[Imports] Fwd: Importing West Virginia State Forests Boundary

Minh Nguyen minh at nguyen.cincinnati.oh.us
Fri Aug 13 16:21:43 UTC 2021


Vào lúc 01:00 2021-08-13, Frederik Ramm đã viết:
> Minh,
> 
> On 12.08.21 21:27, Minh Nguyen wrote:
>> The next time someone proposes a building import, I'll be bracing myself
>> for a debate about the lack of associatedStreet relations in an
>> unrelated address import. 🙈
> 
> I don't think you should make fun of my criticism. The two imports are
> not "unrelated", as they are being executed by the same person within a
> timeframe of less than a year.
> 
> The person has not even seen fit to participate in the discussion,
> instead letting others do "the paperwork" for him, which to me signifies
> a certain unwillingness to take responsibility for their actions.
> 
> Are those the qualities you are looking for in someone who imports data
> in the US?
> 
> The lack of relation use in that previous import might well be a sign of
> the importer being uncomfortable with relations overall. In a boundary
> import, the least you are looking for is that if a protected area
> boundary coincides with an adminisitrative boundary, this is properly
> recorded in OSM, rather than mindlessly throwing in more and more
> overlapping line geometries.

Thanks for explaining the connection that you see. Over the years, I 
have more often advised caution against an import than encouraged one, 
possibly out of the same concerns over data quality and long-term 
management that you have. However, this particular boundary proposal is 
one of the better ones I've been asked about. Attila has offered to do 
the import himself, so it's a moot point now as long as he takes into 
account any technical concerns raised in this discussion.

> Here's an example where this has been done well:
> 
> https://www.openstreetmap.org/relation/5880036 "Scotts Basin Wilderness
> Study Area", which happens to coincide with the boundary of Juab county
> for a bit, and as a result they both share
> https://www.openstreetmap.org/way/392275476.
> 
> Here on the other hand is an almost certainly bad import:
> 
> https://www.openstreetmap.org/#map=16/35.1685/-109.6939
> 
> because whoever has imported the boundary of "Petrified Forest National
> Park" or the "Navajo Nation" (I haven't checked which came first) has
> allowed both to overlap, leading OSM to mistakenly claim that a part of
> the National Park lies inside the Aboriginal Lands. I don't doubt that
> both data sets came from some official source but in my opinion it is
> the duty of the importer, as part of proper conflation work, to fix such
> problems (which may occasionally even mean: do some research!) rather
> than just dumping it into OSM for someone else to care.

There seems to be an assumption that the boundaries aren't defined in a 
way that don't overlap by a smidgen. I know very little about this 
specific example, but as someone who used to live on a property parcel 
that accidentally overlapped with two others, I find it less alarming 
than I probably should. Perhaps Attila could check if there are any 
close calls like this so we could discuss what to do about them here or 
in other forums where locals are present.

-- 
minh at nguyen.cincinnati.oh.us




More information about the Imports mailing list