Andreas and I were speaking earlier about what, in general, we need to do to finish MapBuilder 1.5 and what we're looking for in 2.0, and some ruminations about the project in general.
1.5 - Features Before Final ============================ Andreas: Tim Schaub has finished WMC read/write support for OpenLayers. This is done using Format classes, like for other OGC formats in OL. And many things that we have in MB are currently under development by community members in OL: Loading indicator, measurement tool, layer switcher with legend graphic. Just to mention some. I think if we decide to do MB 2.0, it will only be some glue between OL and our XML/XSL/MVC stuff. But that glue has to be well thought of and designed. A current task for 1.5rc2 would be to port it to OL trunk, so we are prepared for 2.6, which will probably happen before we release 1.5. It will be a little bit tricky to modify MB for OL 2.6, because there is now projection support directly in OL. So all projection code in MB has to be changed, most of it can be dropped. Andreas: Also, I was thinking about changing config.objects initialisation, so if a widget class does not exist, it falls back to WidgetBaseXSL. This would help us to drop a dozen of stub widget classes that do nothing. General Thoughts for 2.0 ======================== M: We should perhaps commence learning of the OL stack, from a porting perspective. A: But there is a way to do the XSL/MVC stuff with OL. We only have to figure out how it can be accomplished in a clean way that deserved to be called Mapbuilder 2.0 A: I suggest that instead of keeping the application state in XML models directly, OL.Format classes are the way to go for XML (de)serialisation. OL keeps a good state of everything in its objects. A: At least the map (upcoming Format.WMC) and layers (Format.GML and others) have this functionality. Not to forget styles (Format.SLD). Now we have a distinction between WFS layers added via the context model, and GML rendering using GmlRendererOL (formerly GmlRendererWZ). With a good design for 2.0, a layer could be loaded using WMC, and a separate layer model could be connected to any vector layer that has to do more than just being displayed. So attaching models to OL objects is something we should maybe think of. Thoughts about MapBuilder Overall ================================= Is seems that development activity has been slowly declining in the past half-year, as it has been noted, we've got more PSC members than active developers. I assume that this is for 3 reasons: 1) Many developers have added features to MB to get up to the point where their needs were satisfied, and with that, didn't see much more room for improvement, and probably felt that there isn't/wasn't much more to contibute. 2) The web mapping world has changed a lot in the last 2 years, OpenLayers and Google Maps hacks make it much easier to get up and running quickly, and doing development. This is probably robbing us of new developers. 3) Any project is going to have a natural out-migration of developers, what we lack is the incoming development. I think the idea behind MB becoming an OL plugin is terrific, and I've started looking into it on my end. What else can we do to spark activity here? Matthew D. Diez Software Engineer - GIS and Meteorology Innovative Emergency Management, Inc. 8550 United Plaza Boulevard, Suite 501 Baton Rouge, LA 70809 (225) 803-9991 ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ mapbuilder-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mapbuilder-devel
