[OSM-talk] [OSM-legal-talk] In what direction should OSM go?

[Anthony]

On Wed, Sep 29, 2010 at 12:09 PM, Ævar Arnfjörð Bjarmason
<avarab at gmail.com> wrote:
> I *don't* mean that they could do it *automatically*. Distributed
> version control systems don't do that either, you always need a human
> to look at the result to see if it's sane.

The problem with imports that have to be kept up to date manually is
that they rarely *are* kept up to date manually.

In any case, I think TIGER was a good import (despite the arguments to
the contrary), but that's because the TIGER import was made at a
completely different stage of OSM development.  If you wanted to
import a road network today, it'd be a whole different story.  The
import itself would be a merge.

Imports today basically fall into the category of being a merge with
existing data, or being an import of data which mappers aren't likely
to be interested in maintaining (and therefore shouldn't be imported
at all).  In the former case, either you've got the sync tools already
in place at the time you make the import, or the import is manual from
the start.

If an import is done manually, I don't even think it really counts as
an import for the purposes of this discussion.  It should be treated
like any other editing.  Don't violate copyrights, don't import things
which aren't verifiable, don't step on other people's tagging schemas,
but other than that do as you please.