[OSM-dev] Pl at ntNet contributing to OpenStreetMap

[Greg Troxel]

My response here is to the part of your question that should have been
on talk@; it's not so much about mechanics and the API but "should you"?


I am a bit skeptical of the wisdom of this endeavor.  I recommend that
you think hard about the relationship between:

  your subcommunity wanting to publish data of particular interest to
  them

  the costs and benefits to the broader OSM community

So I would ask: do you think the data to be added would be viewed as
generally useful to others?  Granted, there are many things in OSM where
the level of detail of one thing is far more than most care about
(e.g. # wires and substation names on power cables), but it is typically
not a significant amount of data.

Adding individual plants (other than trees) could easily  cross into an
overwhelming amount of data.

And, one has to ask: what is the persistence time of these objects in
the real world, and how will the database be kept up to date.  We don't
map "there is this minute a deer standing here" at one extreme, and we
do map massive concrete buildsing that take years to build and are
stable for decades.  Tree are ok by this rule, and individual weeds in
meadow are not.

Separately from the scale issues, anyone contemplating going down this
road should pause and first become an experience OSM mapper who is
familiar with editing and the tagging schemes, sufficiently advanced so
that they can help newbies.  It doesn't come across that you've done this.

Generally, OSM aims to map ~everything, except that this really means
things that a signficant number of people care about, which is partly
about how large they are.  Individual plans are a bit iffy.

There is established tree tagging in OSM, so adding objects with tags
that follow that established scheme is fine.  But you are talking about
a lot of tags that are far more detailed that what is normal practice,
and the normal practice is to project the richer data to the OSM
representation.

There are also questions of spatial accuracy.  OSM doesn't have precise
standards, but it is obvious that using non-GNSS geolocation (that
doesn't achieve comparable accuracy or better) is not ok, and as smaller
features are mapped the required accuracy improves.


 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20220809/b435b1be/attachment.sig>