« Near term goals for TileCache
Spatial Reference: The Web Site »

Update on Near Term Goals for TileCache

Threshold bug: Fixed. 1.9-rc1,2 released, and a user even confirmed that it fixed his threshold problem (yay)! This will probably be the last 1.9 release, and includes significantly improved error reporting.

Refactor for drop in Layers: done, in the ‘refactor’ branch. I think I’ve maintained backwards compatibility at the ‘using a config file’ level, even, though it does need an additional option for mod_python support, so it will definitely be a 2.0. I also added drop in support for caches.

Add ArcXML layer: Done, though I have no clue how people actually use ArcXML, so this support may actually be total crap. I got it to work with the MassGIS site though. (Would love to talk to an actually ArcIMS/ArcXML user and figure out what I should be doing that I’m not.) I bet you’ve never seen ArcIMS look so fast.

Add KML Region Support: Not done, and not really too excited about it. I plan to get the 1.9 final out the door, then release the 2.0 without it, and maybe loop around for the KML region support later.

Overall, not bad.

This entry was posted on Friday, July 20th, 2007 at 2:56 pm and is filed under Locality and Space, TileCache. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

2 Responses to “Update on Near Term Goals for TileCache”

  1. Pete Says:
    July 20th, 2007 at 5:25 pm

    It would help a lot if the tiles you return at that ArcIMS link, had a cache header set to they would be cached on the client browser. My browser is refetching the tiles making it much much slower and put a lot more load on your server.

  2. crschmidt Says:
    July 21st, 2007 at 7:38 am

    Pete —

    Similar to all my TileCache setups, I treat caching as a secondary task, outside of TileCache’s domain. Because most data is updated at regular intervals, it is far easier for the server admins to configure the caching than it would be to bake that functionality into TileCache.

    During testing, such caching would inappropriate, because the tiles that were there would have been changing regularly, and I didn’t want to have to clear my browser cache every time I made a small change in the code.

    I’m still not convinced that my code is at all right — once I am, I’ll change the demo to live on the main TileCache server that MetaCarta hosts — so until that happens, I’m not adding caching headers because it makes my life more difficult, rather than for any other reason.