[mapguide-internals] Viewers and the Future of Web Layout

[Paul Spencer]

I would be keen to have a common framework, but I am starting to get  
emotionally invested in my Fusion work and would have to be very  
convinced before moving away from it :)  I'm not saying our work is  
better or worse than anyone else's ... but I do like it a lot!

Also, I'm anti-java and don't really like the concept that GWT is  
built on (although I can see how it would compelling to a large group  
of users).  I don't know much about atlas, but I have similar  
reservations about .Net :)

Of course, I'm now a mac user so things can change!!!

I would like to see what SL-King is doing as well - and I'm sure that  
everyone would be interested in what I'm doing ;)  It may be that  
there is room, and need, for both - or the community may choose one  
approach over the other for MapGuide, which would be fine for me  
either way.

Unfortunately I'm still stupidly busy and don't really have the time  
right now to get some demos and docs together for the stuff I'm  
doing.  I will try to put this into the priority queue (after Jason!)  
and at least let folks take a look at what is happening in our shop  
on this front.

As a very brief overview, I would say it is most like the OSGeo  
MapBuilder project - we take any web page and associate 'Command's  
from the weblayout with HTML elements using the element's ID and  
dynamically load javascript that implements the command.  We do not  
use anything from the existing AJAX viewer except a couple of  
includes.  There is no need to use frames or iframes anywhere.  Just  
a single javascript file (fusion.js) to include in a web page and a  
call to Fusion.initialize() does the trick in most cases.

Most of the code is javascript.  There are a dozen or so PHP scripts,  
but the architecture is designed to allow these to be replaced by ASP  
or JSP pages without changes to the javascript code.  The biggest  
changes are in my mutilation of the WebLayout ... :)

Right now, we can replicate all the functionality of the AJAX viewer  
except for invoke URL (this is trivial), invoke Script (this is  
actually done, but I stupidly called it something else), and  
supporting tiled base maps.  I hope to use OpenLayers to support  
tiled base maps but I haven't figured out when I'm going to implement  
this.

There are a few new capabilities as well (reference map for  
instance), and some improvements (IMO) on some of the existing  
functionality.  Also, we've put some effort into building a generic  
javascript UI library that provides some slick menus, toolbars,  
trees, tabs, dialogs, splitters etc.  Everything is designed to be  
CSS skinnable, and maximize use of the power of browsers (so hovers  
on buttons are handled as <a> tags with css :hover psuedo classes for  
instance).

Anyway, its hard to describe so I guess I'll just have to try and get  
some code and demos out there!

Cheers

Paul

On 26-Feb-07, at 5:01 PM, Robert Bray wrote:

> Change Subject, was: foss4g workshops
>
> Hmm,
>
> Now I am worried. I was hoping that Fusion would lead to a next  
> generation web layout and we could continue a write once run  
> anywhere type of viewer strategy. Sounds like we really need to get  
> together and  decide on a strategy for moving forward. Deciding to  
> split is not a bad thing, but I want to make sure it is a  
> conscience decision as opposed to the project randomly growing  
> appendages.
>
> From my perspective the pro for keeping it all together is that we  
> can all take advantage of  one another's work. Once we have a  
> flexible client framework then folks can focus on developing new  
> widgets to plug into it. The only potential pro for splitting is to  
> take advantage of best of breed technology from different vendors,  
> but I am not totally convinced this cannot be achieved with the  
> first approach.
>
> What does everyone else think? Paul?
>
> Bob
>
> Jason Birch wrote:
>> Bob wrote:
>>
>>> Hows that RFC for Fusion support coming along anyway :)
>> I'm a little worried about direction with the AJAX viewers.  While  
>> the current viewer meets a large proportion of the users' needs,  
>> it's not very flexible and the map surround elements aren't very  
>> AJAX.
>>  DMSG has been working on Fusion as an alternative, and it looks  
>> like SL-King is looking at building a similar framework for Java  
>> based on GWT.  I could see something coming for .Net that  
>> leverages "atlas" in the same way.
>>  Do we need to talk about this to ensure that these apps are  
>> developed in a particular way (if they're going to be merged into  
>> the MapGuide project)?  Are we OK with different viewers having  
>> different APIs, etc?
>>  Jason
>> --------------------------------------------------------------------- 
>> ---
>> _______________________________________________
>> mapguide-internals mailing list
>> mapguide-internals at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/mapguide-internals
> _______________________________________________
> mapguide-internals mailing list
> mapguide-internals at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/mapguide-internals

+-----------------------------------------------------------------+
|Paul Spencer                          pspencer at dmsolutions.ca    |
+-----------------------------------------------------------------+
|Chief Technology Officer                                         |
|DM Solutions Group Inc                http://www.dmsolutions.ca/ |
+-----------------------------------------------------------------+