On Nov 4, 2010, at 4:14 PM, Markus Innerebner wrote: > >> If you are using geoserver 2.1 you can use a function/stored procedure to >> return "rows" as your source of data for the feature. Surely this would >> survive the multi-user scenario? This is a very nice new feature called SQL >> views that makes it so your layers do not have to be plain tables anymore. > Yes I do. I have to take a look on it. As far as I understand the idea is to > add in the table an additional column, specifying the "owner" of that row and > then to create a view selecting only that specific rows. Correct?
Yes, even better than separate tables. You don't even need a view at this point - just a filter would do it. You could do a view of course, but I don't see any need for it. >> Is there no way to optimize your computation based on the viewport? Because >> if you can, then the performance should be better for the WMS queries and >> you could force them to be zoomed in a bit before running the query if it's >> slow at a high level. > What do you mean with that optimization? The computation is already optimized > a lot (I worked now on this since one year). The expensive part is the the > large number of DB access for reading and writing. > > if you are interested on what I am doing, take a look here: > www.isochrones.inf.unibz.it Interesting. I know nothing about isochrones, but I assuming you do the calculations from the point radiating outward on all possible paths (and then verify the reverse holds true), then I guess using a bounding box wouldn't really help you any. My suggestion was going to be to make sure you limit the amount of data you need to deal with based on the area covered by the viewport, but it's not applicable here. charles_______________________________________________ Users mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/openlayers-users
