No-one has solved conflation in OSM, so there is no mechanism for re-merging branches (I suspect that it's fundamentally very hard for geo data).
If there were usable common conflation techniques there would be many fewer problems with imports. A more viable approach is a post-processed version of OSM data. A couple of the firms doing routing apps do various QA routines before pushing out new releases. The problem is that it's quite easy to do decent post-processing clean-up on data for specific domains (roads, boundaries, retail etc), but one gets back to the conflation problem it it's necessary to merge these together again. Jerry On 9 January 2017 at 12:40, Jez Nicholson <[email protected]> wrote: > has there ever been discussion about code branching in OSM? In git terms, > we are all making changes direct to master. I'm wondering whether small > changes could be automatically approved and large changes would require > peer review first. > > I know that this is a huge change to the base infrastructure, so maybe a > bot that auto-challenges large/sweeping changes? > > ....or is this like in QI when the siren goes off because someone proposes > something that has is a commonly held fallacy? or been discussed ad nauseam > > - Jez > > On Mon, 9 Jan 2017 at 12:29 ael <[email protected]> wrote: > >> On Mon, Jan 09, 2017 at 11:53:51AM +0000, SK53 wrote: >> > Somehow I have been oblivious to the fact that large numbers of polygons >> > tagged natural=heath have been added over the past few months to OSM. >> >> >> I too have encountered at least one armchair mapper who (in my view) >> incorrectly tagged large areas of Bodmin moor with this nonsense. >> I changed this rubbish in the areas that I have directly surveyed, but >> did not think that I should override another mapper who clearly had >> spent a lot of effort in adjacent areas without consultation. >> >> I first noticed this problem several months ago: I cannot remember >> whether I tried to contact the mapper to ask what was happening. >> >> I have had other instances of armchair mappers adding what I regard as >> very dubious landuse tags to areas that I have extensively surveyed. >> When I contacted one of the main offenders; I didn't get a very helpful >> response. >> >> Anyway, I suspect that this is a problem over large areas. >> >> ael >> >> >> _______________________________________________ >> Talk-GB mailing list >> [email protected] >> https://lists.openstreetmap.org/listinfo/talk-gb >> >
_______________________________________________ Talk-GB mailing list [email protected] https://lists.openstreetmap.org/listinfo/talk-gb
