[Osmf-talk] my views on the ODbL

Gervase Markham gerv at gerv.net
Mon Dec 7 14:53:57 UTC 2009


On 07/12/09 04:07, Richard Fairhurst wrote:
> This is an interesting point which, in amongst all the flamage and 
> conspiracy theories, deserves some discussion. (It has already been 
> discussed on osmf-talk, but not everyone, of course, reads that.)

As someone who spends a lot of time doing licensing stuff, I strongly
agree. Of all the arguments and points which have come up in the last
week, this one is the most interesting.

One disadvantage of multiple-licensing schemes for code, e.g. the one
Mozilla uses (MPL, GPL, LGPL) is that downstream users have the ability
to choose a single licence to obey, and have no obligation to release
their changes under all three. This means the uncooperative can (and, in
Mozilla's case, have) avoid recontributing their changes using this
mechanism.

In Mozilla's case, this was already possible when we were MPL-only - the
MPL is a weak copyleft. So we don't get too stressed about it. But for
projects which want a proper copyleft, it removes the ability to always
reintegrate changes others make. Which is the whole point.

In effect, the ODbL and the Contributor Terms is a sort of
dual-licensing. The ODbL is one licence, and "let us relicense it later
if we want" (I summarise for simplicity) in the Contributor Terms is a
second licence, external to the ODbL. And so we hit this dual licensing
problem.

> OSMF's rationale for requiring sign-up, I believe, is that it makes 
> relicensing easier. If copyright laws change drastically in two years' 
> time, OSM has the ability to relicense the entire dataset under a "free 
> or open licence" by a simple "majority vote of active contributors".
> 
> It does, however, mean that data that has been "opened" by ODbL's 
> share-alike terms may not necessarily be available for use in OSM. (I 
> say "may not" because OSMF is at liberty to accept contributions from 
> sources which have not signed the terms, if it wishes.)
> 
> So the decision is which of those is the most important.

A very good summary of the question.

> My feeling is that I would rather have the latter. I would like to see 
> the Contributor Terms written so that the relicensing ability is kept 
> for the first 6 or 12 months after the changeover, just in case a sudden 
> problem is discovered with ODbL which causes us all to rethink.

That sounds like an excellent compromise. Given the relatively slow pace
of license-related change, I would suggest 12 months rather than 6. This
will also be a great incentive for people to keep studying and working
on the licence after we upgrade.

> But bearing in mind that ODbL has an automatic upgrade clause _anyway_ 
> (4.4a), I believe the safeguard is already in place: if laws change, 
> ODbL will change to meet them.

Particularly as the licence is being designed around our needs, and we
have significant input into future changes (which is not true of
CC-BY-SA, and would not be true of a license with many other users).

In this case, having a belt _and_ braces comes with a significant downside.

Gerv




More information about the osmf-talk mailing list