On 01.05.19 22:22, HWANG, MICHAEL  (MICHAEL) wrote:
> I’m looking for feedback/interest in a new open source software project
> idea.  This project is intended to address the problem of systematically
> combining private geo-spatial datasets with OSM data.  There can be
> overlap (in terms of objects) between the private geo-spatial datasets
> and with OSM and the project’s goal is to de-duplicate and to merge all
> objects together to produce a new single, consistent, more complete
> dataset.  The reason why this is needed is that organizations have
> private datasets that will take time/never be pushed into OSM.

It is important to take a moment to look at the reasons here:

* Some datasets might be theoretically suitable for inclusion in OSM,
but the person or organisation using them doesn't have the patience or
resources to commit to a proper import process.

* Some datasets might be of insufficient quality for inclusion in OSM,
but the quality might be sufficient for a particular use case.

* Some datasets might be outright wrong - e.g. a politically defined
"official" set of boundaries that needs to be used to publish maps in a
country but is useless for all other purposes.

* Some datasets might be confidential or copyrighted and therefore not
suitable for contributing to OSM.

If there was a platform that allowed people to mix such data with OSM at
on the data user side, that would be a huge relief for OSM because it
would stop people from pushing low-quality data into OSM "just to make
nice maps" or "just to have the hospitals on their Garmin maps" or so -
they could mix-in questionable data with your toolchain.

Even better if the platform were public in a way that would allow people
to mix-in data provided by others, e.g. someone could at the push of a
button choose to have built-up areas from naturalearthdata.com in their
maps or so.

Lots of technical challenges, of course. In some of the cases above
there might also be license challenges; mixing your own restaurant data
with OSM's and de-duplicating and generating a "new and better" data set
would likely, if you publicly use that, lead to it having to be
published under ODbL so the "confidential/copyrighted" use cases would
have to be carefully checked.

> Would the OSM community be receptive to this sort of project and be open
> for collaboration?  If this is not the proper forum to ask, please let
> me know where else I can go to ask.

It is hard to ask "the OSM community" for anything but I guess here's as
good as anywhere.


Frederik Ramm  ##  eMail frede...@remote.org  ##  N49°00'09" E008°23'33"

dev mailing list

Reply via email to