Hi, Matthias Julius wrote:
One option would be to run osmosis with clipIncompleteEntities=false (the default). That would not increase the burden on your box and still allow the extracts to be merged. Of course, this would leave it up to the data consumer to deal with the incomplete ways and relations.
Yes. I somewhat fear the number of "i am getting a strange error in my program" emails that this would cause.
Merging extracts is a niche operation and not the main purpose of what I am doing. For someone who wants to mix and merge at will, it would be much better to simply divide the world in lots of squares and select from those. The expensive polygon cutting could be dropped completely. It is possible that a process for doing that can be derived from those who do regular Garmin maps. If done properly, it would even be possible to have a web interface where you can select your area of interest, and a matching extract is then merged live from pre-made squares...
In some ways this is even preferable because it would preserve the boundary of the extract. Otherewise you get a fuzzy boundary and the consuming tools have no way of knowing the original boundary.
Yes but if you start looking at this closely then the only thing you know is that the boundary must be somewhere between nodes X and Y so that doesn't get you far.
Do your polygons really overlap? I think it would be ideal if the shared the nodes.
No, that would kill the process. I need to use simplified polygons; cutting out a country with a 15k node polygon would take forever. And it is impossible to simplify the polygons and still make sure that each boundary is contained completely in the respective country file, so they have to overlap.
Bye Frederik -- Frederik Ramm ## eMail [email protected] ## N49°00'09" E008°23'33" _______________________________________________ talk mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk
