On 20.07.2012 18:49, Simon Poole wrote:
> Very simple: the above changesets were black listed because they were
> copy & paste remapping. The nasty side effect of them is that the hid
> any problematic data that they were supposed to fix. With the net result
> that such areas were not remapped prior to the bot processing Switzerland.
Interesting, thanks for the answer. Where does one learn about such things as black listed change sets?
Now I see the problem and I can understand why you are not too happy about the actions of mdk. But let's keep in mind that mdk made his changes from the end of February to March 2012, which was so to speak in the very last minute before the change process that was expected for April. The remapping has been going on a long time before that point and should have pretty much been finished by then.
Instead of focussing on mdk's unfortunate change sets, I think it would be more useful to look at the change set of OSMF Redaction Account, if for no other reason than that they reflect what truly went missing, they seem to be more compact and tend to have smaller bounding boxes.
Alas, even if their bounding boxes might be smaller, they nevertheless cover a considerable area and I still abhor the idea searching all these bounding boxes for missing nodes, ways and relations. Yes, some of the deletions produce visible errors such as unconnected street segments, others leave visible gaps in a street, noticeable for those who know the area. But on a finer level such as tags or routes, the effort of finding out that something went missing at all goes quickly beyond anything that's reasonable.
I understand that the exact locations of nodes and ways and the values of the keys need to be hidden away. But I wonder if it wouldn't be possible to give at least a vague (graphical) hint to the approximate location of deleted ways and relations and a list of deleted keys (without values) in the form of a set of map overlay tiles.
If I remember correctly, one of the arguments why the CCBYSA wasn't working for OSM was that the OSM database would be a collection of "facts" that couldn't be protected. Now you could of course make the same kind of argument about Wikipedia (also under CCBYSA). In neither place you should "creatively" make stuff up. And just as WP uses combinations of words to describe these facts as efficiently and elegantly as possible, we in OSM use combinations of nodes, ways, relations and tags to describe facts. In both cases the "creativity" lies in the efficient and elegant use of our means to describe the underlying "facts". Which is to say that even if the exact description of the fact should or should not be protected by CCBYSA, it should still be possible to give us an approximate pointer to the underlying "fact" that was described and whose description needs to be redone.
Thorsten