > So I would prefer if somebody else would step in. > Doing the translations is fairly easy. I had some time in a train and did a first translation in the .po only and checked it into CVS. I hope it is helpful to some regard. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050811/4644ac69/attachment.bin From bernhard at intevation.de Thu Aug 11 16:47:01 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Thu, 11 Aug 2005 16:47:01 +0200 Subject: recode implementation (was: recode, get sample, and get columns) In-Reply-To: <20050624105124.GA1933@localhost.localdomain> References: <20050624105124.GA1933@localhost.localdomain> Message-ID: <20050811144701.GC29291@intevation.de> Hi Jakson, Am 24. Jun 2005 um 07:51:24 schrieb Jakson A. Aquino: > I developed some tools for Statist while trying to > use the PNAD database. first: Thanks again for these nice efforts. > one to recode a Statist > data file (recode.c). I also wrote the first draft > documentation for the three programs. > recode.c is bigger and more > complex, and, probably, it would be better to keep > it as a separate file. I had a peek at the documentation and have two comments: As you can see from --bernhard, I tried to recreate new long options with the convention that they should start with a "--" and not just a "-". For the new options we might as well use "--" as their is no need for backward compability. I ask myself if using "awk" or python for this kind of reformatting would be better as it is more general and not much more verbose. Or the other way around: What is the advantage of using Statist for this? Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050811/2787f6e4/attachment.bin From bernhard at intevation.de Thu Aug 11 16:51:34 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Thu, 11 Aug 2005 16:51:34 +0200 Subject: CVS: fixed tests; minor manual editing Message-ID: <20050811145134.GD29291@intevation.de> Hi Jakson, yesterday I was in the train for a few hours and had a recent cvs checkout with me, but not the proposed new announcement nor the webpage. Here is the changelog of what I have checked into the CVS just now: + * Fixed the unit tests. + + tests/test_4_1.sh: fixed test: expecting new comment in the output. + + tests/check_memory_handling.c: fixed test for "check" v>=0.9.0. + + * Minor editing of the English manual. + + Minor clean of doc/Makefile: + - Ignoring errors of rm. + - Declared phony targets .PHONY. Also rewrote "all" target. + + Called the operating system GNU/Linux as Linux is only a kernel. + + Made the message recommending Free Software friendlier and + more correct about the security implications. + + added the gnuplot method to create output files. + + updated the build and install instructions. "su -" in general + is prefered over "su". Did you run the automatic tests during development? Just say "make check" in src/ to run them. Automatic tests are nice for development cycle to do test a little -- code a little which is a method also used by extreme programming. When compiling I also found a few interesting warning by the compiler: Comparision ist always true due to limited range of data type. in data.c:1057 recode.c:378 recode.c:466 This makes sense, as '?' == (char)255. So you could leave out the right part of the check. This would be different if char would be unicode character, which has more than 8 bits. I have no experience with unicode in modern C development, but I expect that the comparisons will be completely different anyway. Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050811/f411ef3f/attachment.bin From jaksonaquino at yahoo.com.br Fri Aug 12 13:45:22 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Fri, 12 Aug 2005 08:45:22 -0300 Subject: CVS: fixed tests; minor manual editing In-Reply-To: <20050811145134.GD29291@intevation.de> References: <20050811145134.GD29291@intevation.de> Message-ID: <200508120845.24400.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, I reproduced your changes to the manual into the Portuguese version. in both versions, I also added one line acknowledging your contribution. Thanks! I didn't updated the cvs version of the pdf files, since they will be deleted after the next release of statist. Em Qui 11 Ago 2005 11:51, Bernhard Reiter escreveu: > Did you run the automatic tests during development? > Just say "make check" in src/ to run them. > > Automatic tests are nice for development cycle to do > test a little -- code a little > which is a method also used by extreme programming. In fact, I didn't run the automatic tests. I'll download the check.h file and run them. > This makes sense, as '?' == (char)255. > So you could leave out the right part of the > check. This would be different if char would be > unicode character, which has more than 8 bits. I > have no experience with unicode in modern C > development, but I expect that the comparisons will > be completely different anyway. Yes, your compiler is right, but It curious that I didn't received this warning here. You pointed to an important problem. The test might not work with utf-8. I'll investigate it. I prefer to be allowed to use letters with accents in the column labels. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC/IvSg2jPD2DL3GURAkX1AKCzV1wSS+jcpU63xlTu5/H0paLzdACfUvFK jAy8U0Ibfi82H0iosVpRfc4= =ocAO -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Fri Aug 12 13:31:44 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Fri, 12 Aug 2005 08:31:44 -0300 Subject: recode implementation (was: recode, get sample, and get columns) In-Reply-To: <20050811144701.GC29291@intevation.de> References: <20050624105124.GA1933@localhost.localdomain> <20050811144701.GC29291@intevation.de> Message-ID: <200508120831.54134.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Qui 11 Ago 2005 11:47, Bernhard Reiter escreveu: > As you can see from --bernhard, I tried to recreate > new long options with the convention that they > should start with a "--" and not just a "-". For the > new options we might as well use "--" as their is no > need for backward compability. I agree! Since the first time I saw the statist help text that I perceived the existence of two styles of options. I don't know why I chosen the old style for the new options. Perhaps because the majority of options are in the old style. What a bad criterion! What about if we turn all options to the new style of using "--" for full option name and "-" for abbreviations? The old style could still be accepted by produce a warning that it is deprecated and will no longer be supported in future releases. Only the new style would be documented. > I ask myself if using "awk" or python for > this kind of reformatting would be better as it is > more general and not much more verbose. Or the other > way around: What is the advantage of using Statist > for this? You are correct! I simply don't know how to program with python, and I didn't know the existence of awk until some days ago. That's why I put some tips on the usage of awk and other gnu core utils. I discovered awk while trying to compile statist under Windows. Reading GnuWin32 download page I had interest in downloading core utils, and reading core utils documentation I found "awk" and "join". I even thought about removing recode.c from statist, but I think that the syntax of a "recode_config_file" is still simpler than the syntax of awk, at least for people who don't know how to program in C. Of course the development of recode.c wouldn't have been a priority if I already knew awk. I still didn't find a tool for merging databases. I tried "join", but the results weren't satisfactory. Perhaps I don't have enough knowledge on how to use join. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC/Iing2jPD2DL3GURAkqoAJ0Y9o2gqIVJMnpGYV7xEdQeB3QqwwCeL3eh tSoIzE6YWi9b5FD/7tpTrjI= =9Trv -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Fri Aug 12 15:10:28 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 12 Aug 2005 15:10:28 +0200 Subject: CVS: fixed tests; minor manual editing In-Reply-To: <200508120845.24400.jaksonaquino@yahoo.com.br> References: <20050811145134.GD29291@intevation.de> <200508120845.24400.jaksonaquino@yahoo.com.br> Message-ID: <20050812131028.GA32672@intevation.de> Hi Jakson, Am 12. Aug 2005 um 08:45:22 schrieb Jakson Aquino: > I reproduced your changes to the manual into the > Portuguese version. in both versions, I also added one > line acknowledging your contribution. Thanks! you are welcome. > I didn't updated the cvs version of the pdf files, > since they will be deleted after the next release of > statist. Why not delete it right away then? > Em Qui 11 Ago 2005 11:51, Bernhard Reiter escreveu: > > Automatic tests are nice for development cycle to do > > test a little -- code a little > > which is a method also used by extreme programming. > > In fact, I didn't run the automatic tests. I'll > download the check.h file and run them. Cool, you could even write some tests for your new code. > Yes, your compiler is right, but It curious that I > didn't received this warning here. I am using: gcc-Version 3.3.5 (Debian 1:3.3.5-13) > You pointed to an > important problem. The test might not > work with utf-8. I'll investigate it. I prefer to be > allowed to use letters with accents in the column labels. I think that solving the unicode and utf-8 problem will be a major task that could be deferred. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050812/2455a326/attachment.bin From bernhard at intevation.de Fri Aug 12 15:30:32 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 12 Aug 2005 15:30:32 +0200 Subject: recode implementation (was: recode, get sample, and get columns) In-Reply-To: <200508120831.54134.jaksonaquino@yahoo.com.br> References: <20050624105124.GA1933@localhost.localdomain> <20050811144701.GC29291@intevation.de> <200508120831.54134.jaksonaquino@yahoo.com.br> Message-ID: <20050812133032.GB32672@intevation.de> Hi Jakson, Am 12. Aug 2005 um 08:31:44 schrieb Jakson Aquino: > Em Qui 11 Ago 2005 11:47, Bernhard Reiter escreveu: > What about if we turn all options to the new style of > using "--" for full option name and "-" for > abbreviations? The old style could still be accepted by > produce a warning that it is deprecated and will no > longer be supported in future releases. Only the new > style would be documented. Changes in statist have been slow for a long while and there is no immedeate reasons to force all scripts to the new options format. I would say that we should in any case use the long format for the new long options. We could add "--" versions for the old long options, too and have the warning that at some unknown point in the future the "-" versions will be deprecated. > > I ask myself if using "awk" or python for > > this kind of reformatting would be better as it is > > more general and not much more verbose. Or the other > > way around: What is the advantage of using Statist > > for this? > > You are correct! I simply don't know how to program > with python, and I didn't know the existence of awk > until some days ago. I did not now that you did not know. There is a book called Brian W. Kernighan & Rob Pike, The Unix Programming Environment, 1984, Prentice-Hall. which teaches a lot about how to use small commands to assemble a more complicated operation. > I even thought about removing recode.c from statist, > but I think that the syntax of a "recode_config_file" is still simpler > than the syntax of awk, at least for people who don't > know how to program in C. The question to consider is: Does the additional effort of learning statist's syntax and maintaining the code help more then learing recepies in a standard language? If yes, you can keep the code. Otherwise removing it, will keep statist cleaner. Usually standard languages have knowledge that is useful in other circumstances as well. This is why I prefer python in many cases now, though smaller operations tend to be longer than in awk. > I still didn't find a tool for merging databases. I > tried "join", but the results weren't satisfactory. > Perhaps I don't have enough knowledge on how to use > join. Can you describe what you want to do? As far as I can see "join" will need a matching keyword in both tables. Maybe you want "paste"? $ cat num2 1 2 $ cat let3 a b c $ paste num2 let3 1 a 2 b c Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050812/419807a8/attachment.bin From jaksonaquino at yahoo.com.br Sat Aug 13 00:52:59 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Fri, 12 Aug 2005 19:52:59 -0300 Subject: recode implementation (was: recode, get sample, and get columns) In-Reply-To: <20050812133032.GB32672@intevation.de> References: <20050624105124.GA1933@localhost.localdomain> <200508120831.54134.jaksonaquino@yahoo.com.br> <20050812133032.GB32672@intevation.de> Message-ID: <200508121953.08208.