From mlennert at club.worldonline.be Wed Oct 20 00:12:53 2004 From: mlennert at club.worldonline.be (Moritz Lennert) Date: Wed, 20 Oct 2004 00:12:53 +0200 (CEST) Subject: [Thuban-list] sized point data Message-ID: <32858.83.134.238.100.1098223973.squirrel@vivegnulinux.homelinux.org> Hello, I just browsed through the archives of the development list and discovered that sized point data has been implemented. This is extremely good news ! (and I hope that you did get paid for it ;-) ) I just did a fresh cvs checkout and tried it and it looks promising. Obviously, the inclusion of point size in the classification would make life even more beautiful... And to push you all further to new heights: how about a combination of point size and point color representing two separate classifications (ex: point size=current population size of a city and point color=population growth rate between 1990 and now). Would this be very complicated ? But before I give the impression of ingratitude, I would like to thank you all (and especially Jan) for this great new feature ! Moritz From mlennert at club.worldonline.be Wed Oct 20 00:14:43 2004 From: mlennert at club.worldonline.be (Moritz Lennert) Date: Wed, 20 Oct 2004 00:14:43 +0200 (CEST) Subject: [Thuban-list] tests for Daniel Cavelo's classifiers Message-ID: <32862.83.134.238.100.1098224083.squirrel@vivegnulinux.homelinux.org> Hello, Daniel told me that the lack of test keeps his patches for the classifier from being integrated into the CVS. Could someone explain to me what exactly is needed in terms of tests and how I could contribute ? Thanks ! Moritz From jan at intevation.de Wed Oct 20 07:48:01 2004 From: jan at intevation.de (Jan-Oliver Wagner) Date: Wed, 20 Oct 2004 07:48:01 +0200 Subject: [Thuban-list] sized point data In-Reply-To: <32858.83.134.238.100.1098223973.squirrel@vivegnulinux.homelinux.org> References: <32858.83.134.238.100.1098223973.squirrel@vivegnulinux.homelinux.org> Message-ID: <20041020054801.GD4897@intevation.de> On Wed, Oct 20, 2004 at 12:12:53AM +0200, Moritz Lennert wrote: > I just browsed through the archives of the development list and discovered > that sized point data has been implemented. This is extremely good news ! > (and I hope that you did get paid for it ;-) ) unfortunately not, I was just sitting in the train with my laptop ... > I just did a fresh cvs checkout and tried it and it looks promising. > Obviously, the inclusion of point size in the classification would make > life even more beautiful... hm, actually it is considered by the classification. > And to push you all further to new heights: how about a combination of > point size and point color representing two separate classifications (ex: > point size=current population size of a city and point color=population > growth rate between 1990 and now). Would this be very complicated ? I will move my mind about this feature. I added this to the Roadmap feature proposals anyway since it is definitly interesting. > But before I give the impression of ingratitude, I would like to thank you > all (and especially Jan) for this great new feature ! Hope it helps you for the case you describe some months ago. Best Jan -- Jan-Oliver Wagner http://intevation.de/~jan/ Intevation GmbH http://intevation.de/ FreeGIS http://freegis.org/ From jan at intevation.de Wed Oct 20 07:53:42 2004 From: jan at intevation.de (Jan-Oliver Wagner) Date: Wed, 20 Oct 2004 07:53:42 +0200 Subject: [Thuban-list] tests for Daniel Cavelo's classifiers In-Reply-To: <32862.83.134.238.100.1098224083.squirrel@vivegnulinux.homelinux.org> References: <32862.83.134.238.100.1098224083.squirrel@vivegnulinux.homelinux.org> Message-ID: <20041020055342.GE4897@intevation.de> On Wed, Oct 20, 2004 at 12:14:43AM +0200, Moritz Lennert wrote: > Daniel told me that the lack of test keeps his patches for the classifier > from being integrated into the CVS. Could someone explain to me what > exactly is needed in terms of tests and how I could contribute ? I tried to apply his patches some time ago but the did not apply cleanly to current CVS. I was about to analyse what tests are missing. So first step would be to apply the patch for HEAD. Next, one should think about what are the most important tests and add them to the directory test/ There you can see how the other tests are made - not too