Javier de la Torre wrote:
Hi Chris,
Apologies if we're a bit willy with the download now button. We'll never put an alpha there, but solid betas and RC's go there, as we want people testing the latest releases, to give us feedback. If you have to go stable, then download from the stable link. It's very helpful that you found the postgres error (though I wish I could do something about it and/or reproduce the error, since it closes fine on our installs).
I understand. Fairly enough 
I wrote you a private message regarding this issue and offering my help over chat... I can also give you access to the server where I tried and where the issue happened, maybe that can help you.
Cool, let's do it on monday or tuesday.
It would be great to have a perfect mapping of Google maps projections
at different zoom levels with EPSGs in Geoserver so that they overlay
perfectly (but I don't know if Google also change the projection
depending on the latitude/longitude where you are...)
I think few people in Google even know, at least not the ones I talked to. I can try to get them to find it out, I'd love a script that matches everything up exactly.
I found a very interesting post at (http://cfis.savagexi.com/articles/ 2006/05/03/google-maps-deconstructed) explaining how Google Maps tiling works and how they do with projections. The most curious, and I believe also clever, is that they actually create new projections on every zoom level using pixels as the unit system... this is really good for presentation in computer screens and save them from a lot of headaches (different pixels are differently represented in different screens), but probably this is not that usual in the geography world. Maybe this is the way future maps will work :?
Very interesting. Brent recently made a kml_reflector, which is quite nice, see: http://docs.codehaus.org/display/GEOSDOC/Google+Earth#GoogleEarth-KMLReflector (the code is not released yet, will be soon, if you want a beta we can get one for you).
We could do a similar GMaps reflector, but that does the right switches on the projections for you. If anyone figures out more than the 41001 -> 4326 switch that'd be great, do let us know.
Other interesting thing that we could specify is the matching of Google Maps zoom to the OGC scales idea... I spent half an hour trying to find the Min-MaxScaleDenominator that would allow me to change the presented layer from google maps Zoom 6 to 7, but I would like to change more often, I am discovering how important is to represent different things at different scales!
Do let us know what scales you come up with. Matching with google maps scales would be great. I know that their zoom is based on power of two, but I don't know how that translates in to ogc min/max scale stuff.
We're also going to be working on an online SLD editor, and I think one of the keys is that it will be based on scale. You'll write a new rule for each zoom level, as that's how web map works. In the future we could try to adapt udig's sld editor. Theirs is geared towards desktop, where you can zoom in and control everything right there. With web mapping it's much better to have the pre-set zoom levels, where things look good. If you look at http://sigma.openplans.org you'll see how we present things differently based on the zoom.
Just add in-database SLD filter processing to Geoserver and you have a perfect match to represent big datasets on Google Maps. Is incredible how many hacks are appearing in the Google Maps user community to deal with things that trough OGC standards would be much easier... (or maybe not...)
Yeah, I'm hoping GeoServer can be a tool to make it much easier to serve maps. And indeed not just on gmaps, but also google earth, and even yahoo maps and microsoft ve, ect. So that people will be able to just set up GeoServer and view it in lots of different ways. But to do it right we need to make it easier to do on just _one_ of the ways.
Chris
Javier.
--
Chris Holmes
The Open Planning Project
http://topp.openplans.org