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Sexta 12 Agosto 2005 10:30, Bernhard Reiter escreveu: > I would say that we should in any case use the long format > for the new long options. > We could add "--" versions for the old long options, too > and have the warning that at some unknown point in the future > the "-" versions will be deprecated. It seems that we have only to add the options inside the optrec[MOPT] (in statist.c). For "help", for example, we have: {"-h", &help}, {"-help", &help}, {"-?", &help} In this case, it's just a matter of adding one more "help": {"--help", &help}, and doing the same with the other variables. And, then, update the value of MOPT. We have only to choose the abbreviation letters. Suggestions: - -s --silent - -l --log - -p --noplot - -f --nofile - -t --thist - -B --bernhard - -x --xsample - -c --xcols - -b --nobell > I did not now that you did not know. > There is a book called > Brian W. Kernighan & Rob Pike, > The Unix Programming Environment, 1984, Prentice-Hall. Thanks for the reference. On Monday I'll look for it or for something similar (if it's not there) in the library. > > I even thought about removing recode.c from statist, > > but I think that the syntax of a "recode_config_file" is still simpler > > than the syntax of awk, at least for people who don't > > know how to program in C. > > The question to consider is: Does the additional effort of learning > statist's syntax and maintaining the code help more then > learing recepies in a standard language? > If yes, you can keep the code. > Otherwise removing it, will keep statist cleaner. > Usually standard languages have knowledge that is useful in other > circumstances as well. This is why I prefer python in many cases now, > though smaller operations tend to be longer than in awk. Your reasons are strong. I only regret that I lost my time developing recode.c and you lost yours translating its verbose messages. I agree with your two arguments: awk is useful for tasks other than manipulating statist data files, and recode.c is potentially a source of bugs. One of them you already found. I'm going to made the necessary changes in the source, and in the manual. Regarding Python, I would have to learn it. I'll try it latter, when I have the need of a big script. For now, I'm very impressed with awk speed. > Can you describe what you want to do? > As far as I can see "join" will need a matching keyword in both tables. > Maybe you want "paste"? The problem is (actually, was) that my database had three key variables. The solution was to use awk for creating a variable that is a string concatenation of the three variables. Then, the three variables became one, and I used join to merge the files. I'll add to the manual tips on how to use join, and then, everything that I once knew how to do with a slow dinosaur now I know how to do with the (much faster and free) statist and friends, sort, awk, join, etc... Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFC/ShRg2jPD2DL3GURAgDhAJoCkuMu8Fw/nJK6MSWdsLJ4NThKJwCgrM9J ypneGIPOZDsNO/xv2B2XvJ8= =4BxO -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Sun Aug 14 13:10:45 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sun, 14 Aug 2005 08:10:45 -0300 Subject: options format Message-ID: <200508140810.54947.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Sex 12 Ago 2005 10:30, Bernhard Reiter escreveu: > We could add "--" versions for the old long options, > too and have the warning that at some unknown point > in the future the "-" versions will be deprecated. I added the `--' options and a test to check whether a old long option was passed as argument. But I didn't commit the changes into the cvs yet. Is the following warning OK? And how would it be in German: The option %s will be deprecated in the future. It's better to use -%s instead. Thanks, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC/ya7g2jPD2DL3GURAolLAKDQUe+Le7N0R2Ji59zMzIyhQorsPQCeIHVr fb7r7EwJ+d44o68+KOXqDaY= =svh3 -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Sun Aug 14 14:11:45 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sun, 14 Aug 2005 09:11:45 -0300 Subject: getline and _GNU_SORCE Message-ID: <200508140911.47685.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, I'm trying to avoid the use of the option "rlen". The buffer size for reading lines will be dynamically determined by GNU extension to libc, the function getline. However, I have to add #define _GNU_SOURCE before #include . Will this cause any problem? Anyway, I will create a substitute getline() for MinGW, and CygWin which don't have GNU getline. Under Windows, statist will be able to read at most a fixed amount of bytes (50000) from a line. Under Linux, there will be no limit, and no need of the "rlen" option. The invocation of Statist becomes simpler, and the help on the usage of --xcols and --xsample becomes less verbose. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC/zUBg2jPD2DL3GURAkMTAKCszpujdmvbqwrLsa0fS49JZHpO8ACgjuLx GKJMAPDiYAyV/qSqMP/ChXg= =0JDg -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Sun Aug 14 14:48:51 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Sun, 14 Aug 2005 09:48:51 -0300 Subject: CVS: fixed tests; minor manual editing In-Reply-To: <20050812131028.GA32672@intevation.de> References: <20050811145134.GD29291@intevation.de> <200508120845.24400.jaksonaquino@yahoo.com.br> <20050812131028.GA32672@intevation.de> Message-ID: <200508140948.54031.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, I removed the pdf files from the cvs. Em Sex 12 Ago 2005 10:10, Bernhard Reiter escreveu: > Cool, you could even write some tests for your new > code. I ran the tests, and I found very nice that check unit allow us to test specific funtions. I can try to use it latter. I'm using valgrind to hunt memory leaks, uninitialized variables, buffer overflows, etc... I'll write a test for --xcols, including the recodification with awk, and merging files with join. (I'll use your "test_4_1.sh" as a guide.) I'll only finish the removal of "rlen" option when this test is prepared. > I think that solving the unicode and utf-8 problem > will be a major task that could be deferred. I confirmed that the test can't deal with utf-8 encoding, and I eliminated the accentuated letters from the "if" condition. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFC/z2zg2jPD2DL3GURAiYmAJ4sluOC6s9JaVOacAgE8o4cVcQw9QCdGbr7 6F6qR+YQUsCsiHDsUO972oM= =qP2I -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Mon Aug 15 13:13:42 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 15 Aug 2005 13:13:42 +0200 Subject: CVS: fixed tests; minor manual editing In-Reply-To: <200508140948.54031.jaksonaquino@yahoo.com.br> References: <20050811145134.GD29291@intevation.de> <200508120845.24400.jaksonaquino@yahoo.com.br> <20050812131028.GA32672@intevation.de> <200508140948.54031.jaksonaquino@yahoo.com.br> Message-ID: <20050815111342.GW27886@intevation.de> Hi Jakson, Am 14. Aug 2005 um 09:48:51 schrieb Jakson Aquino: > Em Sex 12 Ago 2005 10:10, Bernhard Reiter escreveu: > > Cool, you could even write some tests for your new > > code. > > I ran the tests, and I found very nice that check unit > allow us to test specific funtions. I can try to use > it latter. The idea is, to write the test for a funcationality or procedure first and then implement it. This way it helps designing good interfaces, because you start thinking about this first. > I'm using valgrind to hunt memory leaks, uninitialized > variables, buffer overflows, etc... I'll write a test > for --xcols, including the recodification with awk, > and merging files with join. (I'll use your > "test_4_1.sh" as a guide.) Do you call "join" or "awk" directly from statist? Usually this would be done by the user. I am unshure if we should add tests for this as statist does not do it. But it would be interesting to have example scripts for people to use. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050815/36a3f80c/attachment.bin From bernhard at intevation.de Mon Aug 15 13:19:30 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 15 Aug 2005 13:19:30 +0200 Subject: getline and _GNU_SORCE In-Reply-To: <200508140911.47685.jaksonaquino@yahoo.com.br> References: <200508140911.47685.jaksonaquino@yahoo.com.br> Message-ID: <20050815111930.GY27886@intevation.de> Am 14. Aug 2005 um 09:11:45 schrieb Jakson Aquino: > I'm trying to avoid the use of the option "rlen". Good idea, it is always best to have this dynamically. > The buffer size for reading lines will be dynamically > determined by GNU extension to libc, the function > getline. However, I have to add #define _GNU_SOURCE > before #include . Will this cause any problem? In general I would hope to conform to C-standards for a simple application like statist. > Anyway, I will create a substitute getline() for MinGW, > and CygWin which don't have GNU getline. If you do this anyway, why not use this routine by default, this would avoid the _GNU_SOURCE extensions. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050815/195adac4/attachment.bin From bernhard at intevation.de Mon Aug 15 13:23:43 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 15 Aug 2005 13:23:43 +0200 Subject: options format In-Reply-To: <200508140810.54947.jaksonaquino@yahoo.com.br> References: <200508140810.54947.jaksonaquino@yahoo.com.br> Message-ID: <20050815112343.GZ27886@intevation.de> Hi Jakson, Am 14. Aug 2005 um 08:10:45 schrieb Jakson Aquino: > Em Sex 12 Ago 2005 10:30, Bernhard Reiter escreveu: > > We could add "--" versions for the old long options, > > too and have the warning that at some unknown point > > in the future the "-" versions will be deprecated. > > I added the `--' options and a test to check > whether a old long option was passed as argument. > But I didn't commit the changes into the cvs yet. > > Is the following warning OK? And how would it be in > German: > > The option %s will be deprecated in the future. > It's better to use -%s instead. Why not The option short format -%s is deprecated and might be removed some time in the future. Use --%s instead. Deutsch: The kurze Form der Option -%s is veraltet und wird eventuell in Zukunft entfernt. Verwenden Sie stattdessen --%s. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050815/500710c1/attachment.bin From bernhard at intevation.de Mon Aug 15 13:30:39 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 15 Aug 2005 13:30:39 +0200 Subject: recode implementation (was: recode, get sample, and get columns) In-Reply-To: <200508121953.08208.jaksonaquino@yahoo.com.br> References: <20050624105124.GA1933@localhost.localdomain> <200508120831.54134.jaksonaquino@yahoo.com.br> <20050812133032.GB32672@intevation.de> <200508121953.08208.jaksonaquino@yahoo.com.br> Message-ID: <20050815113039.GB27886@intevation.de> Hi Jakson, Am 12. Aug 2005 um 19:52:59 schrieb Jakson Aquino: > Em Sexta 12 Agosto 2005 10:30, Bernhard Reiter escreveu: > It seems that we have only to add the options inside the optrec[MOPT] (in > statist.c). For "help", for example, we have: > > {"-h", &help}, {"-help", &help}, {"-?", &help} > In this case, it's just a matter of adding one more "help": {"--help", &help}, I have not checked the code, but it is probably as simple as that. > and doing the same with the other variables. And, then, update the value of > MOPT. We have only to choose the abbreviation letters. Suggestions: > > -s --silent we should not use abbreviation letters, because the unix convention is that -sh would be the same as -s -h and we still allow the -silent usage for backwards compatibility. So for a long time (even after removing the "-long" options, we would not be able to use short one, because there is the danger of old scripts. > > There is a book called > > Brian W. Kernighan & Rob Pike, > > The Unix Programming Environment, 1984, Prentice-Hall. > > Thanks for the reference. On Monday I'll look for it or for something similar > (if it's not there) in the library. It is pretty dated, but shows the design philosphy nicely. As I said I would not do some things as they describe there, but it was a nice read. :) > > The question to consider is: Does the additional effort of learning > > statist's syntax and maintaining the code help more then > > learing recepies in a standard language? > Your reasons are strong. I only regret that I lost my time > developing recode.c and you lost yours translating its verbose messages. Well, this cannot be helped. It needed the additional thoughs to come to the conclusion. > and then, everything that I once knew how to do with a slow dinosaur > now I know how to do with the (much faster and free) statist and friends, > sort, awk, join, etc... Cool. :) Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050815/0398cfb5/attachment.bin From jaksonaquino at yahoo.com.br Tue Aug 16 01:40:28 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 15 Aug 2005 20:40:28 -0300 Subject: CVS: fixed tests; minor manual editing In-Reply-To: <20050815111342.GW27886@intevation.de> References: <20050811145134.GD29291@intevation.de> <200508140948.54031.jaksonaquino@yahoo.com.br> <20050815111342.GW27886@intevation.de> Message-ID: <200508152040.30715.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Monday 15 August 2005 08:13, Bernhard Reiter escreveu: > Do you call "join" or "awk" directly from statist? > Usually this would be done by the user. No, I don't call "join" and "awk" directly from statist. > I am unshure if we should add tests for this as statist > does not do it. But it would be interesting to have example > scripts for people to use. The test included "awk" and "join" because I wanted to check whether the instructions that I wrote on the manual really work as I described. I was testing the user manual, not only statist. The fixed width data file that I created had some empty spaces where they normally are not found in real fixed width data files. This wasn't tested with the PNAD database that I used before to test --xcols. I can create a directory doc/example with some scripts and sample data files (but after the release of 1.3.0). I liked the use of tests. They increase our trust in the code that we are writting. Next time I add a feature to statist I will first write the result as it have to be, and, then, write the code. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFDASfsg2jPD2DL3GURAtImAJ4tCznfhuPU/rcAwY7lBOwmkhjDRgCgxVQf G9rPPscPpD7FZE6ScCQ60oI= =vnm9 -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Tue Aug 16 02:55:01 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 15 Aug 2005 21:55:01 -0300 Subject: binaries and documentation Message-ID: <200508152155.12647.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, I found a script in my computer that ask me some questions and create a deb package. I can, then use alien to create a rpm package, and make both packages available for download after the release. My question is, what about the documentation? We don't have info pages, and the pdf docs might be put into /usr/share/doc/statist-1.3.0/ but I fear that some user migrating from Windows would have difficult to guess where the documentation is. One idea is to add the item "Help" to the main menu. Choosing the item would print a message telling where the documentation can be found. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFDATlug2jPD2DL3GURAo/aAJwJ59tbqXkFfh1L4PfooMnH3vdAjgCg09j6 2Iq1wv0vS7twQzJhEV6CUR4= =jleh -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Grátis - Internet rápida e grátis. Instale o discador agora! http://br.acesso.yahoo.com/ From jaksonaquino at yahoo.com.br Tue Aug 16 03:00:45 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Mon, 15 Aug 2005 22:00:45 -0300 Subject: getline and _GNU_SORCE In-Reply-To: <20050815111930.GY27886@intevation.de> References: <200508140911.47685.jaksonaquino@yahoo.com.br> <20050815111930.GY27886@intevation.de> Message-ID: <200508152200.47440.