[OSM-legal-talk] viral attribution and ODbL

TimSC mapping at sheerman-chase.org.uk
Tue Apr 20 20:56:23 BST 2010

Richard Fairhurst wrote:
> But in five years, we have never been able to obtain clear agreement  
> for this. 
I assume this is based on gatherings of OSM members, mailing list 
discussions, IRC, etc. But I have never been directly asked by OSMF what 
the future license should be. I suspect that the majority of mappers 
will go with the flow, with a small number of vocal supporters on both 
sides. But without polling the members who can't know what the consensus 
is, so perhaps this is something OSMF should do? Until then, the level 
support for PD and ODbL among mappers is pretty much speculation. We 
might regret losing the share-alike people, but we may gain new PD only 
fans. Again, this is an unknown variable.

> OSMF has chosen to follow c) and has managed it, thus far, with a  
> surprising amount of consensus given our fractious community.
I guess I differ with OSMF as I have not seen a viable share-alike 
database license. I know many knowledgable people think ODbL is it. I am 
inclined to agree with the Creative Commons comments on the situation: 
share-alike is not appropriate for databases. Basically option (c) is 
ruled out in my mind, as I agree with CC on this issue.

>  I think  
> they've done well and am grateful to them, and LWG in particular, for  
> the thankless hours they have put in.
True, they do a job that would have made me go crazy. Thanks to all in 
the LWG! But most mappers I talk to regarding ODbL only can say "I trust 
OSMF is doing a good job, so I assume ODbL is OK". People would probably 
go along with anything they propose. But we can't assume the resulting 
license is suitable just because they tried their best. (And I believe 
they are trying their best!)

> I suspect many people  
> would support you; nonetheless, past mailing lists suggest that, if  
> OSMF were to do so, it would almost certainly result in the implosion  
> of the Foundation.
I don't see it as an either/or choice, OSMF can do both. I think SteveC 
mentioned that might be a possibility of hosting both in parallel (or 
did I imagine that)? Did that thought go anywhere? The bin I suspect... 
oh well.

Another possibility is to have a list of approved licenses that more or 
less interoperate (mainly BY type licenses) and put attribution in the 
metadata. Any rendering of the map would then be responsible of display 
of attribution based on the metadata tags. This idea probably won't work 
for SA licenses as they tend not to interoperate. That would be my 
preferred option. Or hosting that with a parallel ODbL database might work.

Thing is it is hard to get a PD organisation independent of OSMF because 
OSMF is meant to reflect the interests of the mappers. If a good chunk 
of mappers want PD, OSMF should respect that. (Admittedly we don't know 
how many mappers want PD, since OSMF didn't ask us.)

> I think your characterisation of ODbL as "bloated" and "confusing" is  
> grossly unfair. 
I was trying to support my point about ODbL being bloated based on this 
thread discussion so far. I would be interested in your (Richard's) and 
Grant's opinion (and anyone else's opinion) on my question regarding 
attribution and can it be stripped off a public domain produced work? My 
interpretation is attribution is not necessarily preserved in PD works, 
therefore that paragraph is bloat (and I would welcome any comments on 
that view).

The accusation of it being confusing is mainly a subjective view from my 
layman reading of the ODbL. I don't really understand contract law, so I 
am not particularly surprised. I am relatively comfortable reading the 
CC licenses (which are of course not appropriate to databases). But the 
difficulty of the full ODbL text being understood is an obstacle. 
Creative Commons also has this view of ODbL. This is only partly 
mitigated by the human readable summary.

Grant Slater wrote:
> Nice simple ODbL summary:
> http://www.opendatacommons.org/licenses/odbl/summary/
> (Created by Matt)
> / Grant
Did you notice my point that the human readable summary seems to be at 
variance with my interpretation of paragraph 4.3? If the human readable 
summary differs from the legal meaning of the full license, that would 
be a problem. (See the second last paragraph of 
) Does anyone in the LWG have a view on that issue?



More information about the legal-talk mailing list