On Thu, Jul 7, 2011 at 1:15 PM, Andrew Harvey <andrew.harv...@gmail.com>wrote:
> > The more who contribute directly to fosm rather than OSM, the less the work > there will be for fosmers dealing with duplicated data resulting from > merges. If it becomes a big problem, I think we should be able to do manual > merges of OSM data into fosm, assuming we have the volunteers. Otherwise we > can just leave OSM data behind if no one is longer to merge it into fosm. > The probability of collisions is quite small in practice. We are able to automatically sync all OSM updates into fosm.org in near real time. Consequenly fosm.org already has more content than OSM and the gap will continue to widen. It will become a massive gulf if OSM ever has the courage to mass delete all non-ODbL licensed content, but I can't see that happening any time soon. The worst case for a collision is an edit in OSM that conflicts with an earlier edit made to the same element in the fosm database. In this case we place the OSM edit in a conflict log and preserve the fosm edit. Other kinds of conflict include the same feature being added to both OSM and fosm independently. This will result in the feature being duplicated in fosm, but it's easy to manually delete such artifacts when they are noticed, retaining whichever is the best one. My largest concern is with piecemeal replacement of non-ODbL licensed content in OSM with inferior quality tracing. This will appear as legitimate edits to the fosm sync process and will result in fosm being degraded needlessly. We've talked about mechanisms for watching areas where this might happen and for users who might be doing this. We can revert such edits in fosm and get the good stuff back providing we notice that it has happened. 80n
_______________________________________________ Talk-au mailing list Talk-au@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk-au