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Monday 15 August 2005 08:19, Bernhard Reiter escreveu: > > Anyway, I will create a substitute getline() for MinGW, > > and CygWin which don't have GNU getline. > > If you do this anyway, why not use this routine by default, > this would avoid the _GNU_SOURCE extensions. The substitute getline function wasn't allocating memory dynamically. But now it is. I included into data.c the function getndelim2() from gnucoreutil/lib/getndelim2.c with some changes and added the following commentary above the function: /* Adapted from GNU coretutils 5.2.1 getndelim2.c */ int getline(char **lineptr, int *linesize, FILE *F){ Can I use the function, and is the above commentary enough? (I already commited the changes into cvs.) Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFDATq9g2jPD2DL3GURAgudAKCsSzND0v6YNAwe1T0zigFtNfGaXACgyxcV 7g9lPC4fEmEKOSk9wjQNEzg= =rqIC -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Tue Aug 16 16:45:22 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 16 Aug 2005 16:45:22 +0200 Subject: getline and _GNU_SORCE In-Reply-To: <200508152200.47440.jaksonaquino@yahoo.com.br> References: <200508140911.47685.jaksonaquino@yahoo.com.br> <20050815111930.GY27886@intevation.de> <200508152200.47440.jaksonaquino@yahoo.com.br> Message-ID: <20050816144522.GA16089@intevation.de> Hi Jakson, Am 15. Aug 2005 um 22:00:45 schrieb Jakson Aquino: > Em Monday 15 August 2005 08:19, Bernhard Reiter > The substitute getline function wasn't allocating > memory dynamically. > > But now it is. I included into data.c the function > getndelim2() from gnucoreutil/lib/getndelim2.c with > some changes and added the following commentary above > the function: > > /* Adapted from GNU coretutils 5.2.1 getndelim2.c */ > int getline(char **lineptr, int *linesize, FILE *F){ > > Can I use the function, yes, the licences seem to be compatible. > and is the above commentary > enough? (I already commited the changes into cvs.) Yes, but you might want to add it to the copyright statement of data.c and the readme, so it is properly documented. Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050816/bbfd8791/attachment.bin From bernhard at intevation.de Tue Aug 16 16:49:22 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 16 Aug 2005 16:49:22 +0200 Subject: binaries and documentation In-Reply-To: <200508152155.12647.jaksonaquino@yahoo.com.br> References: <200508152155.12647.jaksonaquino@yahoo.com.br> Message-ID: <20050816144922.GB16089@intevation.de> Hi Jakson, Am 15. Aug 2005 um 21:55:01 schrieb Jakson Aquino: > I found a script in my computer that ask me some > questions and create a deb package. :) which one did you use? > I can, then use > alien to create a rpm package, and make both packages > available for download after the release. It is probably better then nothing, but hand crafted packages usually are better. They should be fairly easy to do. A Debian version was done some time ago, check the archive of the list. It should provide a start. > My question > is, what about the documentation? We don't have info > pages, and the pdf docs might be put > into /usr/share/doc/statist-1.3.0/ Fine, but why not "/usr/share/doc/statist/"? > but I fear that > some user migrating from Windows would have difficult > to guess where the documentation is. > > One idea is to add the item "Help" to the main menu. > Choosing the item would print a message telling where > the documentation can be found. I think a short manpage refering to the documentation would even be better. A point just pointing to the installed documentation seems to be a bit missleading. There are tools to create a simple manpage. Starting to use texinfo might be a hassle but also a future option, if someone is willing to do the work. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050816/81218fd5/attachment.bin From bernhard at intevation.de Tue Aug 16 16:52:06 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 16 Aug 2005 16:52:06 +0200 Subject: CVS: fixed tests; minor manual editing In-Reply-To: <200508152040.30715.jaksonaquino@yahoo.com.br> References: <20050811145134.GD29291@intevation.de> <200508140948.54031.jaksonaquino@yahoo.com.br> <20050815111342.GW27886@intevation.de> <200508152040.30715.jaksonaquino@yahoo.com.br> Message-ID: <20050816145206.GC16089@intevation.de> Hi Jakson, Am 15. Aug 2005 um 20:40:28 schrieb Jakson Aquino: > Em Monday 15 August 2005 08:13, Bernhard Reiter escreveu: > > Do you call "join" or "awk" directly from statist? > > Usually this would be done by the user. > > No, I don't call "join" and "awk" directly from statist. ah, good. > > I am unshure if we should add tests for this as statist > > does not do it. But it would be interesting to have example > > scripts for people to use. > > The test included "awk" and "join" because I wanted to check whether the > instructions that I wrote on the manual really work as I described. I was > testing the user manual, not only statist. Okay, this is good, though I am considering to move those tests into the doc directory under tests/ or so. > The fixed width data file that I > created had some empty spaces where they normally are not found in real fixed > width data files. This wasn't tested with the PNAD database that I used > before to test --xcols. I can create a directory doc/example with some > scripts and sample data files (but after the release of 1.3.0). This is also fine. > I liked the use of tests. They increase our trust in the code that we are > writting. Next time I add a feature to statist I will first write the result > as it have to be, and, then, write the code. Yes, I also really like it, this is the reason why I started it when redoing some of the memory stuff. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050816/db77182a/attachment.bin From jaksonaquino at yahoo.com.br Tue Aug 16 22:10:03 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Tue, 16 Aug 2005 17:10:03 -0300 Subject: binaries and documentation In-Reply-To: <20050816144922.GB16089@intevation.de> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <20050816144922.GB16089@intevation.de> Message-ID: <200508161710.19889.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 i Bernhard, Em Tuesday 16 August 2005 11:49, Bernhard Reiter escreveu: > Hi Jakson, > > Am 15. Aug 2005 um 21:55:01 schrieb Jakson Aquino: > > I found a script in my computer that ask me some > > questions and create a deb package. > > > :) > > which one did you use? It was a script written by Carlos Morimoto, the author of Kurumin, a Brazilian Linux distro based on Kanoppix (which is based on Debian). I'm not using the script. I just look at it to know what commands were necessary to create another script specific for statist. > It is probably better then nothing, but hand crafted packages > usually are better. They should be fairly easy to do. We can have total control over the debian package with the script. For the second step, the convertion into rpm, I'm asking "alien" to just create the directory and spec file, which have to be manually edited before building the package. The packages are working. I installed and uninstalled them in both Fedora Core 2 and Kurumin. I'm sending the script as an attachment, so you can make suggestions to improve it. The "man page" is inside the script. Although it's not formatted, it's correctly displayed. > A Debian version was done some time ago, > check the archive of the list. > It should provide a start. I tried the link, but I got an error with both firefox and wget. > There are tools to create a simple manpage. > Starting to use texinfo might be a hassle but also a future option, > if someone is willing to do the work. Is there something like tex2texinfo ? Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFDAkgjg2jPD2DL3GURAp8jAJ9q2FuC5x2aD2EZw4zsIk06Yl59TwCfcNDX 9R2Jav5R3XTWbvBdOoiFDdQ= =BCfy -----END PGP SIGNATURE----- -------------- next part -------------- A non-text attachment was scrubbed... Name: create_deb_rpm.sh Type: application/x-shellscript Size: 2416 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050816/2203cc68/create_deb_rpm.sh From bernhard at intevation.de Wed Aug 17 09:48:42 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 17 Aug 2005 09:48:42 +0200 Subject: binaries and documentation In-Reply-To: <200508161710.19889.jaksonaquino@yahoo.com.br> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <20050816144922.GB16089@intevation.de> <200508161710.19889.jaksonaquino@yahoo.com.br> Message-ID: <20050817074842.GF9192@intevation.de> Hi Jakson, Am 16. Aug 2005 um 17:10:03 schrieb Jakson Aquino: > Em Tuesday 16 August 2005 11:49, Bernhard Reiter > escreveu: > > Am 15. Aug 2005 um 21:55:01 schrieb Jakson Aquino: > > It is probably better then nothing, but hand crafted > > packages usually are better. They should be fairly easy to do. > > We can have total control over the debian package with > the script. Still, we could run it once and then have a debian directory we would check into CVS. > I'm sending > the script as an attachment, so you can make > suggestions to improve it. I think it probably lacks the Build dependencies and the dependencies. Also why should this be i386 only? > The "man page" is inside > the script. Although it's not formatted, it's > correctly displayed. Probably because he just installs a formatted one. ;) > > A Debian version was done some time ago, > > check the archive of the list. > > It should provide a start. > > I tried the link, but I got an error with both firefox > and wget. Which one did you try? Maybe we can try contacting the author again. > Is there something like tex2texinfo ? No. For now it is not worth the effort anyway, I think. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050817/70b61434/attachment.bin From bernhard at intevation.de Wed Aug 17 09:58:23 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 17 Aug 2005 09:58:23 +0200 Subject: deb packages (was: binaries and documentation) In-Reply-To: <20050817074842.GF9192@intevation.de> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <20050816144922.GB16089@intevation.de> <200508161710.19889.jaksonaquino@yahoo.com.br> <20050817074842.GF9192@intevation.de> Message-ID: <20050817075823.GA9863@intevation.de> Am 17. Aug 2005 um 09:48:42 schrieb Bernhard Reiter: > Am 16. Aug 2005 um 17:10:03 schrieb Jakson Aquino: > > Em Tuesday 16 August 2005 11:49, Bernhard Reiter escreveu: > > > A Debian version was done some time ago, > > > check the archive of the list. > > > It should provide a start. > > > > I tried the link, but I got an error with both firefox > > and wget. > > Which one did you try? > Maybe we can try contacting the author again. Following up on Carleos info documented in here: http://intevation.de/pipermail/statist-list/2004/000060.html I found out that you can find 1.0.1 packaged at: http://mentors.debian.net/debian/pool/main/s/statist/ We just need to find a Debian sponsor for http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=155073 . Maybe we can ask Carleos for help packaging the new version, too. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050817/ab356f5e/attachment.bin From jaksonaquino at yahoo.com.br Wed Aug 17 10:58:44 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Wed, 17 Aug 2005 05:58:44 -0300 Subject: deb packages (was: binaries and documentation) In-Reply-To: <20050817075823.GA9863@intevation.de> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <20050817074842.GF9192@intevation.de> <20050817075823.GA9863@intevation.de> Message-ID: <200508170558.46583.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, I don't know very much about building binary packages. I just copyied the i386 from the original script. It worked here, but might cause problems somewhere. Em Qua 17 Ago 2005 04:58, Bernhard Reiter escreveu: > Maybe we can ask Carleos for help packaging the new > version, too. Could you please, try to contact him? The better that I can do is reading something, and add some new variables to the control file. Much better would be if the package was built by someone with experience. Of course we don't need the binaries to release statist. We can add them later. Alternatively, we add the binaries and replace them latter. Since Statist is a simple program, I think that the binaries will work in most computers. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFDAvxEg2jPD2DL3GURAlndAJ9ccghLDd8JNd29Z1b4gJn2tmcr2gCfbgvT 74DcgPwLS6gSPIV9KMvDht8= =VT4O -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Wed Aug 17 11:32:45 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 17 Aug 2005 11:32:45 +0200 Subject: deb packages (was: binaries and documentation) In-Reply-To: <200508170558.46583.jaksonaquino@yahoo.com.br> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <20050817074842.GF9192@intevation.de> <20050817075823.GA9863@intevation.de> <200508170558.46583.jaksonaquino@yahoo.com.br> Message-ID: <20050817093245.GC12668@intevation.de> Hi Jakson, Am 17. Aug 2005 um 05:58:44 schrieb Jakson Aquino: > I don't know very much about building binary packages. I just > copyied the i386 from the original script. It worked here, > but might cause problems somewhere. Debian will have instructions. > Em Qua 17 Ago 2005 04:58, Bernhard Reiter escreveu: > > Maybe we can ask Carleos for help packaging the new > > version, too. > > Could you please, try to contact him? If you wish, I will write him an email and copy the list. > Since Statist is a simple program, I think that the binaries > will work in most computers. Yes, I agree. I run it on ppc with GNU/Linux without problems. