[OSM-legal-talk] Deconstructing the "loss of data" claim

[Richard Fairhurst]

Hi all,

A personal view on the perennial PD vs sharealike debate. None of this  
is OSMF thinking, needless to say.

I do agree with John's position that PD is the best solution for  
licensing anything "open". When John says

> Or you take the PD strategy and tell NavTeq that the value of their
> trade secret is rapidly dropping as OSM emerges as an alternative in the
> PD. [...] You mock the culture of control and enforcement.

then I'm nodding vigorously. It's a statement that, IMO, always holds  
true. Let's say Google launches a collaborative mapping tool. Our  
sharealike is as nothing to their money in data-grabbing terms: we win  
because we are "open" and "community-centred", and they are "evil" and  
using you as unpaid labour. (It sounds emotive but I mean it. Among  
opinion-formers, Google is just passing Apple and approaching  
Microsoft in negative PR terms. It's only going to take a couple more  
knols and .cn capitulations before this chattering-class commonplace  
spreads to the mainstream.)

That's why all of my contributions to OSM are PD - not just the  
mapping, but, especially, Potlatch, which is over a year's worth of  
coding. If we were starting again today I'd argue very strongly for PD.

So theoretically I should be standing here arguing for a switch to PD  
(or an attribution-only licence; AFAICT 99.9% of OSMers want  
attribution to the project, that bit's a non-debate).

But I'm not.

Two reasons:

a) I'd say exactly the same for Linux, and
b) we're not starting again

By a) I mean that Linux would be in a _better_ position if it were PD  
(or BSD, or some other attribution-only licence), not GPL. That's my  
opinion, and I think it's the logical result of John's arguments, too.  
Yet nonetheless Linux is doing fine. It's hard to convince a dedicated  
sharealike OSMer that things could be much better when things are  
pretty fantastic as they are.

b) is the big one. We aren't going to get community agreement on  
moving to PD, I'm pretty sure of that.

My rough estimate is that 3% of people are pro-PD (e.g. me, Frederik,  
everyone in Oxford), 3% are pro-sharealike (e.g. Etienne, Steve,  
Oliver), and 94% aren't that fussed (e.g. Andy and everyone who isn't  
subscribed to legal-talk) - somewhere between "we have a licence?" and  
"I know what I generally approve of, but most of all I don't want the  
licence to get in the way".

The 94% would almost certainly still have signed up for OSM if it were  
PD in the first place. But I doubt they'd actively consent to such a  
significant change now. The debate would engender a lot of worry about  
the ramifications of going PD (as this thread shows); and if one of  
the 3% forked OSM to create an always-sharealike alternative, they'd  
take some of the 94% with them. We have something great here. Going PD  
would solve a load of legal problems, but it might, I'm sad to say,  
create some serious community ones.

So with that in mind, I'm very keen to see OSM move to the ODC  
Database Licence. I'll continue to offer my own contributions as PD,  
of course, and hope that we can engineer some way for people to  
extract only the PD contributions. Maybe in time we'd gravitate  
towards "PD is good practice", as we're doing with anonymous editing,  
and we did with tagging segments and their eventual removal: who knows.

But ODC-Database is a superb, practical solution for OSM as a whole.  
It's workable where CC-BY-SA isn't, in that it's (mostly) unambiguous  
as applied to data. It gets OSM data out there, being used, which (for  
right or wrong) CC-BY-SA has prevented in the past. Since the  
sharealike actually _works_ for data we get more stuff contributed  
back. And its future development is more likely to correlate with our  
community's will.

I do appreciate that implementation via contract might not be pretty -  
it offends some people's principles; it might well require a  
clickthrough for download; enforcement would still probably be by PR  
rather than through the courts (bvh's e-mail of 4pm GMT puts it well).  
Yet when it comes down to it, we would still be offering geodata on  
much, much more liberal terms than anyone else. We shouldn't be too  
ashamed that we have to take a few tricks from their arsenal.

cheers
Richard

(and thanks all for the comments in this thread, I'm enjoying reading it)