Re: [OSM-talk] Forests are mappable - was: Re: OTG rule, borders & mountains existing | Re: Crimea situation - on the ground
The problem? Large areas of blank map that, when viewed zoomed out, look to be tree covered areas. Result: Initial mappers tag the large areas as tree covered, ignoring details such as lakes, tree cuttings etc. Some time later details of lakes, tree cuts are added. this may be some years later. This is simple evolution of the map - more detail as time goes by. there is no bad intention with any of the mappers, just trying to give some impression of what is there. Problems arise when the source used it not good enough to really say what is there, a swamp may look like a patch of grass, or heath ... if you simply go by imagery then I think this is bad practice. On 13/2/20 8:13 am, Pierre Béland via talk wrote: Hi Mateusz The link below shows north of Canada areas, where the wood landcover correspond in general to Canvec imports. The blank areas are mostly not mapped yet except some lakes and infrastructures. https://www.openstreetmap.org/#map=5/55.740/-79.804 But for Labrador, the contributors have made the choice to import Canvec excluding the wood landcover. If someone wants to test how easy it is to add the wood landcover, there is quite some work to do there creating multipolygons with inner roles for lakes, cuts for Power lines, etc. https://www.openstreetmap.org/#map=9/53.4595/-63.9679 Pierre Feb 12 r 2020 13 h 30 min 57 s UTC−5, Mateusz Konieczny wrote : Hard to say more or verify without getting specific location but I cleaned up some places mangled by badly done forest imports - for example border area that was hit with multiple low quality imports. But it sounds exactly like https://www.openstreetmap.org/user/Abbe98/diary/28368 that was solved by splitting unreasonably large relation in parts (by deleting it and remapping) ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Forests are mappable - was: Re: OTG rule, borders & mountains existing | Re: Crimea situation - on the ground
Hi Mateusz The link below shows north of Canada areas, where the wood landcover correspond in general to Canvec imports. The blank areas are mostly not mapped yet except some lakes and infrastructures.https://www.openstreetmap.org/#map=5/55.740/-79.804 But for Labrador, the contributors have made the choice to import Canvec excluding the wood landcover. If someone wants to test how easy it is to add the wood landcover, there is quite some work to do there creating multipolygons with inner roles for lakes, cuts for Power lines, etc. https://www.openstreetmap.org/#map=9/53.4595/-63.9679 Pierre Feb 12 r 2020 13 h 30 min 57 s UTC−5, Mateusz Konieczny wrote : Hard to say more or verify without getting specific location but I cleaned up some places mangled by badly done forest imports - for example border area that was hit with multiple low quality imports. But it sounds exactly like https://www.openstreetmap.org/user/Abbe98/diary/28368 that was solved by splitting unreasonably large relation in parts (by deleting it and remapping) ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
Re: [OSM-talk] Forests are mappable - was: Re: OTG rule, borders & mountains existing | Re: Crimea situation - on the ground
While I second Mateusz, the obvious solution for data users who may want to get rid of them in OSM is to filter them out. Yves ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
[OSM-talk] Forests are mappable - was: Re: OTG rule, borders & mountains existing | Re: Crimea situation - on the ground
Feb 12, 2020, 01:54 by pierz...@yahoo.fr: > > > > pierz...@yahoo.fr > > > If we could keep the wood landcover outside of OSM, it would greatly > > > simplify mapping of such areas and dramatically reduce the Mulipolygons > > > problems where huge multipolygons are created with inner for lakes and > > > all the problems related to this. > > On Feb 11 18 h 49 min 26 s UTC−5, Mateusz Konieczny via talk wrote: > > ??? just do not create unreasonably large multipolygons (or split > >existing, > > possibly undo import if it makes area uneditable and do it right). > > Your answer seems to be that it is possible to map appropriately with the > current rules. Or maybe not, but anyway, let simply ignore these areas, not > find appropriate solution to add these areas to OSM. For north of Canada > alone, > the superficy is closed to the size of Europe> . > I am pretty sure that it is possible to map forests in OSM. Can you link area where forests are somehow not mappable and cause you to propose deleting tree coverage data from OSM? You claimed that we should consider deleting tree coverage data from OSM and create a separate project of it (what is mostly offtopic in this thread). I responded that forests are mappable in OSM. And poor data quality in an unspecific location is not a good reason to do something like that. Why proposing solution for the problem would be ignoring it? Hard to say more or verify without getting specific location but I cleaned up some places mangled by badly done forest imports - for example border area that was hit with multiple low quality imports. But it sounds exactly like https://www.openstreetmap.org/user/Abbe98/diary/28368 that was solved by splitting unreasonably large relation in parts (by deleting it and remapping) And yes Canada is large and we may never finish mapping forests. It does not mean that we should delete all forest data. We will also never finish mapping shops, opening hours and many other things. ___ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk