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050817/801567e1/attachment.bin From jaksonaquino at yahoo.com.br Wed Aug 17 21:57:20 2005 From: jaksonaquino at yahoo.com.br (Jakson Aquino) Date: Wed, 17 Aug 2005 16:57:20 -0300 Subject: deb packages (was: binaries and documentation) In-Reply-To: <20050817093245.GC12668@intevation.de> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <200508170558.46583.jaksonaquino@yahoo.com.br> <20050817093245.GC12668@intevation.de> Message-ID: <200508171657.22942.jaksonaquino@yahoo.com.br> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Bernhard, Em Qua 17 Ago 2005 06:32, Bernhard Reiter escreveu: > > > Maybe we can ask Carleos for help packaging the new > > > version, too. > If you wish, I will write him an email and copy the list. I made some improvements in the script. But I think that it still not fulfilling all prerequisites for being accepted by Debian people as an official Debian package. It would be good to find someone to improve it. Best, Jakson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (GNU/Linux) iD8DBQFDA5agg2jPD2DL3GURAkDyAKCsfFjnEEMvqHFENJszeUPc6nWAhwCgxL46 L3BZlLjVeGhhA21DYFuFvyo= =YlRl -----END PGP SIGNATURE----- _______________________________________________________ Yahoo! Acesso Gr?tis - Internet r?pida e gr?tis. Instale o discador agora! http://br.acesso.yahoo.com/ From bernhard at intevation.de Fri Aug 26 19:33:22 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 26 Aug 2005 19:33:22 +0200 Subject: Statist 1.3.0 released! Jakson Aquino the new maintainer! Message-ID: <20050826173322.GA18719@intevation.de> Lightweight statist(ical) application with new maintainer and release! ---------------------------------------------------------------------- Statist 1.3.0 is available! This is a development quality release introducing a release numbering scheme similiar to Linux (the famous kernel): Odd numbers in second position mark experimental versions. Even numbers indicate stability. With this release Bernhard Reiter is handing over the task of maintaining and publishing Statist to Jakson Alves de Aquino. Jakson already was the main force behind this 1.3.0 release. Bernhard had published Statist from 1997 on, taking over from its original author Dirk Melcher. The statist community likes your feedback, e.g. on the mailinglist available from the homepage. What is Statist? ================ Statist is a small and portable statistics program written in C. It is terminal-based, but can use Gnuplot for plotting purposes. Scripting is possible and Statist handles big datasets reasonably well on small machines. It is also easy to use due to its interactive menu, being a very handy tool also for those who do not use statistics every day. Statist is Free Software under GNU GPL v>=2. THE DOCUMENTATION ================= The new statist user manual is available in both English and Portuguese. The German documentation is partially outdated, but it contains information about the statistical functions and the way the program works that are not covered by the new user manual. REQUIREMENTS ============ Statist is very portable. It can be compiled in pratically any system with any C compiler. However, Statist needs Gnuplot to create graphics, and the system must have libintl and libiconv to compile the program with support for languages other than English and German. Some GNU core utils -- gawk and sort -- are useful additions to prepare databases for use with Statist. DISTRIBUTION ============ Statist source code and some pre-compiled binaries from Statist's homepage: http://www.usf.uos.de/~breiter/tools/statist/index.en.html (It is likely that Statist will move to a new home in the near future.) NEW FEATURES IN VERSION 1.3.0 ============================= - Gettext support, with message catalogs for German, Italian, Portuguese, and Spanish. - English and Portuguese documentation. - The use of the option -delrow was replaced with a new system. Now, the rows are deleted during analyzes (when necessary), and we are using the smallest double number as missing value. It was necessary a lot of data manipulation to do some analyzes with the old system. - Added new features: (1) extract columns from fixed width data file; (2) extract a sample of rows from a file. - Small adaptions in the code to keep it compiling on DOS/Windows. Gettext support is possible if compiled with MinGW or CygWin. The use of pipes is possible with pgnuplot, which is part of new versions of gnuplot. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050826/523731b3/attachment.bin From bernhard at intevation.de Fri Aug 26 19:44:19 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 26 Aug 2005 19:44:19 +0200 Subject: Statist 1.3.0 released! Jakson Aquino the new maintainer! In-Reply-To: <20050826173322.GA18719@intevation.de> References: <20050826173322.GA18719@intevation.de> Message-ID: <20050826174419.GB18719@intevation.de> Am 26. Aug 2005 um 19:33:22 schrieb Bernhard Reiter: > > Lightweight statist(ical) application with new maintainer and release! > ---------------------------------------------------------------------- > > Statist 1.3.0 is available! > > This is a development quality release > introducing a release numbering scheme similiar to Linux (the > famous kernel): Odd numbers in second position mark experimental > versions. Even numbers indicate stability. I have also submitted this breaking news to freshmeat.net. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050826/ac5a96a5/attachment.bin From jaksonaquino at yahoo.com.br Tue Aug 30 10:40:21 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 30 Aug 2005 05:40:21 -0300 Subject: changes for 1.3.1 Message-ID: <20050830084021.GA897@k6> Hello! Below is the list of new changes for Statist 1.3.1 that are already in the cvs: * Eliminated the 60 columns limit: Now the limit to the number of columns depends on the the maximum number of files that the operating system can open simultaneously. Details: + New functions: void inflate_MCOL(); char * get_default_label(int i); void create_columns(int amount); void free_column(int i); void delete_column(int i); + Functions significantly modified: char *makefilename(int i); (was: char *makefilename(int, char*)) void erasetempfiles(); int parsecomment(char *comment); char * get_default_label(int i); (was: void create_label(int)) void readsourcefile(char *sourcename); + The global variable n_alias is no longer necessary: The comparison between the number of labels and the number of columns is now made only inside the function readsourcefile(). After this, n_alias has the same value of ncol. Thus, we don't need the two variables. * Changes in "tests" directory: + Added: test_xcols.sh, xcols.config, and xcols.dat + Makefile: added test_xcols among the possible tests, and added comments about the pre-requisites for running the tests. * Fixed memory leak in myexist(). * Fixed bug in plot_command(): The function was testing only if the user typed "quit". However, gnuplot quits whenever the first letters in a line might unambiguously be interpreted as "quit" or "exit". Gnuplot also skips leading blank characters. The bug was causing statist run an infinite loop when the wrong letters were sent as commands to gnuplot. Now, Statist is overzealous. It closes the pipe with gnuplot whenever the line first letters are either "q" or "ex". * Added the option --version: The usage tip was detached from VERSION_INFO. * Updated po files. * Added function ls(): The function is called when the user is asked to enter a file name, and, as the name suggests, it lists the content of the current directory. * Fixed bug in equal_rows(): The function was testing vn. The correct is to test nn. The bug was introduced while correcting the missing values problem and would manifest itself only when the user "Read another file" with different number of rows, but any user knowing what he was doing would never run analysis using columns of unrelated databases (unless he was testing Statist robustness). * Fixed bug in plot.c: xtics were disappearing after running the code: "set xtics %g, %g, %g". The command "set xtics" wasn't enough to reset the "start", "increment", and "end" values previously defined. Now we are using "set xtics autofreq". This change might only cause problems for people using very outdated versions of gnuplot, because the option "autofreq" is part of gnuplot since Dec 1998. * Replaced all gnuplot commands "set novar" with "unset var": The commands "set novar" were deprecated by gnuplot since Aug 17, 1999. * Name of gnuplot data file is no longer fix: If the name of the gnuplot data file already exists, a number is added to the name. Now it's possible to use the gnuplot command "replot" with multiple instances of statist. This closes the bug report opened on Sep 12, 1997. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050830/101a9a05/attachment.bin From jaksonaquino at yahoo.com.br Tue Aug 30 12:37:12 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 30 Aug 2005 07:37:12 -0300 Subject: Remove German messages? Message-ID: <20050830103712.GA1655@k6> Hello! Some German messages in the source code are either outdated or untranslated. It's unproductive to write strings in two languages simultaneously. It's also a potential cause of errors, even for German programmers, because they have to remember of updating the German string whenever the English one is changed. In my case, because I can't write in German, when I change an English message, I can't update the German one. Differently of what happens with the use gettext, if a message becomes outdated, there is no easy way of identifying it latter. Unless there is some person interested in maintaining the German messages inside the source code updated, I think it's better to remove them from the source code. Does anyone has objections? If not, I can remove the messages in the next days. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050830/2d97f7b4/attachment.bin From beyer at fli-leibniz.de Tue Aug 30 14:06:15 2005 From: beyer at fli-leibniz.de (Andreas Beyer) Date: Tue, 30 Aug 2005 14:06:15 +0200 Subject: Remove German messages? In-Reply-To: <20050830103712.GA1655@k6> References: <20050830103712.GA1655@k6> Message-ID: <43144BB7.6020903@fli-leibniz.de> I fully agree. I introduced the mechanism with the macros for dual languages. Before that time, statist was completely German. The reasons that I didn't use gettext were (1) it was more simple to implement and (2) gettext was not supported on all platforms at that time. But I always had the future use of gettext in mind. Andreas -- Dr. Andreas Beyer beyer at fli-leibniz.de www.fli-leibniz.de/~beyer From bernhard at intevation.de Tue Aug 30 15:26:15 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 30 Aug 2005 15:26:15 +0200 Subject: Remove German messages? In-Reply-To: <20050830103712.GA1655@k6> References: <20050830103712.GA1655@k6> Message-ID: <20050830132615.GN13049@intevation.de> Am 30. Aug 2005 um 07:37:12 schrieb Jakson A. Aquino: > Unless there is some person interested in maintaining the > German messages inside the source code updated, I think it's > better to remove them from the source code. Does anyone has > objections? If not, I can remove the messages in the next > days. No objection, just an addition for completeness: Most of the problems could also be solved if we could just create the german po file directly out of the source code. So the alternatives to maintain this would be to write a script doing just that. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050830/15445c28/attachment.bin From jaksonaquino at yahoo.com.br Tue Aug 30 17:45:21 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 30 Aug 2005 12:45:21 -0300 Subject: Remove German messages? In-Reply-To: <20050830132615.GN13049@intevation.de> References: <20050830103712.GA1655@k6> <20050830132615.GN13049@intevation.de> Message-ID: <20050830154521.GA796@k6> Hi Bernhard, On Tue, Aug 30, 2005 at 03:26:15PM +0200, Bernhard Reiter wrote: > No objection, > just an addition for completeness: > > Most of the problems could also be solved if > we could just create the german po file directly out of > the source code. > So the alternatives to maintain this would be to write > a script doing just that. I think that the extraction of the German strings directly from the source code is just one step of a longer process. The steps that I have in mind are: (1) Identify what German strings in the source code are outdated in relation with their English counterparts. (2) Translate and/or correct these messages in the source code. (3) Run the script to extract the German and English messages and put then in the po file. Probably, the step 3 is the easiest. I think it can be solved with gettext + grep. Anyway, it's simpler to translate a po file using poedit than to do the translation directly inside the source code, because poedit checks immediately whether we did any spell or printf mistake. One advantage of translating directly in the source code would be that the messages are contextualized by the code around them. But with poedit (or kbabel) we can see the context if necessary. Other advantage of having only one language in the source code is that it becomes smaller and easier to read and fix. It's easier to run the code in our mind if it is entirely in the screen. Of course, if you prefer, we can try the scripts. Would you like to write them? Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050830/27d6a508/attachment.bin From bernhard at intevation.de Tue Aug 30 19:32:30 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 30 Aug 2005 19:32:30 +0200 Subject: Remove German messages? In-Reply-To: <20050830154521.GA796@k6> References: <20050830103712.GA1655@k6> <20050830132615.GN13049@intevation.de> <20050830154521.GA796@k6> Message-ID: <20050830173230.GG13049@intevation.de> Hi Jakson, Am 30. Aug 2005 um 12:45:21 schrieb Jakson A. Aquino: > On Tue, Aug 30, 2005 at 03:26:15PM +0200, Bernhard Reiter wrote: > > No objection, > > just an addition for completeness: > Of course, if you prefer, we can try the scripts. Would you > like to write them? I also believe it is too much work, this is why I did not object. Because I thought about that script possibility for a while, I wanted to add it to the eternal archives, though I realise it is not worth the effort and I do not have the time to try. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050830/81f287f5/attachment.bin From jaksonaquino at yahoo.com.br Tue Aug 30 20:54:57 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 30 Aug 2005 15:54:57 -0300 Subject: Remove German messages? In-Reply-To: <20050830173230.GG13049@intevation.de> References: <20050830103712.GA1655@k6> <20050830132615.GN13049@intevation.de> <20050830154521.GA796@k6> <20050830173230.GG13049@intevation.de> Message-ID: <20050830185457.GA711@k6> Hi Bernhard, On Tue, Aug 30, 2005 at 07:32:30PM +0200, Bernhard Reiter wrote: > Because I thought about that script possibility for a while, > I wanted to add it to the eternal archives, > though I realise it is not worth the effort and I do not have the > time to try. Then, since you and Andreas agree with me, I'll remove the German strings. I see the fact that statist is internationalized as indicative that it's a good made in Germany software. If I was German, I would be even more proud of seeing statist working in many languages than of hearing statist speaking German in any system it was compiled in. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050830/ecd12002/attachment.bin From jaksonaquino at yahoo.com.br Wed Aug 31 14:52:34 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Wed, 31 Aug 2005 09:52:34 -0300 Subject: German messages in po file only Message-ID: <20050831125234.GA553@k6> Hello! I removed the German messages from the source code. I also reindented the functions histogram() and standard(), and removed some empty spaces from the source code. Thus, it's better to run diff with the option -b to get only what really changed. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050831/e93218cd/attachment.bin From jaksonaquino at yahoo.com.br Wed Aug 31 14:53:07 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Wed, 31 Aug 2005 09:53:07 -0300 Subject: procs.c: part_corr() Message-ID: <20050831125307.GB553@k6> Hello! While running partial correlation with 5 variables, valgrind reported the following errors: ==2558== Use of uninitialised value of size 8 ==2558== at 0x804E9FA: part_corr (procs.c:90) ==2558== by 0x804BDCA: regress_menu (menue.c:499) ==2558== by 0x804A380: main_menue (menue.c:104) ==2558== by 0x804965C: main (statist.c:232) ==2558== ==2558== Use of uninitialised value of size 8 ==2558== at 0x804EAAF: part_corr (procs.c:91) ==2558== by 0x804BDCA: regress_menu (menue.c:499) ==2558== by 0x804A380: main_menue (menue.c:104) ==2558== by 0x804965C: main (statist.c:232) ==2558== ==2558== Use of uninitialised value of size 8 ==2558== at 0x804EF22: part_corr (procs.c:97) ==2558== by 0x804BDCA: regress_menu (menue.c:499) ==2558== by 0x804A380: main_menue (menue.c:104) ==2558== by 0x804965C: main (statist.c:232) To fix the bug, I replaced (procs.c, line 71): if (ncol == 4) { with: if (ncol >= 4) { Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050831/6a15aefd/attachment.bin From jaksonaquino at yahoo.com.br Wed Aug 31 16:57:12 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Wed, 31 Aug 2005 11:57:12 -0300 Subject: equal_freq() and compare_freq() Message-ID: <20050831145712.GA570@k6> Hello! The functions equal_freq() and compare_freq() were causing segmentation fault when there were more than 60 classes of frequency. Bug fixed: if(nclass > MCLASS){ out_err(MAT, ERR_FILE, ERR_LINE, _("Test aborted: There are more than %i classes of frequency.\n"), MCLASS); return; } Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050831/0dba5932/attachment.bin From beyer at fli-leibniz.de Wed Aug 31 17:42:32 2005 From: beyer at fli-leibniz.de (Andreas Beyer) Date: Wed, 31 Aug 2005 17:42:32 +0200 Subject: equal_freq() and compare_freq() In-Reply-To: <20050831145712.GA570@k6> References: <20050831145712.GA570@k6> Message-ID: <4315CFE8.3040904@fli-leibniz.de> Hi Jakson, You could also create a change-log file to record the changes you make. I would use the mailing list only for anouncing major changes (e.g. new features or removing features). But, that's only my opinion. Maybe Bernhard sees that differently. Jakson, keep going with this great work! Andreas Jakson A. Aquino wrote: >Hello! > >The functions equal_freq() and compare_freq() were causing segmentation >fault when there were more than 60 classes of frequency. Bug fixed: > > if(nclass > MCLASS){ > out_err(MAT, ERR_FILE, ERR_LINE, > _("Test aborted: There are more than %i classes of frequency.\n"), > MCLASS); > return; > } > >Best, > >Jakson > > > >------------------------------------------------------------------------ > >_______________________________________________ >Statist-list mailing list >Statist-list at intevation.de >https://intevation.de/mailman/listinfo/statist-list > -- Dr. Andreas Beyer beyer at fli-leibniz.de www.fli-leibniz.de/~beyer From jaksonaquino at yahoo.com.br Thu Sep 1 10:21:22 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Thu, 1 Sep 2005 05:21:22 -0300 Subject: equal_freq() and compare_freq() In-Reply-To: <4315CFE8.3040904@fli-leibniz.de> References: <20050831145712.GA570@k6> <4315CFE8.3040904@fli-leibniz.de> Message-ID: <20050901082122.GA509@k6> Hi Andreas, On Wed, Aug 31, 2005 at 05:42:32PM +0200, Andreas Beyer wrote: > You could also create a change-log file to record the changes you make. > I would use the mailing list only for anouncing major changes (e.g. new > features or removing features). I agree with you, and the changes are being registered in the file CHANGES. Until some months ago the traffic in this list was very low. So it wasn't necessary to think in splitting it two or three. But now I'm afraid that too much traffic in the list might lead some people to unsubscribe the list or to use a filter to delete all messages automatically as they arrive. I think that it would be better to have two separate lists, perhaps, three. One for release announcements, and, perhaps, pre-release announcement of the need of help to translate strings. Other for developers, where we would announce changes applied into the cvs. I think this is important because it becomes easier to all developers to check each other work, and correct things before the changes go to a released version. I have sent the last email, for example, because I'm not sure that my solution to the bug is really the best one, and because someone might think in a more informative and precise error message than what I wrote. In fact, I think we don't need a real list for developers. We can use our email software to send messages to three or four people simultaneously. The current statist-list would be used by users to report bugs and ask questions. Developers would report bugs in the developer's list. Major improvements in the cvs version of Statist could also be reported here. Bernhard will move statist to gforge system, which will make it easier to setup things such as lists and forums. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050901/8606a463/attachment.bin From jaksonaquino at yahoo.com.br Thu Sep 8 00:48:47 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Wed, 7 Sep 2005 19:48:47 -0300 Subject: New behavior for getcols()? And new printcol()? Message-ID: <20050907224847.GA859@k6> Hello! I added three new menu items for statist. Statist now also has its ~/.statistrc. All changes are described in the file CHANGES. The item "Data management | List data of columns" is a new version of "List data of one column". The original item is still there, but I think that we have to choose one of them. This new item lists as many columns simultaneously as the user chooses, and allow the jumping to any row. It even has an advantage over graphical spreadsheets: we can see side by side columns that are in the extremes of a table with many columns. "List data of columns" uses a new version of getcols(), and I propose the replacement of this function. Currently, the user has to answer how many columns he will use in an analysis before choosing the columns. However, in social sciences, when we begin to explore a new data base, we usually don't know very well what to do: sometimes we only know what variables we will use when we begin to choose them. Thus, getcols2() has a new behavior. The user begins immediately to choose the variables and finish the list of columns with a blank line. Statist, and not the user, counts how many variables were chosen. I still have to add to getcols2() the possibility of limiting the number of columns, as required by some analysis (for example, "Partial linear correlation", which works only with 3, 4 or 5 columns). The only disadvantage of the new getcols() that I see is that it will break old scripts. But these scripts will not begin to yield wrong results (what would be a serious problem). They will stop working, what is easy to fix. We might add a warning in the release announcement about that. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050907/0c5b1282/attachment.bin From carleos at uniovi.es Thu Sep 29 22:58:22 2005 From: carleos at uniovi.es (Carleos Artime) Date: Thu, 29 Sep 2005 22:58:22 +0200 Subject: deb packages In-Reply-To: <200508171657.22942.jaksonaquino@yahoo.com.br> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <200508170558.46583.jaksonaquino@yahoo.com.br> <20050817093245.GC12668@intevation.de> <200508171657.22942.jaksonaquino@yahoo.com.br> Message-ID: Jakson Aquino skribis: > Em Qua 17 Ago 2005 06:32, Bernhard Reiter escreveu: >> > > Maybe we can ask Carleos for help packaging the new >> > > version, too. >> If you wish, I will write him an email and copy the list. > > I made some improvements in the script. But I think that it > still not fulfilling all prerequisites for being accepted by > Debian people as an official Debian package. It would be > good to find someone to improve it. Hi all! First, sorry for taking so much time to answer. I'm very busy now, like all of you, I suppose... 1. I have not used "statist" for a long time. But building the Debian package was quite easy, becauses it uses GNU tools (configure & make). 2. I just received a message, that my ITP (intent to package) to Debian is closed, because it was too old and no Debian developer was interested in it. Sure it helped that I did not continue to build packages for the new versions. See http://bugs.debian.org/155073 3. If there is an interest, I can "reopen" that "bug" and try to build the new "statist". I remember that last time we discussed whether there should be a German or English version by default. By following this list, I presume that gettext support was improved thereafter, and now packaging "statist" should be even easier. Cheers, -- _________________________________________________________________________ Carlos Enrique Carleos Artime , GNU.org , Dep-to de Statistiko kaj Plejbonigiko, / \ kaj Matematika Didaktiko ((__-^^-,-^^-__)) Universitato de Oviedo `-_---' `---_-' `--|o` 'o|--' Retposhto: carleos at uniovi.es \ ` / Telefono: +34 985 181 904 ): :( Fakso: +34 985 181 902 :o_o: Adreso: Escuela de Peritos, 33203 Hispanio "-" From bernhard at intevation.de Fri Sep 30 02:27:32 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 30 Sep 2005 02:27:32 +0200 Subject: deb packages In-Reply-To: References: <200508152155.12647.jaksonaquino@yahoo.com.br> <200508170558.46583.jaksonaquino@yahoo.com.br> <20050817093245.GC12668@intevation.de> <200508171657.22942.jaksonaquino@yahoo.com.br> Message-ID: <20050930002732.GB21526@intevation.de> Hi Carleos, Am 29. Sep 2005 um 22:58:22 schrieb Carleos Artime: > First, sorry for taking so much time to answer. I'm very busy now, > like all of you, I suppose... no problem, it is good to read you now! > 1. I have not used "statist" for a long time. But building the Debian > package was quite easy, becauses it uses GNU tools (configure & make). What was the reason for not using it? Especially: Can we make you a user again? > 2. I just received a message, that my ITP (intent to package) to Debian > is closed, because it was too old and no Debian developer was > interested in it. Sure it helped that I did not continue to build > packages for the new versions. See http://bugs.debian.org/155073 > > 3. If there is an interest, I can "reopen" that "bug" and try to build > the new "statist". It is Jakson's say, but I'd say: Go ahead! > I remember that last time we discussed whether there > should be a German or English version by default. By following this > list, I presume that gettext support was improved thereafter, and now > packaging "statist" should be even easier. Yes, Jakson completely enabled gettext, so Statist is completely internationalised. Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20050930/095d9cf5/attachment.bin From jaksonaquino at yahoo.com.br Fri Sep 30 14:08:02 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Fri, 30 Sep 2005 09:08:02 -0300 Subject: deb packages In-Reply-To: <20050930002732.GB21526@intevation.de> References: <200508152155.12647.jaksonaquino@yahoo.com.br> <200508170558.46583.jaksonaquino@yahoo.com.br> <20050817093245.GC12668@intevation.de> <200508171657.22942.jaksonaquino@yahoo.com.br> <20050930002732.GB21526@intevation.de> Message-ID: <20050930120802.GA2363@yahoo.com.br> Hi Carleos, and Berhard, I'm glad that Carleos might build the Debian package. Although I have made it for release 1.3.0, it really wasn't following Debian rules, and I still don't know how to make it correctly. Thanks Carleos! Best, Jakson On Fri, Sep 30, 2005 at 02:27:32AM +0200, Bernhard Reiter wrote: > Hi Carleos, > > Am 29. Sep 2005 um 22:58:22 schrieb Carleos Artime: > > First, sorry for taking so much time to answer. I'm very busy now, > > like all of you, I suppose... > > no problem, it is good to read you now! > > > 1. I have not used "statist" for a long time. But building the Debian > > package was quite easy, becauses it uses GNU tools (configure & make). > > What was the reason for not using it? > Especially: Can we make you a user again? > > > 2. I just received a message, that my ITP (intent to package) to Debian > > is closed, because it was too old and no Debian developer was > > interested in it. Sure it helped that I did not continue to build > > packages for the new versions. See http://bugs.debian.org/155073 > > > > 3. If there is an interest, I can "reopen" that "bug" and try to build > > the new "statist". > > It is Jakson's say, but I'd say: Go ahead! > > > I remember that last time we discussed whether there > > should be a German or English version by default. By following this > > list, I presume that gettext support was improved thereafter, and now > > packaging "statist" should be even easier. > > Yes, Jakson completely enabled gettext, so Statist is completely > internationalised. > > Best, > Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20050930/14014380/attachment.bin From jaksonaquino at yahoo.com.br Wed Oct 26 11:34:05 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Wed, 26 Oct 2005 07:34:05 -0200 Subject: Translations for 1.3.1 Message-ID: <20051026093405.GA2012@localhost.localdomain> Hello! The strings for the next release of Statist probably will not change any more, and I would be grateful if we had volunteers to translate the new strings. The current status of translations is the following: pt_BR.po: 441 translated messages. de_DE.po: 391 translated, 24 fuzzy, 26 untranslated. es_ES.po: 384 translated, 24 fuzzy, 33 untranslated. it_IT.po: 371 translated, 25 fuzzy, 45 untranslated. Best regards, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051026/bcebffb3/attachment.bin From bernhard at intevation.de Thu Nov 3 11:55:47 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Thu, 3 Nov 2005 11:55:47 +0100 Subject: Translations for 1.3.1 In-Reply-To: <20051026093405.GA2012@localhost.localdomain> References: <20051026093405.GA2012@localhost.localdomain> Message-ID: <20051103105547.GJ11911@intevation.de> Am 26. Oct 2005 um 07:34:05 schrieb Jakson A. Aquino: > de_DE.po: 391 translated, 24 fuzzy, 26 untranslated. I checked in the result of a short po-edit run: de_DE.po: 400 translated, 18 fuzzy , 23 untranslated -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051103/3d744502/attachment.bin From jaksonaquino at yahoo.com.br Fri Nov 4 08:31:53 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Fri, 4 Nov 2005 05:31:53 -0200 Subject: Translations for 1.3.1 In-Reply-To: <20051103105547.GJ11911@intevation.de> References: <20051026093405.GA2012@localhost.localdomain> <20051103105547.GJ11911@intevation.de> Message-ID: <20051104073153.GC1459@localhost.localdomain> Thanks, Bernhard, Translations made until tomorrow will be in the next version, since I plan to upload the files on Sunday. Best, Jakson On Thu, Nov 03, 2005 at 11:55:47AM +0100, Bernhard Reiter wrote: > I checked in the result of a short po-edit run: > > de_DE.po: 400 translated, 18 fuzzy , 23 untranslated -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051104/4fa1b7e6/attachment.bin From bernhard at intevation.de Fri Nov 4 21:49:24 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 4 Nov 2005 21:49:24 +0100 Subject: Translations for 1.3.1 In-Reply-To: <20051104073153.GC1459@localhost.localdomain> References: <20051026093405.GA2012@localhost.localdomain> <20051103105547.GJ11911@intevation.de> <20051104073153.GC1459@localhost.localdomain> Message-ID: <20051104204924.GA20088@intevation.de> Am 4. Nov 2005 um 05:31:53 schrieb Jakson A. Aquino: > Translations made until tomorrow will be in the next > version, since I plan to upload the files on Sunday. A German translation is completed in CVS now. Might be rough at some edges. E.g. I did not fully grasp what "Valid %" meant in a fequency diagram. > On Thu, Nov 03, 2005 at 11:55:47AM +0100, Bernhard Reiter wrote: > > I checked in the result of a short po-edit run: > > > > de_DE.po: 400 translated, 18 fuzzy , 23 untranslated -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051104/8a99027e/attachment.bin From bernhard at intevation.de Fri Nov 4 21:50:19 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 4 Nov 2005 21:50:19 +0100 Subject: 1.3.1 things I have noticed Message-ID: <20051104205019.GB20088@intevation.de> Hi Jakson, some things I have noticed: a) --xcols does not seem to go well with --silent. I have noticed when trying run_statist.sh in your test_xcols.sh script. b) I have noticed a bug while noticing a), even if the test failed, it did not properly report it. I have fixed this in CVS. c) Jakson: You should add yourself on the welcome message as copyright holder for 2005. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051104/761196ae/attachment.bin From jaksonaquino at yahoo.com.br Fri Nov 4 22:57:12 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Fri, 4 Nov 2005 19:57:12 -0200 Subject: Translations for 1.3.1 In-Reply-To: <20051104204924.GA20088@intevation.de> References: <20051026093405.GA2012@localhost.localdomain> <20051103105547.GJ11911@intevation.de> <20051104073153.GC1459@localhost.localdomain> <20051104204924.GA20088@intevation.de> Message-ID: <20051104215712.GA1253@localhost.localdomain> Hi Bernhard, Thanks for the translation! On Fri, Nov 04, 2005 at 09:49:24PM +0100, Bernhard Reiter wrote: > Might be rough at some edges. > E.g. I did not fully grasp what "Valid %" meant in a fequency diagram. It's the percentage of data points in relation to the number of valid data points (total - missing). For example, if the column has 113 rows, but 13 are missing values, "%" will be (100 * N/113) and "Valid %" will be (100 * N/100). Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051104/fe02ac08/attachment.bin From jaksonaquino at yahoo.com.br Sat Nov 5 00:25:57 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Fri, 4 Nov 2005 21:25:57 -0200 Subject: 1.3.1 things I have noticed In-Reply-To: <20051104205019.GB20088@intevation.de> References: <20051104205019.GB20088@intevation.de> Message-ID: <20051104232557.GA1478@localhost.localdomain> Hi Bernhard, On Fri, Nov 04, 2005 at 09:50:19PM +0100, Bernhard Reiter wrote: > a) > --xcols does not seem to go well with --silent. > I have noticed when trying run_statist.sh in your test_xcols.sh > script. --xcols is hard coded to be used like: statist --xcols config_file fixed_width_data_file dest_file and not like: statist [options] --xcols config_file fixed_width_data_file dest_file Although this sequence of commands is documented in the --help message, in the man page, and in the user manual, something more flexible would be better. Currently, the parameters passed to the program must be exactly the 4 above and in the same sequence. I can add this problem to KNOWN_BUGS and try to make it more flexible for the next release. > c) Jakson: You should add yourself on the welcome message > as copyright holder for 2005. Do you mean something like the following?: ------------------ jakson@:src$ ./statist --version STATIST --- Version 1.3.1-cvs --- (C) Dirk Melcher 1997-1999 (C) Jakson Aquino 2005 Published by Bernhard Reiter 1998-2005. STATIST comes with ABSOLUTELY NO WARRANTY; it is Free Software under GNU GPL. Read the file COPYING for details. ------------------- Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051104/412207a0/attachment.bin From jaksonaquino at yahoo.com.br Sun Nov 6 14:23:11 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Sun, 6 Nov 2005 11:23:11 -0200 Subject: Release Announcement: STATIST 1.3.1 Message-ID: <20051106132310.GA1426@localhost.localdomain> Lightweight statist(ical) application release! ---------------------------------------------------------------------- Statist 1.3.1 is available! This is a new development quality release The statist community likes your feedback, e.g. on the mailinglist available from the homepage. What is Statist? ================ Statist is a small and portable statistics program written in C. It is terminal-based, but can use Gnuplot for plotting purposes. Scripting is possible and Statist handles big datasets reasonably well on small machines. It is also easy to use due to its interactive menu, being a very handy tool also for those who do not use statistics every day. Statist is Free Software under GNU GPL v>=2. THE DOCUMENTATION ================= The statist user manual is available in both English and Portuguese. The German documentation is partially outdated, but it contains information about the statistical functions and the way the program works that are not covered by the new user manual. REQUIREMENTS ============ Statist is very portable. It can be compiled in practically any system with any C compiler. However, and the system must have libintl and libiconv to compile the program with support for languages other than English. If gnuplot is installed, statist can create graphics. Some GNU core utils -- gawk and sort -- are useful additions to prepare databases for use with Statist. DISTRIBUTION ============ Statist source code and some pre-compiled binaries are available at its homepage: http://wald.intevation.org/projects/statist/ INCOMPATIBLE CHANGES IN VERSION 1.3.1 ===================================== * The user has no longer to tell statist how many columns will be chosen. This new behavior might break most scripts written for previous versions of statist. If you are running statist under scripts, you should either adapt your scripts or not upgrade for version 1.3.1. NEW FEATURES IN VERSION 1.3.1 ============================= * Statist can deal with value labels and column titles: with the option --labels, statist reads column and value labels from a file. Example: You might have a column with the name "tvhours", and the title "How many hours do you watch TV every day?". * Eliminated the 60 columns limit: Now the limit to the number of columns depends on the the maximum number of files that the operating system can open simultaneously. * You can edit a configuration file (~/.statistrc) to customize some features of statist. * Menu item "Preferences": You can also customize some features while statist is running, and save your new preferences. Comments in the ~/.statistrc file are preserved. * Statist will list the content of current directory when the user is asked to enter a file name. * Menu item "Export data base as fixed width data file". * Replaced "List data of one column" with "List data of columns". It's possible to see data of many columns simultaneously. * Improved the menu option "List names of columns". It's possible to show not only column names, but also column titles and value labels. * New descriptive menu items: "Frequency table", and "Compare means". BUGS FIXED IN VERSION 1.3.1 ============================= * "Data manipulation / Invert values (1/x)": the inverted values were not corresponding to the original ones because rows with missing values were being deleted before the inversion. * Gnuplot graphics: the xtics were disappearing from the graphics after some analysis. The new command sent by statist to gnuplot avoids this. * Name of gnuplot data file is no longer fix: If the name of the gnuplot data file already exists, a number is added to the name. Now it's possible to use the gnuplot command "replot" with multiple instances of statist. This closes the bug report opened on Sep 12, 1997. * Uninitialized variables were being used when running "Partial linear correlation". * "Miscellaneous / Enter gnuplot commands": The function was testing only if the user typed "quit". However, gnuplot quits whenever the first letters in a line might unambiguously be interpreted as "quit" or "exit". Gnuplot also skips leading blank characters. The bug was causing statist to run in an infinite loop when the wrong letters were sent as commands to gnuplot. Now, Statist is overzealous. It closes the pipe with gnuplot whenever the line first letters are either "q" or "ex". * "Multiple linear correlation" and "Cross-validation of multiple linear regression": The values of the columns "resi" and "pred_" were in the wrong rows if there were missing values. * "Probit analysis": Now there is a limit to the number of error messages printed in the screen if the user chooses the wrong variables. Without this limit, the number of errors printed were a problem with very big data files (thousands of rows). * "Log-transformation (base 10)" and "Log-transformation (natural logarithm)" are no longer aborted if there are values equal or lower than zero. The results become missing values. * Now, to pass more than one file as argument for statist is fatal error only if statist is running in silent mode. If in interactive mode, the user will only receive a warning. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051106/c6713d14/attachment.bin From bernhard at intevation.de Mon Nov 7 12:18:48 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 7 Nov 2005 12:18:48 +0100 Subject: Translations for 1.3.1 In-Reply-To: <20051104215712.GA1253@localhost.localdomain> References: <20051026093405.GA2012@localhost.localdomain> <20051103105547.GJ11911@intevation.de> <20051104073153.GC1459@localhost.localdomain> <20051104204924.GA20088@intevation.de> <20051104215712.GA1253@localhost.localdomain> Message-ID: <20051107111848.GA20418@intevation.de> Am 4. Nov 2005 um 19:57:12 schrieb Jakson A. Aquino: > Thanks for the translation! You are welcome of course. :) > On Fri, Nov 04, 2005 at 09:49:24PM +0100, Bernhard Reiter wrote: > > Might be rough at some edges. > > E.g. I did not fully grasp what "Valid %" meant in a fequency diagram. > > It's the percentage of data points in relation to the number > of valid data points (total - missing). For example, if the > column has 113 rows, but 13 are missing values, "%" will be > (100 * N/113) and "Valid %" will be (100 * N/100). Ah, then I got it right. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051107/db33e9fe/attachment.bin From bernhard at intevation.de Mon Nov 7 12:21:34 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 7 Nov 2005 12:21:34 +0100 Subject: 1.3.1 things I have noticed In-Reply-To: <20051104232557.GA1478@localhost.localdomain> References: <20051104205019.GB20088@intevation.de> <20051104232557.GA1478@localhost.localdomain> Message-ID: <20051107112134.GB20418@intevation.de> Hi Jakson, Am 4. Nov 2005 um 21:25:57 schrieb Jakson A. Aquino: > On Fri, Nov 04, 2005 at 09:50:19PM +0100, Bernhard Reiter wrote: > > a) > > --xcols does not seem to go well with --silent. > > I have noticed when trying run_statist.sh in your test_xcols.sh > > script. > > --xcols is hard coded to be used like: > statist --xcols config_file fixed_width_data_file dest_file > > and not like: > statist [options] --xcols config_file fixed_width_data_file dest_file > > Although this sequence of commands is documented in the > --help message, in the man page, and in the user manual, > something more flexible would be better. Currently, the > parameters passed to the program must be exactly the 4 above > and in the same sequence. I can add this problem to > KNOWN_BUGS and try to make it more flexible for the next > release. Yes, this clearly is a bug and should be different. > > c) Jakson: You should add yourself on the welcome message > > as copyright holder for 2005. > > Do you mean something like the following?: > > ------------------ > jakson@:src$ ./statist --version > > STATIST --- Version 1.3.1-cvs --- > (C) Dirk Melcher 1997-1999 > (C) Jakson Aquino 2005 > Published by Bernhard Reiter 1998-2005. > STATIST comes with ABSOLUTELY NO WARRANTY; it is Free Software under GNU GPL. > Read the file COPYING for details. > ------------------- Yes, you can make this "(C)" for me, too, as I ended up writing some code over the years. Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051107/0f4cfc02/attachment.bin From bernhard at intevation.de Mon Nov 7 12:31:09 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 7 Nov 2005 12:31:09 +0100 Subject: Release Announcement: STATIST 1.3.1 In-Reply-To: <20051106132310.GA1426@localhost.localdomain> References: <20051106132310.GA1426@localhost.localdomain> Message-ID: <20051107113109.GF20418@intevation.de> Jakson: Am 6. Nov 2005 um 11:23:11 schrieb Jakson A. Aquino: > Lightweight statist(ical) application release! > Statist 1.3.1 is available! > This is a new development quality release Congratulations! Good Work! Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051107/0d7256d2/attachment.bin From jaksonaquino at yahoo.com.br Mon Nov 7 18:45:40 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Mon, 7 Nov 2005 15:45:40 -0200 Subject: 1.3.1 things I have noticed In-Reply-To: <20051107112134.GB20418@intevation.de> References: <20051104205019.GB20088@intevation.de> <20051104232557.GA1478@localhost.localdomain> <20051107112134.GB20418@intevation.de> Message-ID: <20051107174540.GA1177@localhost.localdomain> Hi Bernhard, On Mon, Nov 07, 2005 at 12:21:34PM +0100, Bernhard Reiter wrote: > > --xcols is hard coded to be used like: > > statist --xcols config_file fixed_width_data_file dest_file > > and in the same sequence. I can add this problem to > > KNOWN_BUGS and try to make it more flexible for the next > > release. > > Yes, this clearly is a bug and should be different. I added it to the KNOWN_BUGS before the release. Is it enough if statist accepts the following: statist [options] --xcols config_file fixed_w_file dest_file [options] Or even more flexibility is desirable? > > > c) Jakson: You should add yourself on the welcome message > > > as copyright holder for 2005. > > > > Do you mean something like the following?: > > > > STATIST --- Version 1.3.1-cvs --- > > (C) Dirk Melcher 1997-1999 > > (C) Jakson Aquino 2005 > > Published by Bernhard Reiter 1998-2005. > > STATIST comes with ABSOLUTELY NO WARRANTY; it is Free Software under GNU GPL. > > Yes, you can make this "(C)" for me, too, as I ended up writing some code > over the years. Then, we'll have: STATIST --- Version 1.3.1-cvs --- (C) Dirk Melcher 1997-1999 (C) Bernhard Reiter 1998-2005. (C) Jakson Aquino 2005 STATIST comes with ABSOLUTELY NO WARRANTY; it is Free Software under GNU GPL. But what about the "Published by" expression? We still have to include it? Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051107/2736d1b7/attachment.bin From bernhard at intevation.de Sat Nov 12 17:37:15 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sat, 12 Nov 2005 17:37:15 +0100 Subject: 1.3.1 things I have noticed In-Reply-To: <20051107174540.GA1177@localhost.localdomain> References: <20051104205019.GB20088@intevation.de> <20051104232557.GA1478@localhost.localdomain> <20051107112134.GB20418@intevation.de> <20051107174540.GA1177@localhost.localdomain> Message-ID: <20051112163715.GD18725@intevation.de> Am 7. Nov 2005 um 15:45:40 schrieb Jakson A. Aquino: > On Mon, Nov 07, 2005 at 12:21:34PM +0100, Bernhard Reiter wrote: > Is it enough if statist accepts the following: > > statist [options] --xcols config_file fixed_w_file dest_file [options] > > Or even more flexibility is desirable? should be fine enough, though, if you want --xcols to be an option with three parameters, this picture would look like: statist [options] :) > Then, we'll have: > > STATIST --- Version 1.3.1-cvs --- > (C) Dirk Melcher 1997-1999 > (C) Bernhard Reiter 1998-2005. > (C) Jakson Aquino 2005 > STATIST comes with ABSOLUTELY NO WARRANTY; it is Free Software under GNU GPL. > > But what about the "Published by" expression? We still have > to include it? No, I think it is fine when the history is preserved in the documentation. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051112/ac0f0859/attachment.bin From jaksonaquino at yahoo.com.br Mon Nov 14 00:22:57 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Sun, 13 Nov 2005 21:22:57 -0200 Subject: colors Message-ID: <20051113232257.GA6950@localhost.localdomain> Hello all! I'm developing a color scheme for statist. The use of colors will be optional. The color must be activated by either a command line option (--color) or an option in the preferences file (statistrc: color = yes). Below is what I added to my ~/.statistrc. I also attached to this mail a html file with the resulting output of using this color scheme. So far my creativity was only enough to use 5 colors. Any suggestion of a different color scheme is welcome. Best, Jakson # file: ~/.statistrc: # Color scheme. The color names have the format # , where the attribute part of the name # is optional, and might be "bright", "dim", "blink" or # "underln"; and the colors might be "red", "yellow", # "green", "blue", "cyan", "magenta", "white", or "black". # Warning: "dim", "blink", and "underln" might not work as # expected in some terminal emulators. cl_error = brightred cl_instructions = yellow cl_headers = brightwhite cl_menu_separator = blue cl_line_num = magenta -------------- next part -------------- A non-text attachment was scrubbed... Name: statist_output.html.gz Type: application/octet-stream Size: 3015 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051113/3c67d44a/statist_output.html.gz -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051113/3c67d44a/attachment.bin From jaksonaquino at yahoo.com.br Tue Nov 15 09:26:12 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 15 Nov 2005 06:26:12 -0200 Subject: "STATIST" in gnuplot graphics Message-ID: <20051115082612.GB1282@localhost.localdomain> Hi, Statist always includes in its graphics the word "STATIST". I think that the title of graphics and tables should only include information about the analysed data, not about the software used to create them. Thus, my suggestion is that the expression "STATIST: " should no longer be part of the title of graphics. Any objections? Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051115/5da8024b/attachment.bin From beyer at fli-leibniz.de Tue Nov 15 09:42:58 2005 From: beyer at fli-leibniz.de (Andreas Beyer) Date: Tue, 15 Nov 2005 09:42:58 +0100 Subject: "STATIST" in gnuplot graphics In-Reply-To: <20051115082612.GB1282@localhost.localdomain> References: <20051115082612.GB1282@localhost.localdomain> Message-ID: <43799F92.4000107@fli-leibniz.de> I am a fan of removing all unnecessary information from graphs. So, please go ahead and remove it. However, you might want to allow the user to define a string in the preferences that will preceed the specific title. Thus, it would be possible to have titles like this: "My study's title: This graph's title" The first part would be defined in the preferences. Andreas Jakson A. Aquino wrote: >Hi, > >Statist always includes in its graphics the word "STATIST". >I think that the title of graphics and tables should only >include information about the analysed data, not about the >software used to create them. Thus, my suggestion is that >the expression "STATIST: " should no longer be part of the >title of graphics. Any objections? > >Best, > >Jakson > > -- Dr. Andreas Beyer beyer at fli-leibniz.de www.fli-leibniz.de/~beyer From jaksonaquino at yahoo.com.br Tue Nov 15 10:24:07 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 15 Nov 2005 07:24:07 -0200 Subject: "STATIST" in gnuplot graphics In-Reply-To: <43799F92.4000107@fli-leibniz.de> References: <20051115082612.GB1282@localhost.localdomain> <43799F92.4000107@fli-leibniz.de> Message-ID: <20051115092407.GA1580@localhost.localdomain> Hi Andreas, On Tue, Nov 15, 2005 at 09:42:58AM +0100, Andreas Beyer wrote: > I am a fan of removing all unnecessary information from graphs. So, > please go ahead and remove it. > However, you might want to allow the user to define a string in the > preferences that will preceed the specific title. Thus, it would be > possible to have titles like this: > "My study's title: This graph's title" > The first part would be defined in the preferences. Good Idea! I'll do it. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051115/f7d070bb/attachment.bin From bernhard at intevation.de Tue Nov 15 16:00:15 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Tue, 15 Nov 2005 16:00:15 +0100 Subject: "STATIST" in gnuplot graphics In-Reply-To: <20051115082612.GB1282@localhost.localdomain> References: <20051115082612.GB1282@localhost.localdomain> Message-ID: <20051115150015.GC12179@intevation.de> Am 15. Nov 2005 um 06:26:12 schrieb Jakson A. Aquino: > Thus, my suggestion is that > the expression "STATIST: " should no longer be part of the > title of graphics. Any objections? No. From jaksonaquino at yahoo.com.br Tue Nov 15 17:59:24 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 15 Nov 2005 14:59:24 -0200 Subject: "STATIST" in gnuplot graphics In-Reply-To: <20051115150015.GC12179@intevation.de> References: <20051115082612.GB1282@localhost.localdomain> <20051115150015.GC12179@intevation.de> Message-ID: <20051115165924.GA3745@localhost.localdomain> Hi! On Tue, Nov 15, 2005 at 04:00:15PM +0100, Bernhard Reiter wrote: > Am 15. Nov 2005 um 06:26:12 schrieb Jakson A. Aquino: > > Thus, my suggestion is that > > the expression "STATIST: " should no longer be part of the > > title of graphics. Any objections? > > No. OK! The changes are in the cvs. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051115/aa6fe755/attachment.bin From bernhard at intevation.de Wed Nov 16 14:41:45 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 16 Nov 2005 14:41:45 +0100 Subject: colors In-Reply-To: <20051113232257.GA6950@localhost.localdomain> References: <20051113232257.GA6950@localhost.localdomain> Message-ID: <20051116134145.GA6339@intevation.de> Am 13. Nov 2005 um 21:22:57 schrieb Jakson A. Aquino: > I'm developing a color scheme for statist. The use of colors > will be optional. Good idea. Make sure that it will not contain vital information as many people cannot distinquish colors. Note that it probably is hard to get this right in a compatible way, because all the different terminal capabilities. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051116/577348a0/attachment.bin From bernhard at intevation.de Wed Nov 16 15:14:43 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Wed, 16 Nov 2005 15:14:43 +0100 Subject: CVS update: statist/po In-Reply-To: <20051115152457.36399101F02@lists.intevation.de> References: <20051115152457.36399101F02@lists.intevation.de> Message-ID: <20051116141443.GD7093@intevation.de> Am 15. Nov 2005 um 16:24:57 schrieb statist-cvs at intevation.de: > Modified Files: > Makefile > Log Message: > * Added "-f" to avoid error messages if the directory already is clean Usually this is not the right fix for Makefile. Use "-" in front of the command to ignore errors, -f is potentially dangerous, e.g. if the replacement of the variables fails. Bernhard From jaksonaquino at yahoo.com.br Thu Nov 17 08:00:00 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Thu, 17 Nov 2005 05:00:00 -0200 Subject: colors In-Reply-To: <20051116134145.GA6339@intevation.de> References: <20051113232257.GA6950@localhost.localdomain> <20051116134145.GA6339@intevation.de> Message-ID: <20051117070000.GA1235@localhost.localdomain> On Wed, Nov 16, 2005 at 02:41:45PM +0100, Bernhard Reiter wrote: > Am 13. Nov 2005 um 21:22:57 schrieb Jakson A. Aquino: > > I'm developing a color scheme for statist. The use of colors > > will be optional. > > Good idea. > Make sure that it will not contain vital information as many people > cannot distinquish colors. OK! I'll not add any feature that depends on colors. And the colors are really optional. They must be explicitly activated. > Note that it probably is hard to get this right in a compatible way, > because all the different terminal capabilities. Yes, I only know to produce colors in the GNU/Linux terminal emulators, and I don't think that colors are really important. I only added them to make the output more funny. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051117/471e4389/attachment.bin From jaksonaquino at yahoo.com.br Thu Nov 17 08:01:11 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Thu, 17 Nov 2005 05:01:11 -0200 Subject: CVS update: statist/po In-Reply-To: <20051116141443.GD7093@intevation.de> References: <20051115152457.36399101F02@lists.intevation.de> <20051116141443.GD7093@intevation.de> Message-ID: <20051117070111.GB1235@localhost.localdomain> Hi Bernhard, On Wed, Nov 16, 2005 at 03:14:43PM +0100, Bernhard Reiter wrote: > Am 15. Nov 2005 um 16:24:57 schrieb statist-cvs at intevation.de: > > Modified Files: > > Makefile > > Log Message: > > * Added "-f" to avoid error messages if the directory already is clean > > Usually this is not the right fix for Makefile. > Use "-" in front of the command to ignore errors, > -f is potentially dangerous, e.g. if the replacement of the > variables fails. Thanks for noting the problem. I eliminated the "-f" from po/Makefile and doc/Makefile in my working directory. However, I didn't remove the "-f" from src/Makefile because it was there since 1.0.1. Then, I looked more carefuly at the output of the main Makefile (make clean) and noted that what is causing make to stop isn't the lack of "-f", but the undefined variable $(MAKEFLAGS) in the src/Makefile. This variable causes no problem if "make" is typed in src, but it's expanded to "w" when I type "make clean" in the root directory. I don't where this "w" comes from. Should the $(MAKEFLAGS) be removed? Perhaps we could transfer this discussion to the new statist-developers forum at wald.intevation.org. