[Osmf-talk] License with or without virus

Matt Amos matt at asklater.com
Fri Dec 4 16:06:35 UTC 2009


Jaak Laineste wrote:
>  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.

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

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.

>  Making parallel with the open software world, I have a feeling that 
> there is  trend to have less GPL projects, and it is more “trendy”  to 
> have non-viral projects. For my own commercial/enterprise projects any 
> GPL software usually means automatically no-go decision;  so in theory 
> the virus clause should open (free) more software but in practice it 
> puts it into the chains and limits usage a lot.

the share-alike-ness makes it more difficult to combine OSM with other 
datasets, sure. that's got to be part of the decision process and is one 
of the reasons CC cite for recommending data projects go CC0.

i think it's important that OSM data can be augmented with 3rd party 
data, especially if it's non-geographic data that is outside of OSM's 
scope like subjective review data. we've discussed this on the 
legal-talk list before and i think we have a practical way of allowing 
"linkage" similar to the way LGPL works.

>  As a contributor, I personally would not mind removing share-alike 
> principle from the new OSM license. Even if it would mean that 
> Google/TA/NT/etc could therefore use my work in their products. It would 
> be actually great to see that work of my little hands will be in Google 
> Maps or in my TomTom Go, and loosening share-alike term would be the 
> fastest (or maybe the only) way to make this happen.

personally, i don't quite get the "i don't want PD because i don't want 
people to profit from my free work" argument. people are already 
creating businesses based around OSM (whether profitable or not). my 
preference is that there should be some "cost" to using OSM data, 
whether that's sharing-alike or just prominent attribution.

as you say, i'd feel really proud if i went to google maps and there was 
an OSM attribution sitting in the corner.

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. (in the interests of a full and frank debate i 
have to say that i'd prefer if we didn't import these datasets anyway)

>  I was also thinking of some kind of dual-licensing scheme (so OSMF 
> would get a lot of $ from Google to remove share-alike term), but for 
> OSM-like projects this would be probably just undoable. I do not know 
> any successful examples of this.

yeah. bringing money into a collaborative project is always going to get 
complicated, and is probably why there aren't any successful examples.

cheers,

matt




More information about the osmf-talk mailing list