Thanks for your lights Martin. Is there any way for me to help you go further? From my point of view, I can write user requirements, but hardly help you on a technical implementation. Cheers, Régis
> -----Message d'origine----- > De : Martin Dobias [mailto:wonder...@gmail.com] > Envoyé : mardi 11 février 2014 10:33 > À : HAUBOURG > Cc : qgis-dev > Objet : Re: [Qgis-developer] Fwd: [SoC] OSGeo Google Summer of Code > 2014 - ideas needed! > > Hi > > On Fri, Feb 7, 2014 at 4:33 PM, Régis Haubourg <regis.haubourg@eau-adour- > garonne.fr> wrote: > > Hi, > > qgis is still very weak concerning pure cartographic uses cases with > > size varying symbols, and have no working legend for this, and more > > generally for all data defined symbology. > > I've already shared my ideas with Martin Dobias who seems to have > > legend refactoring in mind. > > > > Since this is critic for a cartographic tool, and since data defined > > symbologies can't be guessed before real data is rendered, we have > > something that needs some deep investigations. > > > > Putting this in a developpement contract seems dangerous in such early > > stage. It be could either be a topic for the next hackfest, or a GsoC. > > > > Martin, any opinion on this? Others? > > It could be a GSoC project if we knew what do we want to implement. > Currently we just have a use case in the mind ("generate nice legend for > symbols with data-defined scale"). But we have not yet elaborated the > concept how this would be done. The mentioned use case is probably not > the only scenario where it would be nice to have customized legend. So we > need to think more about other possible use cases (what to do if there are > combinations, e.g. data-defined scale, data-defined color, data-define > rotation, categories) and how the user would define their requirements. > Without a good design, we can end up with a poor solution that satisfies just > the simplest use case, but fails with all others. I do not feel confident > proposing such project without having a clear idea as that may lead to high > risk of failure. > > Regards > Martin _______________________________________________ Qgis-developer mailing list Qgis-developer@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/qgis-developer