[OSM-talk] Zero tolerance on imports
danielsabo at gmail.com
Sun Feb 20 09:44:42 GMT 2011
On Feb 20, 2011, at 1:03 AM, Serge Wroclawski wrote:
> On Sat, Feb 19, 2011 at 10:40 PM, Daniel Sabo <danielsabo at gmail.com> wrote:
>> On Feb 19, 2011, at 4:03 PM, Richard Fairhurst wrote:
>>> From what I can tell (talk-ca postings etc.) 'sammuell' is a fairly inexperienced OSMer who presumably thinks "this is how things are done". It isn't. How do we stop this impression taking hold? How do we explain that imports are _not_ welcome except as a last resort, and if you do them, you _must_ follow a very, very rigorous set of guidelines?
>> Maybe you don't like it, but you are not the entire OSM community. Yes, in this case someone overwritten what I presume was good surveyed data with an import was stupid. But in general the fact that data was gathered by a government surveyor with tools an order of magnitude more accurate than ours does not, IMHO, make it less "worthy" of being in OSM.
> David, how would you propose to measure consensus from the OSM community?
I assume you were replying to me? At present I think it's pretty much impossible: e.g. the license debate :). We just don't have enough OSMers on one communication channel to make real policy about these things, so I expect were stuck with the Wikipedia model right now where the few people with enough technical knowledge to revert things try to police everyone else.
>> From my limited experience working with a bunch of conflicting government shapefiles is a HUGE pain in the ass. Doing quality imports where we correct conflicts based on actual research (on the ground or otherwise) IS adding value.
> The process of reconciliation is painful and difficult, and,
> unfortunately, rarely is the importer the one making the corrections,
> especially in large scale imports. It's not reasonable for someone to
> survey all the airstrips in the world, but neither is it reasonable
> for a single person to verify all the streets in a city.
> I say this both as a previous "bad importer" and as someone who fixes
> errors on the map.
> Since you feel strongly about imports, how would you go about
> preserving their value while reducing the kinds of issues brought up
> on the list (conflicts, removal of ground surveyed data, etc.)
Really it should be the importer making those fixes if at all possible, so personally I would be much faster to revert things. If we could email back someone their changesets as an OSM file and say "it was nice of you to try importing this but could you not erase that lake over on the left?" I think we could even do it without chasing too many would be importers away.
A major limitation is also that there's really no way to collaborate on something like this without dumping it into the main map. There have been previous talks about making some repository were people could put OSM formatted data they want imported and then let other people gradually copy it into the map but I don't think it's ever gotten off the ground.
Which I think is basically why the airports import didn't get reverted. There was bad data but also a lot of good data and no better option for separating the two than just leaving it all in the map.
More information about the talk