Hi, Sebastian Klein wrote: > Compared to that, the JOSM format is quite minimalistic. It just adds a > action=* attribute and thats it. But it reflects much better, how the > application "thinks". And it is very similar to the "standard" osm > format and thus, can be processed by osmosis.
Actually this is something that many people trip over: "oh, it looks just like a normal OSM file, let me use it as one" - and then things don't work because of negative IDs, or (even worse) they use osmosis to merge two files with negative IDs (of course the IDs are used for different objects in both), or they process the files in an application that does not understand the way JOSM records object deletions (and then are surprised that all the deleted objects show up again)... > We could use the osmChange format, but then we would need to put the > normal osm data and the osc change data in a zip-archive. I'm not sure > this would be an improvement. Probably right; if we wanted to use osc alone we would again have to do dirty tricks like adding a "create" record for each existing object or so. The more I think about it the less I like using an existing format. I think that using <osm>...</osm> has problems (noted above), and you are right in saying that <osmChange> is not better. If JOSM used a special JOSM format which other software would not recognize, then we could provide import and export functions or converters that would make sure that things are handled correctly, e.g. that if you export from JOSM to an <osm> file all objects marked as deleted will not be exported, and so on. People would not be tempted to think "oh, it's an OSM file, I can use it with the XYZ application". Bye Frederik -- Frederik Ramm ## eMail [email protected] ## N49°00'09" E008°23'33" _______________________________________________ josm-dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/josm-dev