difficult in practice but some brain work to have them useful. test_color.py is a simple one to get the idea. Ah well, this is actually better for thuban-devel :-) Best Jan -- Jan-Oliver Wagner http://intevation.de/~jan/ Intevation GmbH http://intevation.de/ FreeGIS http://freegis.org/ From mlennert at club.worldonline.be Wed Oct 20 09:26:48 2004 From: mlennert at club.worldonline.be (Moritz Lennert) Date: Wed, 20 Oct 2004 09:26:48 +0200 (CEST) Subject: [Thuban-list] tests for Daniel Cavelo's classifiers In-Reply-To: <20041020055342.GE4897@intevation.de> References: <32862.83.134.238.100.1098224083.squirrel@vivegnulinux.homelinux.org> <20041020055342.GE4897@intevation.de> Message-ID: <32775.83.134.237.97.1098257208.squirrel@vivegnulinux.homelinux.org> On Wed, October 20, 2004 7:53, Jan-Oliver Wagner said: > On Wed, Oct 20, 2004 at 12:14:43AM +0200, Moritz Lennert wrote: >> Daniel told me that the lack of test keeps his patches for the >> classifier >> from being integrated into the CVS. Could someone explain to me what >> exactly is needed in terms of tests and how I could contribute ? > > I tried to apply his patches some time ago but the did not apply cleanly > to current CVS. I was about to analyse what tests are missing. Funny, I just applied the patches to a fresh CVS checkout yesterday and it worked without any problem... But you are right, I'll resubscribe to thuban-devel... Moritz From mlennert at club.worldonline.be Wed Oct 20 09:44:44 2004 From: mlennert at club.worldonline.be (Moritz Lennert) Date: Wed, 20 Oct 2004 09:44:44 +0200 (CEST) Subject: [Thuban-list] sized point data In-Reply-To: <20041020054801.GD4897@intevation.de> References: <32858.83.134.238.100.1098223973.squirrel@vivegnulinux.homelinux.org> <20041020054801.GD4897@intevation.de> Message-ID: <32803.83.134.242.37.1098258284.squirrel@vivegnulinux.homelinux.org> On Wed, October 20, 2004 7:48, Jan-Oliver Wagner said: > On Wed, Oct 20, 2004 at 12:12:53AM +0200, Moritz Lennert wrote: > >> I just did a fresh cvs checkout and tried it and it looks promising. >> Obviously, the inclusion of point size in the classification would make >> life even more beautiful... > > hm, actually it is considered by the classification. Either I didn't see how, or I was not clear enough. What I meant was that presently it seems that I have to set the size by hand. I can't just tell thuban that I want each point to be displayed automatically in a size that is proportional to a given column value (this could be a special case of "unique values" but where the classification does not result in a color ramp, but in a "size ramp"). The former is ok if I have a few sites or only few different sizes, but if I have several hundred "sites" then the latter becomes necessary. For the implementation of the proportionate size it would also be necessary to be able to determine a scale factor, i.e. a constant by which to multiply the size of all sites in order to adapt the sizes to the map. >> And to push you all further to new heights: how about a combination of >> point size and point color representing two separate classifications >> (ex: >> point size=current population size of a city and point color=population >> growth rate between 1990 and now). Would this be very complicated ? > > I will move my mind about this feature. > I added this to the Roadmap feature proposals anyway since it is > definitly interesting. Actually if the "automatic" size is implemented seperately from the classifier, and the classifier remembers the sizes, then it would be possible to reclassify the different point (colors) according to any given column. At present, using the iceland_sample_size session, when I reclassify the cultural landmarks, they all go back to the same size and I have to reset the size manually. If the classifier could remember the sizes (given either manually or through some automatic proportionate mechanism, then the combination of sizes and colors should be no problem. > >> But before I give the impression of ingratitude, I would like to thank >> you >> all (and especially Jan) for this great new feature ! > > Hope it helps you for the case you describe some months ago. It does. We are getting closer to having something that I might be able to sell to my colleagues. Many thanks again (including to the Deutsche Bahn ;-)) ! Moritz