El Miércoles, 22 de Abril de 2009, Frederik Ramm escribió: > A proper monitoring function has to use the changeset bbox as an index > only and then check whether the changeset *really* contains something in > the area of interest to the subscriber. This is more expensive in terms > of CPU power and I'm not sure if we want to burden the API with it. It > could be done externally if changesets are distributed as OSM diffs are > today.
Maybe we could calculate the convex hull of the changeset and use that with some postgis magic. It should be quicker than comparing all the elements in the changeset to the bbox, and would not produce so many false positives. (Yes, yes, I know, patches are welcome) -- ---------------------------------- Iván Sánchez Ortega <i...@sanchezortega.es> Un ordenador no es un televisor ni un microondas, es una herramienta compleja.
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk