[Imports] Florida Landuse Import

[Kendall Mullenhour]

Afternoon all,

After some discussion on Slack and Discord and seeing the import digests
flood my emails for the past four days, I think the data is pretty good,
just there was some issues with how fast it was being important, and from
what I've heard on the Discord chat the "public comment period" for imports
(as we call it over here at the state DOT when planning large road
projects) is usually a minimum of 14 days/2 weeks, when this import started
going 4 days into that 2 week period.

Yesterday on the Discord channel there was something brought up about the
natural=wood/landuse=forest mismatch on some features. From what I have
heard, landuse=forest is supposed to align to landuse (like a pine tree
farm for example) instead of natural features (so natural=wood would go
across different landuses). "Coniferous Plantations" is a good use for a
landuse=forest + leaf_type=needleleaved. Another thing brought up that I
found is that there were duplicate nodes found by another OSM user b_jazz.
Not sure how the duplicate nodes would pop up but it is something worth
mentioning since importing large blocks of landuse at a time would make it
very hard to check everything for errors.

If original importer is indeed going to abandon the import I could probably
pick it back up with some major changes:
-Split the entire project up into 67 parts (1 part per county), and have
each part going on the tasking manager (like it is with my other countywide
projects, import and non-import). TM cell sizes would probably be the same
size as my Bibb County Waterways manual mapping project (OSMUS task #243),
but folks would be able to split a cell into four in areas with complicated
landuse (like in the cities).
-Feedback from the previous import attempt would be worked into the newly
resurrected project (such as the landuse=forest/natural=wood scenario
brought up on Discord #imports channel)
-2 week public comment period (solicited on mailing lists, Discord, and
Slack) prior to two test sections (1 rural, 1 urban) being uploaded Both
test sections would be in the Panhandle and placed in a way that would
allow testing of tie-ins to existing landuse or other landcover features.
Raw data and processed data (before upload) would be available on wiki page
for download so that others can chime in on how to avoid errors and other
areas of improvement.
-There will be another 1-2 week public comment period solicited on the same
communication channels after both test sections are live. Any issues found
then would be corrected before moving to a larger test section.
-After a 3rd comment period, the full import will commence, one county at a
time, from importation to validation. Any errors caught in the validation
phase in each county would be corrected on the spot and suggestions to
implement bugfix implemented in the next phases.

Hopefully we can get something worked out on this, I think the
landuse/landcover import would be beneficial for the state of Florida and
improve map quality for this state. Please let me know on this mailing
list, or on the Discord or Slack #imports channel (my username on both is
asdfjkll).

Thanks,
-Kendall Mullenhour (username RollTideRoll_asdfjkll at
https://www.openstreetmap.org/user/RollTideRoll_asdfjkll)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/imports/attachments/20220228/dc35624a/attachment-0001.htm>