On 18 October 2012 23:05, Frederik Ramm <frede...@remote.org> wrote: > In a recent message, to talk-it > (http://lists.openstreetmap.org/pipermail/talk-it/2012-September/030778.html), > Paul writes > > "We recognize that the line between an import and assisted mapping is not > currently clearly defined; however all the cases I have seen recently > clearly were on the import side of that line." > > So he calls it "assisted mapping", I called it "using a variety of legal > third-party material in the mapping process", we could also call it "a > manually verified, small-scale import". > > These things are ok and while it is not currently written, DWG does not > enforce separate accounts for them. If that is any help, we can try to sit > down together and try to clarify the line between "assisted mapping" and > "import". > > There are many reasons why we want mass imports clearly separated from > normal, human-contributed data. We got burnt by this in the license change > in Poland, where we had to spend massive amounts of time sorting between > "good" and "bad" changesets contributed by the same account.
This is off topic in this thread, but I'd like to set the record straight. Who do you refer to as "we" when you say you had to spend any time sorting those changes? The LWG and the rest of the contributors to the license change have done nothing at all to understand what data in Poland was compatible with the new license and which of it wasn't. You might have noticed (or not) that pretty much every sentence in LWG minutes referring to this data has a factual error of some sort, especially the ones quoting any numbers. Really. This was on such a scale that the day before the redaction started (Tuseday morning) I was asking people on #osm-dev including LWG members and the bot operator, if any of the decisions made by LWG to that time had in fact been taken into account. What I learnt was that the final rules to be applied by the bot were such that over 50% of the data to be redacted by the bot in Poland was in fact compatible with ODbL, while at the same over 50% of the data incompatible with ODbL would be left in the database. Completely nothing had been done to that time. (And even then there was not much will to do anything right. I was told that if I wanted to provide drop-in code to fix the basic problems, then I had about 12 hours to do so -- that was another statement that made Michael's Collins' "reasonable effort" and "due diligence" simply laughable. On that same day, the operator of the bot had literally said that it was not his task to read LWG's meeting minutes and implement those plans. A week later when it turned out that a human mistake caused too many objects to be redacted (mind you those objects have not been unredacted to this day -- a joke on the automated edits guideline where you're supposed to have the tools to undo anything you do), another community member had come to #osm-dev to ask about this and was quoted an hourly rate for programming work by one of the OSMF board members for repairing the destruction done to that contributor's work) So I'll appreciate it if you can avoid saying that you'd spent time sorting through any changesets (or have you and that work was simply not used?). I had agreed to provide the whitelists and blacklists needed for the bot to approximate what would be a license-based redaction because I felt responsible to both the Polish OSM community and the authors of the CC-By-SA data to be removed, for what the OSM project does with those people's contributions. I had quit the OSMF to avoid the responsibility for their movements. But at that time I had already spend *weeks* of programming work trying to help the OSMF destroy less by writing the equivalent of the redaction bot to go through UMP edits history. And because of that decision I have even spent a couple hours this last weekend helping the OSMF do *more* destruction to free geographic data in the OSM database, instead of doing something productive. Now coming back to the question of dedicated import accounts, I don't see they make a lot of difference. They're not a huge burden to the importer, but they neither do solve any problem that the source tagging doesn't solve better already. If you want to redact the OSM database ignoring basic facts and information that has been provided to you clearly and repeatedly by different people then not much is going to help you. Still the UMP imports usually have required days to weeks of manual work before each changeset uploaded because of the data model differences and I think you could easily put them in the "assisted import" category. Which would mean that there's more manual work in them than 3-rd party and using either a separate account, or entire-changeset tagging, would cause more false negatives than not. You could do that work in smaller changesets but you'd lose the atomicity or "bisectability" in git speak, where you'd have a map state in between the beginning of the work and the end, which is unroutable and unusable. We've put great effort in avoiding that by doing a lot of preparation for each upload and working together with other local mappers. Cheers _______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/listinfo/talk
