On 18/09/12 at 20:51 +0100, Lester Caine wrote: > I think that one of the problems here is that if a large block of > data is uploaded in one 'commit' it is difficult to know if it IS a > manually edit, or something that has been created automatically > off-line, and is being slipped in to bypass the bot rules.
I don't buy that argument. I think that what matters is that changesets/commits are logically split, not split just to keep them below some size limit and avoid raising eyebrows. When integrating cadastre data, contributors work on a /commune/ by /commune/ basis (a /commune/ is an administrative division, basically similar to a city). Why should contributors have to artificially split a commune into several changesets? It's much more convenient to process the whole commune at once. Alternatively, if this was software development, what should probably be done is: 1. commit the raw conversion for the vectorized cadastre, before the cleanup 2. clean up and upload modified buildings after the cleanup 3. add roads, etc. and upload That would split the integration of a commune's cadastre, into several logical commits, but I suspect that this would raise even more eyebrows due to the nature of the first commit. Or are you suggesting that the first commit should be made with a separate account, but the two following commits should be done with a normal account? That would generate many more changes, as many buildings need to be fixed. Lucas _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
