Hi, On 12/06/2011 12:11 PM, 80n wrote:
We've been taking OSM content and applying it to a separate database which also has direct contributions. This is the opposite direction to what you are contemplating, but the same principles apply.
Well, only if one leaves out a lot of important detail! If one were to operate a separate "leaf" server and feed the collected edits back into OSM in the way you describe, one would at least encounter three problems:
1. The person resolving the conflicts is not the person who has made (any of) the conflicting edits. Who will do the work and will they have the necessary information to do the right thing?
2. All uploads to OSM would appear under one user ID. Therefore OSMers could not identify contributors shielded behind the leaf server which is undesirable (e.g. one could not send them messages, or block them individually if they are found to violate copyright or so). This could be solved by making the leaf server use OAuth against OSM's server but that would require some work and make your approach "import to OSM using your favourite load tool" a bit more difficiult.
3. Extra work would be required to modify the local database with the IDs assigned by the OSM server during the upload process, or else you have each new object twice - one locally created and one arrived via a diff load from OSM later.
In practice, the risk of edit collisions is very small, and when they do happen it is non-destructive and easy to resolve with no special skills.
In my OSM work, edit conflicts happen frequently. I guess the risk of edit conflicts increases with the amount of users.
You might find there is political resistance to this idea from some OSM people, but nobody can stop you from doing it.
Nobody can stop you from running your own database in the way you, 80n, do; but if someone were to run this the other way round and upload collcted edits to OSM under one and the same user account then they can, and likely would, be stopped.
Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09" E008°23'33" _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