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051117/62c7d87f/attachment.bin From bernhard at intevation.de Fri Nov 18 15:47:23 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Fri, 18 Nov 2005 15:47:23 +0100 Subject: CVS update: statist/po In-Reply-To: <20051117070111.GB1235@localhost.localdomain> References: <20051115152457.36399101F02@lists.intevation.de> <20051116141443.GD7093@intevation.de> <20051117070111.GB1235@localhost.localdomain> Message-ID: <20051118144723.GA6486@intevation.de> Hi Jakson, Am 17. Nov 2005 um 05:01:11 schrieb Jakson A. Aquino: > Thanks for noting the problem. I eliminated the "-f" from > po/Makefile and doc/Makefile in my working directory. Cool. > However, I didn't remove the "-f" from src/Makefile because > it was there since 1.0.1. > > Then, I looked more carefuly at the output of the main > Makefile (make clean) and noted that what is causing make to > stop isn't the lack of "-f", but the undefined variable > $(MAKEFLAGS) in the src/Makefile. This variable causes no > problem if "make" is typed in src, but it's expanded to "w" > when I type "make clean" in the root directory. I don't > where this "w" comes from. Should the $(MAKEFLAGS) be > removed? I cannot say without examine it myself. We should note it as TODO somewhere. > Perhaps we could transfer this discussion to the new > statist-developers forum at wald.intevation.org. Most forums will take more time to use then a mailinglist for me. I have not seen the "statist-developers" forum there anyway. Bernhard From jaksonaquino at yahoo.com.br Sat Nov 19 18:31:21 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Sat, 19 Nov 2005 15:31:21 -0200 Subject: CVS update: statist/po In-Reply-To: <20051118144723.GA6486@intevation.de> References: <20051115152457.36399101F02@lists.intevation.de> <20051116141443.GD7093@intevation.de> <20051117070111.GB1235@localhost.localdomain> <20051118144723.GA6486@intevation.de> Message-ID: <20051119173121.GA15966@localhost.localdomain> Hi Bernhard, On Fri, Nov 18, 2005 at 03:47:23PM +0100, Bernhard Reiter wrote: > > where this "w" comes from. Should the $(MAKEFLAGS) be > > removed? > > I cannot say without examine it myself. > We should note it as TODO somewhere. I added the item to statist "Tasks" at wald with the lowest priority flag. Congratulations: The gforge system is working very well! > > Perhaps we could transfer this discussion to the new > > statist-developers forum at wald.intevation.org. > > Most forums will take more time to use then a mailinglist for me. > I have not seen the "statist-developers" forum there anyway. The "statist-developer" is visible to me only when I login in wal.intevation.org. I can't see it while I'm seen by wald as a "guest". The mailing list is better to me too. I only made the suggestion because some people in this list might no be interested in the details of the development. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051119/40b2dadf/attachment.bin From bernhard at intevation.de Mon Nov 21 10:59:10 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 21 Nov 2005 10:59:10 +0100 Subject: CVS update: statist/po In-Reply-To: <20051119173121.GA15966@localhost.localdomain> References: <20051115152457.36399101F02@lists.intevation.de> <20051116141443.GD7093@intevation.de> <20051117070111.GB1235@localhost.localdomain> <20051118144723.GA6486@intevation.de> <20051119173121.GA15966@localhost.localdomain> Message-ID: <20051121095910.GD26610@intevation.de> Hi Jakson, Am 19. Nov 2005 um 15:31:21 schrieb Jakson A. Aquino: > On Fri, Nov 18, 2005 at 03:47:23PM +0100, Bernhard Reiter wrote: > I added the item to statist "Tasks" at wald with the lowest > priority flag. Congratulations: The gforge system is working > very well! thanks. > > > Perhaps we could transfer this discussion to the new > > > statist-developers forum at wald.intevation.org. > > > > Most forums will take more time to use then a mailinglist for me. > > I have not seen the "statist-developers" forum there anyway. > > The "statist-developer" is visible to me only when I login > in wal.intevation.org. I can't see it while I'm seen by wald > as a "guest". I should see it then, but I cannot. > The mailing list is better to me too. I only > made the suggestion because some people in this list might > no be interested in the details of the development. So far we do not have much traffic on this list, so we can still use it as a catch all. If we get more traffic, we can think about an additional developers list. Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051121/20856d66/attachment.bin From jaksonaquino at yahoo.com.br Tue Nov 22 08:09:05 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Tue, 22 Nov 2005 05:09:05 -0200 Subject: examples Message-ID: <20051122070905.GA1280@localhost.localdomain> Hello All! I've created some data files that I think can help new users to learn how to prepare statist data files. To avoid copyright problems, all files contain unreal data. I'm thinking of adding an "examples" directory to statist tree. So, users will have something to run right after downloading statist. I'm sending the directory as attachment, and any suggestion of different files are welcome. Best regards, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: examples.tar.gz Type: application/octet-stream Size: 4602 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051122/51d0ea8e/examples.tar.gz -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051122/51d0ea8e/attachment.bin From jaksonaquino at yahoo.com.br Wed Nov 23 08:56:44 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Wed, 23 Nov 2005 05:56:44 -0200 Subject: change statist behavior? (y/N) Message-ID: <20051123075644.GA1432@localhost.localdomain> Hi, When the user has a question of the type (y/N), he must choose "y" explicitly to answer "yes". This is ok, because other software with text interface do the same (example: apt-get). However, unlike apt-get, statist aborts the operation if the user just press . I think that considering a as "No" would be more intuitive. I have no strong preference for the new proposed behavior, and I would like to know what more experienced users of Linux and Unix think is the best behavior. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051123/65f254a2/attachment.bin From bernhard at intevation.de Thu Nov 24 14:29:19 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Thu, 24 Nov 2005 14:29:19 +0100 Subject: change statist behavior? (y/N) In-Reply-To: <20051123075644.GA1432@localhost.localdomain> References: <20051123075644.GA1432@localhost.localdomain> Message-ID: <20051124132919.GD6811@intevation.de> Am 23. Nov 2005 um 05:56:44 schrieb Jakson A. Aquino: > When the user has a question of the type (y/N), he must > choose "y" explicitly to answer "yes". This is ok, because > other software with text interface do the same (example: > apt-get). However, unlike apt-get, statist aborts the > operation if the user just press . I think that > considering a as "No" would be more intuitive. I think it is fine, if the default action will be choosen if is hit. In the case of "(y/N)" this should be the answer "no". -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051124/a4deabed/attachment.bin From jaksonaquino at yahoo.com.br Wed Nov 30 07:44:41 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Wed, 30 Nov 2005 04:44:41 -0200 Subject: avoiding fatal errors Message-ID: <20051130064440.GA1351@localhost.localdomain> Hi, Next version of statist will have colors! This is the most visible change for users. The number of fatal errors when running in interactive mode was also reduced to a minimum. Wrong parameters, including the name for a non existent file, still are treated as fatal errors. But after started, statist should never quit abruptly due to either an user error or any other reason. If a fatal error happens, it's a bug and must be fixed. GNU recommends that failure to allocate memory should not be treated as fatal errors in programs running interactively. I think this suggestion very reasonable, but this wasn't implemented in statist yet. This is the only fatal error that allowed to happen in interactive mode, and I think that I can even add it to the known bugs. The current status of translations in the cvs is the following: Language Translated Fuzzy Untranslated pt_BR 451 0 0 de_DE 439 5 7 es_ES 382 29 40 it_IT 369 29 53 I think that the current messages will not change until the next version. Best, Jakson -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051130/fa3fcab4/attachment.bin From bernhard at intevation.de Sun Dec 11 17:21:44 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Sun, 11 Dec 2005 17:21:44 +0100 Subject: de_DE translation completed Message-ID: <20051211162144.GC1949@intevation.de> Hi Jakson, just now I have completed the de_DE translation in CVS. Best, Bernhard -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051211/86d53c06/attachment.bin From jaksonaquino at yahoo.com.br Mon Dec 12 02:51:52 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Sun, 11 Dec 2005 23:51:52 -0200 Subject: Statist 1.3.2 Released Message-ID: <20051212015152.GA1913@localhost.localdomain> Lightweight statist(ical) application release! ---------------------------------------------------------------------- Statist 1.3.2 is available! And now with colors! This is a new development quality release The statist community likes your feedback, e.g. on the mailinglist available from the homepage. What is Statist? ================ Statist is a small and portable statistics program written in C. It is terminal-based, but can use Gnuplot for plotting purposes. Scripting is possible and Statist handles big datasets reasonably well on small machines. It is also easy to use due to its interactive menu, being a very handy tool also for those who do not use statistics every day. Statist is Free Software under GNU GPL v>=2. THE DOCUMENTATION ================= The statist user manual is available in both English and Portuguese. The German documentation is partially outdated, but it contains information about the statistical functions and the way the program works that are not covered by the new user manual. REQUIREMENTS ============ Statist is very portable. It can be compiled in practically any system with any C compiler. However, and the system must have libintl and libiconv to compile the program with support for languages other than English. If gnuplot is installed, statist can create graphics. Some GNU core utils -- gawk and sort -- are useful additions to prepare databases for use with Statist. DISTRIBUTION ============ Statist source code is available at its homepage: http://wald.intevation.org/projects/statist/ INCOMPATIBLE CHANGES IN VERSION 1.3.2 ===================================== * When answering a question of the kind "y/N", if the user press , it will be assumed that the answer is "No". Until version 1.3.1, pressing would cancel the action. This new behavior may break some scripts. NEW FEATURES IN VERSION 1.3.2 ============================= * Colorized output. The colors are optional and customizable. This feature is not available if statist is compiled under DOS/Windows. * Added directory "examples", with some data files that will make statist easier for first time users. * Graphics labels no longer include the string "STATIST: ". Instead, users can easily add their own (optional) string, which can include single quotes, double quotes and line breaks ("\n"). * Almost all fatal errors were converted into normal errors while statist is running interactively. Wrong command line options, including wrong data file name, still is considered a fatal error because this is the quickest way of giving to users a second chance of passing the correct parameters. However, after succesfuly started, statist should never quits abruptly. The only exception is if it runs out of memory. * Commands sent to gnuplot are saved and are listed when the user chooses "Enter gnuplot commands". Users no longer need to know gnuplot syntax to make simple changes in the graphics (labels, colors, etc...). BUGS FIXED IN VERSION 1.3.2 ============================= * The options --xcols and --xsample no longer must be the first parameter passed to statist. This closes the bug reported on November 4, 2005. * Replaced the function tmpnam() with tmpfile() to avoid gcc warning that statist uses a dangerous function. * The probit analysis was opening the already opened file statist.log and, then, closing it. Fixed. * Fixed bug in "List data of columns". Big values could cause buffer overrun. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051211/1a1ac160/attachment.bin From jaksonaquino at yahoo.com.br Mon Dec 12 03:21:31 2005 From: jaksonaquino at yahoo.com.br (Jakson A. Aquino) Date: Mon, 12 Dec 2005 00:21:31 -0200 Subject: de_DE translation completed In-Reply-To: <20051211162144.GC1949@intevation.de> References: <20051211162144.GC1949@intevation.de> Message-ID: <20051212022131.GA2154@localhost.localdomain> Hi Bernhard, Thanks for the translation! Best, Jakson On Sun, Dec 11, 2005 at 05:21:44PM +0100, Bernhard Reiter wrote: > Hi Jakson, > > just now I have completed the de_DE translation in CVS. > > Best, > Bernhard > -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://www.intevation.de/pipermail/statist-list/attachments/20051212/3efaf1de/attachment.bin From bernhard at intevation.de Mon Dec 12 12:34:24 2005 From: bernhard at intevation.de (Bernhard Reiter) Date: Mon, 12 Dec 2005 12:34:24 +0100 Subject: Statist 1.3.2 Released In-Reply-To: <20051212015152.GA1913@localhost.localdomain> References: <20051212015152.GA1913@localhost.localdomain> Message-ID: <20051212113424.GP7304@intevation.de> Am 11. Dec 2005 um 23:51:52 schrieb Jakson A. Aquino: > Statist 1.3.2 is available! And now with colors! Congratulations on the new release! -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://www.intevation.de/pipermail/statist-list/attachments/20051212/c300a205/attachment.bin