[Osmf-talk] Contributor Agreement is Dual Licensing

Gervase Markham gerv at gerv.net
Tue Dec 15 02:21:24 UTC 2009

On 13/12/09 16:55, Frederik Ramm wrote:
>> How is the ODbL better than PD (or PD-attribution) in terms of
>> allowing us to use data from other people?
> I don't think the general motivation for share-alike is that you want to
> be able to gobble up into your database whatever other people add. You
> just want it to be free for everyone to use!

The general motivation for share-alike is that you and everyone else can 
share modifications. If you can't import the modifications of others who 
have followed your licensing terms, then the share-alike isn't working. :-)

> Anyone could set up a database where he imports all of OpenStreetMap
> *plus* every OSM derived database anyone has ever made.

But surely the aim of OSM is to provide the best world geodata set 
possible. Then why is OSM not the place which contains all such 
(relevant) databases?

> I am starting to be anti-import as well. If there are other datasources,
> let us link to them instead of becoming the bitrot centre of the world.

But surely that means that everyone who wants to use the combined data 
has to do all the merging, rather than it being done once, correctly?

I'm sure there are plenty of technical ways of comparing a new version 
of data that was previously imported with the current state of OSM, and 
getting the equivalent of a "diff" so OSM can be updated.

> But the question is rather academic anyway since there are unlikely to
> be other ODbL licensed data sources.

Then why _isn't_ it the OpenStreetMap Licence?

>> The position seems to have changed from "if we go to ODbL rather than
>> PD, we preserve share-alike" to "actually, we don't, but share-alike's
>> not important anyway. Go get all the data again! It won't take long."
> I thought you were talking about problems when re-incorporating
> downstream additions into OSM. What has that got to do with having to
> get data *again*?

"Again" as in "these other people have gone and gathered it, but now you 
have to go and gather it as well (again; a second time) in order for it 
to be included".

> We have established - well 80n has challenged that in his post this
> morning but I still maintain that it is correct - that most OSM
> contributors have nothing that can be protected. Not by copyright
> because it is factual, and not by database law because it is not a
> database.

If I am mapping a curved road, the route of the road is factual. The 
particular points I place, and the trade off between number of points 
and fidelity, and how I model the curve using straight lines, is 
creative. Multiply that by a hundred or a thousand roads and you 
definitely (IMO) have a creative work.

How do I best tag things using the possible tags available to me, when 
there are two or more options (footway=bridleway, access=horse, 
whatever)? That's creative. Unless you want to argue that there should 
only be a single defined way to tag anything - and I'm pretty sure 
that's a position you don't hold ;-)


More information about the osmf-talk mailing list