On 30 July 2010 16:26, Ross Scanlon <i...@4x4falcon.com> wrote: > > Total time 6 minutes > > Hundreds of hours, yeah right. > > The program I've been talking about uses osm2pgsql and mapnik so I'm well > aware of them. > > If your smart you could probably add the emergency data without having to > totally rerun osm2pgsql. >
I think you are missing the point that Brian was making. If the data has to change dramatically, it means one or two man weeks to do the change at work for our engine at work. The change in itself is simple, but that means testing if none were missing, that everything is running fine, regenerating a database earlier than expected which in itself takes a significant amount of time. Testing is something quite important and we have tons of test to make sure that the database is working as expected before we release to production. While I follow this mailing list, I am pretty sure that many people working in the OSM ecosystem is not following the change that fast. It means that every one doing an app needs to do some significant work to make sure that those new settings are taken into consideration. In the meantime, you have effectively broken many applications, which is something which is very bad. That would go against any kind of rationalisation that you want to do since you would break things quite a bit. It is not the image that we want to be showing to the world. Ultimately, IF it was the right solution, John solution of using double tags might be a solution to lessen the impact. However, I don't like those tags, as they don't really make sense in many cases. Before rushing, we should really evaluate a bit more if they make sense. Emilie Laffray
_______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
