[Imports-us] Rural SE Ohio Sidewalks Import

[Greg Troxel]

Malcolm Meyer <mmeyer at ovrdc.org> writes:

> Ok, that all sounds doable, but the thing that you said that makes me
> pause is 'the code needed to transform the data'. I have been under
> the assumtion that I can just import something like geojson into JOSM
> or another editor, assuming all the feature properties match. Or maybe
> there is a geojson to xml converter or something. The imports will be
> in chunks, so 2,000 features at a time or something. Is this not the
> case? Is there not an 'import' method using one of the desktop
> editors? If not then could you point me in the direction of the actual
> import process. All the other communication stuff I have no problem
> with.

I said code, because usually people write scripts in python or something
to read shapefiles with properties from some traditional GIS and them
emit OSM data with tags, and the scripts encode the logic.  If you have
actually produced data in OSM format already, except that it's in
geojson instead of osm xml, but it follows the OSM tagging schema,
that's another story.

Using GIS tools instead of scripts is ok, too, but the point is that you
should explain enough so that someone else can take your raw data and
run the transformation pipeline.   If you have multiple people you'll
want this written down anyway, so I hope this is mostly just
transparency and not a lot of extra work.

The big point is that you have a dataset and OSM does too and just
"squirt X into OSM" in general risks duplicate objects; conflation is
hard.  So to succeed you need a story about that, and the process just
makes you write that down for others to read.

> Overall, considering I could probably map these in a few days directly
> into OSM, this while process seems a bit absurd. If we get another

The import guidelines attempt to deal with much larger scale things, and
we sort the world into mapping by hand and mechanical edits.

> intern I may have them do just that, though it may take them a bit
> longer. The problem is we cannot use the direct OSM iD mapping as our
> only solution due to the license. All of the data we create must be in
> the public domain - so hence creating it ourselves outside OSM then
> doing the import. However I didnt realize the import process would be
> so complicated.

I am pretty sure that you can check a box on registering that your edits
are public domain, and I am 99% sure that extracting your own work (not
subsequently edited by others) from the DB after editing and using it as
PD is ok.   A question for talk-legal.  Certainly OSM should encourage
governments to do things like this; I think most US mappers are quite
happy for government data to be PD even if round-tripped through OSM.
What most are not ok with is proprietary copies of the DB with other
data being used, and denying those improvements to the commons.

> But once I figure all this out, when I am ready to import another
> city's sidewalks, do I have to do all this stuff again, or can I just
> use this same wiki/communication/procedure, etc for the remainder of
> our cities once they are complete?

Generally, once an import passes muster, then continuing with the same
kind of source data and the same processing pipeline is ok.

In your case, it sounds like this is large scale hand mapping with some
different tools, and right on the edge.  The process and discussion have
evolved to avoid problems.   So really it's about explaining what you
are doing in enough detail that people who have seen the problems can
point them out before they happen.

So how about you first publish the data you are talking about, and start
a page that explains where it came from, and how you decided on tagging
and what standards you are following, and people can have a look?   I
think most here would like to encourage governments putting in data,
especially if there's an intent to maintain it.

You probably also should think about what happens as you update your
local data, and then it's different from what's in OSM.  That could be
because OSM is your old data and doing an updated into OSM is ok, or
because someone in OSM edited it, in which case overwriting their change
is not ok.

Also, separately from the import, you can use the OSM tools with your
own database and your own license.  But I think getting the data into
OSM is a good thing.

Keep in mind that I'm one person, and while I think my thoughts are
mainstream, I'm not speaking authoratively for the group.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 180 bytes
Desc: not available
URL: <http://lists.openstreetmap.org/pipermail/imports-us/attachments/20160813/765dc144/attachment.sig>