[OSM-legal-talk] Progressing OSM to a new data Licence regime

[Richard Fairhurst]

Thanks for all the comments so far. This is an OSMF response drawn up  
by those of us who've been taking this forward (myself, Steve, Andy),  
but we are, of course, not lawyers and so it's not a definitive legal  
answer.

As per Steve's e-mail, we would intend to assess all the feedback and  
take it to Jordan Hatcher as we proceed with the licence.


David Earl wrote:
> Steve, how do we avoid the situation where e.g. someone who  
> disagrees the new license has run a bot over all of Cambridge to  
> tweak things (as has indeed happened to many of the ways) or who has  
> 'tidied up' bits of my mapping so all my surveying is now labelled  
> with their name. Does all of my mapping of Cambridge get deleted  
> because someone has later modified my work in a trivial way?  
> (Conversely, can I just select a big area, and add a new tag to  
> transfer the data to my name and cause someone who doesn't agree the  
> new license to be retained?)

Firstly, rights only exist in the data where a "substantial"  
contribution has been made - so a trivial contribution like this would  
be unlikely to affect the relicensing.

Secondly, we do of course have full history of all data (albeit some  
of it archived), so your original Cambridge data (assuming you agreed  
to relicense) would still be available to OSM.


Tom Evans wrote:
> If we're going to do this anyway, can we not allow users to mark  
> their preference as public domain too?  It seems a significant  
> number of OSM participants may be perfectly happy to have their data  
> given away PD, and storing an option per user would make this  
> possible.  We're going to do that work of asking each user anyway,  
> so why not let each user mark themselves as
> one of:
>
> a) Public Domain
> b) Open Database License
> c) CC by SA (the default now)

Thus far everyone who signs up to the project has already agreed to  
CC-BY-SA. We don't believe it's a viable option for data licensing  
going forward so would not seek to offer it as a choice.

The idea of storing an optional 'PD?' preference per user is an  
interesting one and we'd welcome feedback as to whether there'd be the  
demand for this. Effectively this would be requesting that OSMF, or  
anyone else, creates a public domain database from your contributions.

Of course, if you agree to PD, your data can be relicensed in any way  
so this would also imply acceptance that it can be distributed under  
ODC-Database (or anything else).


Ivan Sanchez Ortega wrote:
> I don't quite understand the difference between ODL-factual
> and Public Domain.
>
> I mean, what would be different if, instead of ODL-DB + ODL-factual,  
> we used ODL-DB + PD ???

This would certainly be an option. However, ODC-Factual is an explicit  
grant of rights and is applicable in countries where there is no  
formal concept of the public domain.


Frederik Ramm wrote:
>    I would like to know more about our talks with the CC folks.  
> After all, the people at Creative Commons have a lot of respect from  
> the world-wide community and I read on the wiki page that:
>
> "... We would have liked Creative Commons to have offered a   
> sharealike/attribution data licence that we could adopt. However,   
> their position is that map data should be dedicated to the public   
> domain, ..., the OSMF board does not believe this route is in the   
> project's best long term interests."
>
> So they actually listened to us and to the doubts we (some of us)  
> have about CC0, but still they take the position that map data  
> should be CC0. I am sure they didn't just say so on a whim, they  
> must have had good and solid reasons for that, and the OSMF must  
> have even better and more solid reasons for rejecting their  
> suggestion.
>
> [...]
> Why, exactly, does CC recommend CC0 even after they have thoroughly  
> looked at our situation, and on what basis does the OSMF board  
> reject the CC suggestion?

Creative Commons' position is this (summarised from an e-mail to OSMF  
by John Wilbanks, head of Science Commons, to which CC has expressly  
delegated its policy on data licensing):

A licensing approach based on intellectual property (i.e. a  
copyright-based licence) is not applicable for factual recordings. The  
EU Database Directive may apply but CC believes this is bad law and  
does not want to use it.

A contractual approach could work in theory, but is only binding on  
those who sign it, and CC believes any "bad guys" wanting to find a  
way around the licence would do so.

OSMF disagrees significantly with this assessment of a contractual  
approach. Commercial geodata (TeleAtlas, Navteq etc.) is protected  
this way. We also believe that the experience of free software/open  
source licences is that enforcement is largely achieved by public  
pressure: the GPL is enforced more through bad publicity for alleged  
violators (or "bad guys") than by actual legal action.

OSMF believes that CC's position is essentially formed from their work  
with scientific data, and that their recommended approach of "norms"  
is laudable but less likely to be practical in the geodata market.

If there was a community consensus to move OSM towards a CC0-like  
approach then we would, of course, give it further consideration but  
we believe this is not the case.

We should say that CC have very generously given of their time to  
communicate their position and the reasons for forming it, and we do  
appreciate that.



Tim (chippy) wrote (and TomH followed up similarly):
> If I distribute web mapping of my special company data and OSM, I  
> don't have to give my data back to OSM, right? In other words,  
> there's no requirement to distribute, but also, if the map images  
> are distributed, then it doesn't have to be under the same licence.
>
> If I zip up the shapefiles used, and put them on my server for folks  
> to download, then these would come under the same licence and be  
> able for OSM to benefit from them?
>
> How about putting my propriety data and OSM together locked within  
> an in-car sat nav system. Would this be classed as distribution of  
> the database? What should my company do in this case?

These should be answered at the "differences" page at
http://wiki.openstreetmap.org/index.php/Open_Data_License
(which wasn't originally linked from the FAQ - apologies)

First case: if the work is derivative then you do have to give your  
data back. We are asking the licence authors to add a further  
provision to make it explicit that the data has to be contributed  
back. This is probably the aspect of the licence over which we've had  
most discussions and we are determined to get it right.

Second case: yes, they would come under the same licence (again,  
assuming that the result is derivative).

Third case: there's an explicit clause (4.6) that a derivative  
database protected by technological measures (such as one sealed in a  
satnav) must also be made available in unrestricted form.



Richard (for OSMF)