[mapguide-internals] Tile maps...

[Jason Birch]

Yes, I was thinking about this, but was also thinking that managing selections on two maps might be a bit annoying to deal with :)  As long as Paul's doing it I guess I don't have much to complain about...  That was a guarantee for the first integration, right Paul? :)
 
I think that with any serious tiling implementation, a requirement would be to be able to hit multiple hosts for the tiles in round-robin for a single user, even if the hosts are all answered by the same server.  The two-active-connections limit that most users face really slows down tile maps if they're fetching all of the tiles from a single host.  Especially when one of those pipes is being held up by an active JS script.
 
Personally, I would like to see the tiling stuff re-implemented as a rest-based filter / error handler on top of Apache/IIS/etc.  This would allow the straight files to be served natively/statically with incredible performance if they exist (with no map agent getting in the way), and would allow the map agent to only deal with requests that haven't already generated images.  It would also provide better utilization of distributed caching, with CTAGs, expiry headers, and the like all being there by default or easily configurable in the web server.
 
>From Paul's comments, I guess it would be possible to have an internal MapGuide WMS service consumed by something like TileCache, and then published out to Fusion from there... if it's possible to restrict WMS to a single host.  My problem with this is that my most common layer is probably the parcel base, and I'll need maptips, etc...  
 
I guess I could publish the visible base as a tileset, and then publish an invisible layer on top of it for maptips and selection?  Unless invisible polygons are optimised out of the selection process?
 
Jason

________________________________

From: mapguide-internals-bounces at lists.osgeo.org on behalf of Trevor Wekel
Sent: Fri 2007-09-14 4:35 PM
To: MapGuide Internals Mail List
Subject: RE: [mapguide-internals] Tile maps...



Hi Jason,

A possible enhancement would be to define two map definitions for a
"map".  The first definition would define the overlay/dynamic layers and
the second definition would be used only for the tiles.  I think they
are separate layers in the new client framework so it may be possible to
implement.

There would be some other implementation details like selection and
query that would have to be worked out.  You may need to hit two maps to
find the selected/queried object.  This could be a little tricky but I
don't think it would be impossible.  I suspect all the maps would also
have to use the same coordinate system.

Version 1.2.0 of the GETTILEIMAGE operation takes the resource id of the
map definition, the base map layer group name, the tile row and column
and the scale index.  Authentication can be specified using sessionid or
username/password.

A further enhancement to the "two layer" / "two map" approach would be
to code up two different URLs for the overlay/dynamic and tiled layers.
This would allow us offload the base tiles to a completely separate
server and web extensions.

Thanks,
Trevor

-----Original Message-----
From: mapguide-internals-bounces at lists.osgeo.org
[mailto:mapguide-internals-bounces at lists.osgeo.org] On Behalf Of Jason
Birch
Sent: Friday, September 14, 2007 4:38 PM
To: MapGuide Internals Mail List
Subject: [mapguide-internals] Tile maps...

Hi all,

I've been thinking (not a good sign) about how tile maps are implemented
in MapGuide.  I'm seeing some real limitations on how useful this is
going to be for me because of the association of tile sets with specific
maps. 

With the new Fusion framework, I intend to have several dozen maps
(themes), each using one of three tile set definitions.  With the
current inclusion of tile groups in map definitions, this means that
I'll have to have separate copies of these tile groups for each one of
my maps.

I would really like to see some way of abstracting tile groups from map
definitions.  Either through a new resource type that defines the tile
group, or by allowing tiled maps to be referenced from regular maps,
rather than redefined in each one.

Thoughts?

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