[Talk-us] Clear Creek County in Colorado has a broken county boundary

[stevea]

Taylor Smock wrote:

> https://data.colorado.gov/Transportation/Counties-in-Colorado/67vn-ijga The `About` tab /claims/ that the data is Public Domain, but
> I would double check, just in case. (Some people I've talked to in county governments thought public domain meant the data was
> available for the public, /not/ that the copyright status was public domain, so I've learned to double check). 

Thank you for that link, Taylor.  While I am not an attorney, I am confident enough that when a state government adds metadata of "License:  Public Domain" I can discern these data are ODbL-compatible.  So I essentially re-drew OSM's previously-broken (significant chunks missing) polygon for Clear Creek County based on these data (and using the missing segment whole and unchanged).  This also required some massage of the data for part of the boundary shared with Gilpin County.  I attributed the link above in the changeset's source comment.

Especially in the area of James Peak (node/358916927), there are a number of boundaries here (James Peak Wilderness Area, James Peak Protection Area, Grand County, Continental Divide) which remain quite messy, with significant errors that look to be around 20 to 30 meters.  Indeed, the Continental Divide itself (way/385331055) is tagged FIXME=improve precision.

While I don't often publicly complain about OSM's data, thanks to this dataset submission to talk-us, I have noticed that a fair amount of county boundary data in Colorado, while extant in OSM, have serious drift and accuracy errors — hundreds of meters or even kilometers.  As there exist ODbL-compatible data which are clearly superior to OSM's current dataset, I recommend a dedicated editor (or team) properly import these, conflating with existing data where necessary (there appear to be many shared ways for edges of forest and wilderness boundaries).  This necessarily will be careful work, but OSM will be much better for it.

I don't know how extensive these sorts of data errors are in other states.  I stumbled down this rabbit hole while using http://layers.openstreetmap.fr/?zoom=5&lat=40.2&lon=-97.5&layers=0B000FFFFFFFTFTTTTTTTFFFFFF to quality-check county-boundary tagging.  There are still some oddities (several counties in California plus a score or so sprinkled around the lower 48) where I still do not quite understand why they remain in apparent error, despite my "healing" some previously-broken boundaries.  It could be I don't fully understand this renderer's tiling schedule.  But while this rendering has improved some of those broken counties due to my improvements, some of them stubbornly refuse to better render, despite seeming correct (and correctly-tagged) polygons.  Hm.

I'm "slowly watching" this (county boundaries which appear to have decayed to brokenness, then repairing them), but I wanted to both share my activities more widely with the US OSM community, as well as thank Taylor for his quick reply to my request for state-issued county boundary data:  I appreciate the pointer.

SteveA
California