[OSGeo-Discuss] FOSS4G & Tile WMS

[Chris Holmes]

Note that the list in question is at: 
http://lists.eogeo.org/mailman/listinfo/tiling

It would be great if we could get openlayers, ka-map, mapbuilder and 
udig (and anyone else!) to all implement a common tiling scheme, and 
then serve it up from geoserver, mapserver and mapguide.  Would be very 
cool if we could generalize ka-map's caching code to be generic for any 
WMS, able to crawl and cache a WMS, or to cache on the fly as clients 
request.  We can use squid in front of geoserver or mapguide, but I 
think there are some nice advantages to a dedicated caching piece - a 
client could set it up to spare a server, and we could eventually link 
to WFS-T to expire the cache when there's been an update in a certain area.

If we get a spec going I imagine we could pretty easily get it adopted 
by OGC, in the same manner as GeoRSS.  We can certainly incubate at 
osgeo/eogeo.

Chris

Steven M. Ottens wrote:
> Hi Jody, all
> 
> I'm holding a technical session of 30min called 'Creating a high
> performance webmapping site
> <http://www.foss4g2006.org/contributionDisplay.py?contribId=127&confId=1>'
> (nr 127) where I will discuss the implementation of tiling in the edugis
> project, using mapbuilder, mapserver and apache's mod_cache. I hope this
> will be a starting point for a WMS-C/tiling discussion. I don't know of
> any other conferences about tiling, but I would be interested to visit them.
> I hope to visit the ka-map workshop who also has tiling. But I can
> imagine the ka-map people like to discuss ka-map and not wms-c/tiling.
> Also a discussion around beers is a very good idea ;).
> I haven't yet done a proper performance test with the new setup and I'm
> not too hopeful that I can manage that before lausanne, but we'll see.
> 
> Regards,
> Steven
> 
> ps. I accidently send the same mail from the wrong account, sorry if you
> receive it twice
> 
> Jody Garnett wrote:
>> Shorter, Cameron wrote:
>>> Jody,
>>> With regards to tiling for the client, Steven Ottens is currently
>>> incorporating tiling into Mapbuilder and has been coordinating to some
>>> extent with OpenLayers and Kamap.  They have an email list set up -
>>> can't
>>> remember what it is off the top of my head.
>>>   
>> Indeed discussion about this email list, and the associated link was
>> the motivation for the idea. Their
>> wiki page seems to revole around questions not decisions.  Justin
>> wants an easy service to implement
>> for his GeoServer presentation, and setting up a uDig rendering
>> workflow to make use of it is dead
>> easy.  The interesting question is where to find the the description
>> of the tile breakdowns so requests
>> can be made in a manner that may be cached.
>>
>> So I would rather implement something; and then talk with code and a
>> profiler, rather then wade through
>> performance tradeoffs via email.  If Mr. Ottens has made some
>> decisions so much the better, if not we can
>> make some and compare notes & performance.
>>
>> I did think it may be fun to have something to talk shop over at
>> FOSS4G and Justin and I have managed to work
>> with the other projects mentioned (yet). It would be interesting to
>> see if building a standards in code is a good
>> way to start things off.
>>
>> Cheers,
>> Jody
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: discuss-unsubscribe at mail.osgeo.org
>> For additional commands, e-mail: discuss-help at mail.osgeo.org
>>
> 
> 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: discuss-unsubscribe at mail.osgeo.org
> For additional commands, e-mail: discuss-help at mail.osgeo.org
> 
> 
> !DSPAM:1003,44e493d4170071429667743!
> 

-- 
Chris Holmes
The Open Planning Project
http://topp.openplans.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cholmes.vcf
Type: text/x-vcard
Size: 269 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/discuss/attachments/20060819/4b5daaad/attachment.vcf>