[Imports] IBGE license statement

Fernando Trebien fernando.trebien at gmail.com
Tue Jul 9 16:55:17 UTC 2013


Hmm I used a "source=IBGE" tag on the areas but not on the nodes. I've
already imported about 4000 areas and this is about half of the
complete data set.

I agree that this somewhat inefficient but IMHO this it's more
relevant for larger imports (contrary to what I'm doing, I've seen
very large data sets imported without any tag cleanup or geometry
cleanup whatsoever). I've put quite a lot of thought on the data
cleanup process and I thought it was a good idea to use the tag
because:
- previous IBGE imports have also used this tag (but they were done
years ago and I'm not sure if the community was as concerned with
attribution as it is today, I've joined the community long after the
change to ODbL)
- it seemed safer because communications with IBGE demonstrate that
they are "interested" in attribution (I'm trying to get them to be
more clear about this but they're not very "responsive")
- it helps local mappers (experienced or not) easily ascertain how
(in)accurate the data is (their data comes from various cartography
offices likely to have different practices); the tag displays even in
Potlatch, whereas a changeset tag requires quite a lot of
investigation to be discovered (I've only recently figured out how,
and I consider myself an experienced user)
- the import takes up around 10 MB in uncompressed OSM files (less
than 1.5 MB as PBF) and not using such tags would yield less than 10%
in space savings (most of the data consists of untagged node
definitions and of the lists of node IDs in way definitions)

So, I believe (though I may be wrong) that a little bit of
inefficiency in this case can be justified by the need to get
inexperienced users (even newcomers in areas currently without a
mapping community) to consciously act on the data to improve it.

On Sun, Jul 7, 2013 at 6:33 AM, Andy Allan <gravitystorm at gmail.com> wrote:
> On 6 July 2013 22:45, Eric Ladner <eric.ladner at gmail.com> wrote:
>> Worst case, put a "source:IBGE" on every imported item (which technically
>> should be there anyway).
>
> No, please don't do that.
>
> As discussed on other threads, the correct place to provide
> attribution is on http://www.openstreetmap.org/copyright and the
> associated wiki page, http://wiki.openstreetmap.org/wiki/Contributors
>
> If you want to provide sources for the entities in the bulk import,
> then do so using a tag on the changeset. It uses approximately
> 1/50,000th of the space in the database.
>
> Further, since it's only the first version of the entity that comes
> from the bulk import, that source for that version can be
> automatically traced via the changeset, and saves us waiting 6 months
> before adding it to the ever-growing list of
> tags-to-automatically-strip-in-editors.
>
> In summary, the entities created in a bulk import should contain the
> absolute minimal set of tags. The changesets are an appropriate place
> for any necessary metadata regarding the import.
>
> Thanks,
> Andy
>
> _______________________________________________
> Imports mailing list
> Imports at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/imports



-- 
Fernando Trebien
+55 (51) 9962-5409

"The speed of computer chips doubles every 18 months." (Moore's law)
"The speed of software halves every 18 months." (Gates' law)



More information about the Imports mailing list