[Osmf-talk] License with or without virus

Jaak Laineste jaak at nutiteq.com
Fri Dec 4 18:11:28 UTC 2009

> there are really two questions here:
> 1) is ODbL a better license than CC BY-SA?
> 2) what's the best license for OSM?

 Yes, I know that the ODbL transfer is your main question in the list right
now. And it looks that there are good reasons enough to do it. 

 But when I look how hard is the transfer between licenses, which look in
essence more or less identical for me, then what it would take to have
really a major change, like removal of "the virus" term? So maybe at least
discussion about it can be started as soon as possible.

> >  Is there actually consensus about share-alike clause? I guess that
> > contributors love it and users hate it, just as with GPL/LGPL
> software
> > license options. The quick poll showed no clear preference among
> > contributors.
> looking at it another way - the existing license is (intended to be)
> share-alike, so clearly any preference any existing contributors have
> isn't strong enough to stop them contributing.

I'm not so sure about it. There are several options why people contribute
according to current license, like:
a) they knowingly accept the terms
b) they do not care about the terms
c) there are no alternatives

> while discussing (2) is certainly worthwhile, i think we shouldn't hold
> up our progress on (1). after answering (1) we can come back to (2),
> which is a much more wide-ranging, open-ended question.

Fair enough for me.
> Going non-viral because it's "trendy" is not a very strong argument (to
put it mildly ;-) )  

 This was not really my argument :) I see that there are very good specific
reasons to prefer non-viral licenses in the open software world. It seems
that generally the IP contribution protection with viral terms just does
really not work for software (at least as much as it has been hoped for).
Databases are different in many aspects, but general philosophy why viral
terms are used in CC-SA/ODbL look quite similar for me, and the problems of
it look also similar. Basically I keep away of GPL software when I need to
integrate it with something. With the databases and OSM I see even more
benefits in integrations, therefore also dangers. 

> it also seems there are many data providers (naptan and canvec spring
> to
> mind) where they are willing to donate data to OSM, but would like to
> take corrections back into their own dataset. clearly the share-alike
> clause makes this very difficult for them. perhaps even an attribution
> clause would be too much. 

>- Let's say OSM is not going viral. In that case we cannot import datasets
that require SA. So where potentially missing out datasets. (e.g. Canada)

 I do not know about Canada case, but these sentences look contradicting for
me. Matt sais that Canvec would benefit from non-viral license, Henk refers
contrary also about Canada? Or do I miss here something? 

 According to my own experiences with local data vendors (I have also worked
for a commercial mapping company for about 12 years) I cannot see really a
case why a vendor should in principle require SA. I can imagine more cases
where SA would be a blocker. Maybe there are some useful already virally
licensed data sources for us, but I'm afraid we would open much more doors
by choosing the other option.


More information about the osmf-talk mailing list