On Thu, Mar 3, 2016 at 11:02 AM, Steve Friedl <st...@unixwiz.net> wrote:
> Hi all, > > I’ve been updating all the cities in Orange County California to have fully > segmented relationalized boundaries, such that cities sharing a common > border share a single way in each of their relations; this eliminates > overlapping ways. It’s been very tedious but it's really getting cleaned > up. > > So a few questions: > > First: The individual relations – city, county, national forest, etc. – all > have full information tags about the entity, but how should the way members > themselves be tagged? > > I’ve seen some OSM notes that say the ways can be tagged too, but I’m not > sure how. There’s no common name, and there may not even be the same admin > level (a city boundary on the border of a county would have two admin > levels > for the given way. > > My experience is that if the relation is fully tagged, and one of the ways > is tagged with the same info, we see duplicate city names along the border, > as the renderer takes the name from both the relation and the way. > > I am not sure I see any value *other than* tagging it as a boundary, with > no > other information. But I’d really like to do this right. > The Ways can actually be without tags as the information is fully described in the Relations. Depending on the situation, it may make sense to utilize existing map features such as roads or rivers in the administrative boundary relations. > Second: I’m not sure how to handle quasi-enclaves. Orange county is made > of > many cities, and a few cities contain some small *county* regions - think > of > them as donut holes. I don't know how to handle this. > > The Orange County relation is tagged with > > boundary = administrative > border_type = county > admin_level = 6 > > The city of Westminster relation (fully inside Orange County) is tagged > with > > boundary = administrative > border_type = city > admin_level = 8 > > Within Westminster is a "donut hole" , and the Westminster relation has it > as a role=inner. > > Question: should that same donut hole be tagged role=outer in the Orange > County relation? > > It just doesn't feel right to have a role=outer fully within another > role=outer, but that's the only way I can think of to handle this. > > Or do we just assume that because this donut hole has been excluded from > the > City of Westminster, it's automatically part of the next outer item (in > this > case, Orange County). The renderer doesn't identify any parts of the > donut-hole boundary. > > The hole in question: > http://www.openstreetmap.org/#map=17/33.73488/-117.98422&layers=N it's > just > to the west of Star View Elementary School There are many of these in > Orange County. > I think you are misunderstanding the administrative boundary situation. All areas within the city limits of Westminster are also contained in Orange County. In other words, when you enter Westminster, you are not leaving Orange County. Similarly, all areas of within Westminster are also contained within the state of California. You are not leaving California when you enter Orange County and you are also not leaving California when you enter Westminster. When you leave Westminster, you are entering an area that is in Orange County. If that area is not contained within another city or town, then it is said to be in an unincorporated area of Orange County. So you would be traveling from an incorporated area of Orange County to an unincorporated area of Orange County, but all still within Orange County. The holes in the Westminster boundary do not pertain to the Orange County boundary and they shouldn't be part of the Orange County relation. Does that make sense? Peter
_______________________________________________ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
