[Imports] Best practices for address imports

Frederik Ramm frederik at remote.org
Sun Nov 24 19:09:19 UTC 2013


On 24.11.2013 19:18, Carol Kraemer wrote:
> Can we get definitions for "manual community import,"

A community import is one where the import is not done by one central
entity but small parcels of the data-to-be-imported are made available
for individual community members to tackle as they see fit, in their own
time, in areas they are familiar with, and where they will be expected
to check imported data, comparing it to their own knowledge of the area
and other suitable data sources they are familiar with from their prior
mapping practice.

A community import doesn't have a fixed timeline and relies on the
enthusiasm of individuals to get it done, and because familiarity with
the area is required, even the most enthusiastic individual will not be
able to do everything by themselves.

Any mapper can be a "community importer" but you would normally be
expected to have prior OSM experience to participate. Someone who has
just joined OSM to participate in a community import will face a very
steep learning curve and unless they are very good & diligent this might
defeat the purpose of a community import.

In a community import, the individual who imports a parcel of data takes
(a certain degree of) responsibility for the uploaded content. This
means that if challenged later about the content of the import, the
uploader cannot say:

"Don't ask me, I just grabbed this parcel of data that Fred hat provided
on his download page and loaded it up in JOSM!"

and neither

"Don't ask me, this is just the county's building footprints as they
were published, why would I know about that peculiar cluster of
buildings there?"

Because the data has been reviewed and checked by the uploader before
uploading, they will usually have noticed any peculiarities, and fixed
problems before they hit "upload".

As a consequence, to achieve the same throughput (in terms of time
needed to import a certain number of objects), a community import will
have to enlist much more participants than a scripted import by a
central entity, but will also make sure that many more eyeballs are
involved to make sure the data isn't rubbish.

A comunity import isn't necessarily good but it has a better chance to
be good than an automated import.

Both types of imports require detailed prior discussion and consultation.


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

More information about the Imports mailing list