[fusion-users] RE: [fusion-dev] fusion 1.1 planning

[Jason Birch]

Is Fusion using the MapGuide MapAgent JSON output implemented in:

http://trac.osgeo.org/mapguide/wiki/MapGuideRfc25

yet?  I've managed to get a bit lost trying to figure out what is going
on, but as far as I can tell it's doing requests without specifying
content type, and post-processing the XML into JSON.  I'd imagine that
this would have substantial performance benefits.

If the json format output by MapGuide is the same as that used by the
internal xml2json stuff, this probably wouldn't be a huge change.  But
it may need to be configurable if supporting < 2.0 MapGuide releases is
important?

Jason

-----Original Message-----
From: Paul Spencer
Subject: [fusion-dev] fusion 1.1 planning

We'd like to start planning for a fusion 1.1 release this summer  
sometime.  I don't have a specific feature set or time frame in mind  
for a release, but it would be nice to have it before FOSS4G (Sept  
2008).

My primary goals for a 1.1 release are:

1) improve performance (and reduce library size),
2) take advantage of more OpenLayers features,
3) API cleanup and documentation

Currently, Mike is working on the following aspects (some are already  
complete):

* upgrade to OpenLayers 2.6 final
* move from Prototype.js framework to OpenLayers equivalents (freeing  
us up to move away from Prototype.js and Scriptaculous in the Jx  
library)
* make more extensive use of OpenLayers handlers/controls in place of  
existing widget code that does the same thing
* implement support for other OpenLayers layer types through the  
MapGroup tag in the ApplicationDefinition
* improved single file builds (including a 'compiled application'  
capability)

I will be working on a new version of Jx that will use MooTools  
instead of Prototype.js and Scriptaculous which will give Jx a smaller  
final size and more cinematic effects.  I will be updating Fusion to  
use the new Jx for 1.1 as well.

The net effect of all these changes will be to reduce the size of the  
code base (streamlining initial start up) and remove extraneous code  
(probably not a noticeable runtime effect yet).

There will also be some bug fixes of course, but I don't see any  
substantial work going into new widgets unless there are some new  
contributors willing to get involved.