[Imports] Florida Landuse Import

[Kevin Kenny]

On Mon, Feb 28, 2022 at 12:19 PM Mike N <niceman at att.net> wrote:

> On 2/28/2022 12:03 PM, Brian May wrote:
> > For US related import issues is the idea of using imports-us mailing
> > list outdated?
>
>    I don't believe that it is, although many prefer to kick around ideas
> on Slack first.   Imports are so tricky to get right that they must be
> exposed to the maximum number of people, including all mailing lists,
> Slack and Discord.  And though there is no formal wait time requirement
> before starting, 2 full weeks seems to be a minimum.
>
>
Imports need lots of eyeballs. I've been known to cut corners slightly on a
small, low-conflict project, but only slightly - usually to do a tiny
fraction of the import in very few changesets so that people have something
more concrete to comment on, being prepared to revert or revise as needed.

I like discussing them on Slack because people can collaborate in
quasi-real-time, but then I make it a point to memorialize the outcome of
the discussion using the mailing lists and Wiki.

I've been calling my current project of minor civil division boundaries in
New York an audit and revision, rather than an import, because it's all
semi-manual patchwork.  All the boundaries were there in the TIGER import.
Except that about a third of the townships were inexplicably missing - I
don't know what happened to the Adirondacks, the Alleganies and the Niagara
Frontier, but the townships just weren't in the TIGER import.  A lot of
them had topological problems, and some were way off - by a half-mile in
one case.  But none of the work is mindless importing, it's lots of manual
'copy-and-paste from a shapefile; glue the corners together, make sure that
mappers haven't modified the data since import, verify the metadata,
replace geometry, maybe introduce temporary patchwork to preserve topology
at the margins, upload'. By the time I've done all that, it hardly
qualifies as a mechanical edit, it's at most mechanically aided in that I
didn't need to trace all the nodes myself.

Even at that, I gave a 'heads up' warning on imports, talk-US, and Slack,
and waited until I at least got the nod from a few prominent mappers (the
roster of commenters included OSM-US board members and DWG members).

The job is just municipality boundaries, not anything nearly as complex as
landcover, and I still estimate that it'll take me a couple more months to
finish. I don't want even to think about importing landcover to what I'd
consider an acceptable standard of data quality.


-- 
73 de ke9tv/2, Kevin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/imports/attachments/20220228/24deb349/attachment.htm>