[Osmf-talk] Results of OSMF Member Vote

[Matija Nalis]

On Mon, Dec 28, 2009 at 09:47:54AM +0100, Frederik Ramm wrote:
> Matija Nalis wrote:
>> However, if we do get the same (or similar) turnout on the actual
>> relicensing question sent to OSM contributors (due to happen soon?), that
>> would AFAICT mean that contributions of all those ~45% people who didn't
>> bother or didn't want to answer the question, as well as those ~11% who were
>> against will have their data REMOVED from the map
>
> I don't view the matter with quite the same urgency as you do.
>
> If a significant amount of data is not relicensed, and if OSMF should go  
> ahead nonetheless (which I doubt it will),

Ah, but that is what I'm calling a vote for. To see what is that acceptable
damage threshold to OSMF members. Shall we proceed if 70% of the data will
be destroyed ? If 30% is destroyed ? What is that percentage ?

Note that it must be decided eventually anyway - I just think it should be
decided in advance, as later that might be that thinking "we're so close to
the target, we shouldn't let this one remaining thing to stop us now when
we've come so far" that even a reasonable people will succumb to (as shown
many times in history).

I'm not saying that I myself should proclaim that "anything above 10% data
loss in unacceptable" (that would be foolish, really) but just that the OSMF
membership should vote (and soon) on that and decide what that "acceptable
data loss percentage" should be.

Would you agree that such a vote would be a good thing (or do you think it
would be bad, and why) ?

> then people will simply run the old data set on some server of their
> own[1]. There will, at worst, be some ripples and ruptures in the
> community. If certain countries or areas are hard hit by people not
> relicensing, we might see regional splits/forks, where the local community
> continues to use the CC-BY-SA licensed data set and the little excerpts
> that have been relicensed go to rot on the main OSM data base.

I have several problems with that. 

Firstly, I'm not so sure alternate servers would be so easy to set up (at
least in Croatia, especially in this times of economic crisis, even if
people with enough will and time and knowledge to do it could be found
[which is a problem in itself!], they won't have access to hardware and
bandwidth to do such a setup).

Would OSMF really set up such a servers (to accommodate those not complying
with OSMF decision to relicense) and why ? If we want to have data under
ODbL, helping others to license their work otherwise would be working
against ourselves, no ?

Secondly, IMHO any fork would damage both sides greatly. The forks are
necessary evils to be used only in most grave situations. And unless one of
the sides is very minor in comparison to the other, the losses could be too
big, enough to destroy both forks (just look what happened to *BSD folks,
they never recovered to their days of glory). If *multiple* forks are
created, the damage would go up exponentially.

Thirdly, I do not think that the people who *really* want to remain with
CC-BY-SA would be the problem. While they do exist, I think they're minority
(in OSMF poll, it that's any indication, they were less than 6%, and it
might be even lower if voting "against" meant losing all your data from the
map)

IMO there are 4 major groups in such a fallout:

1) people who somewhat prefer CC0 or other PD-alike license (but mostly not
   so much that they would lose data over that issue - after all they have
   been contributing under CC-BY-SA all this time)

2) people who prefer copyleft (SA) license (possibly strongly, and some
   might refuse to relicense under non-copyleft license. But some minor part
   might not like ODbL)

3) people who don't care (enough or anymore or at all) and/or uncontactable
   people (including people who might have relicensed gladly but are now dead etc)

4) people who would support ODbL (more or less willingly; ie. any of the
   above 3 groups) but whose data will be deleted because it depends on
   someone else's data which will be removed.

Now, most of pro-copyleft license people (group 2) would agree that ODbL
would be more successful at that task of protecting copyleft than CC-BY-SA,
and would probably have just minor grudges against ODbL due to it's
complexity or multiple laws or other problems, but I guess most of them
would go with the flow.

Group 1 might prefer CC0, but have already shown that most of them will go
with the flow (most would probably be equally unsatisfied with both CC-BY-SA
and ODbL)

Now, it the numbers of OSMF vote are any measure, we'll have about 6% of
those against (minority of groups 1 and 2 who are fine with CC-BY-SA but
against ODbL for some reason) and a very big group 3 (45% -- while I do hope
that number would be somewhat reduced by implicit threat of destroying their
work if they don't relicense).

But what remains is unknown extent of group 4. They might be anything from
0% (almost impossible) to 99% (also not very likely). And they would be most
enraged, because even with their full cooperation they will be counted under 
"collateral damage". And this group will grow with each OSM contributor who
doesn't (or couldn't) relicense (for any reason, which means users from all
groups 1-4)

While those 6% (anti-ODbL users from groups 1 and 2) might benefit from
forks (combined with group 3 it they fork under CC-BY-SA), most of the 
group 4 won't - they're losers no matter if they embrace some fork or stick 
with "original" OSM (as they've probably edited both datasets). 
And through no fault of their own, and there isn't anything they could do. 

So those who cooperated the most with work of others will be the most
damaged (most of their work will be in group 4), while those who ignored
work of others will be most safe *not* to be in group 4. Essentially, such a
outcome would weed out those willing to cooperate...

So I think groups 3 and 4 would be the biggest problem.

> But my guess is that there will be a strong gravitational pull from OSM
> proper, and after a while everything will be back to normal.

there might be a pull back for those ~6% (who willingly didn't like ODbL),
but there is nothing that can pull back much larger groups 3 and 4 (by
their definition !)

> It might be worse in some areas, and better in others; it is too early  
> to tell. It may be rough seas, but I don't think the ship's going to 
> sink.

If too many of the sailors jump to evacuation boats (different regional
forks) there might not be enough sailor remaining on the main ship to save
it from sinking. 

On the other hand if main ship survives, many of the evacuation boats might
not make it back to the main ship by the time they decide to so (and if they
decide to do so). Or they might diverge too much by that time to be accepted
back.

> Ironically, those places where we have imported public domain data  
> (remember, that evil non-license which would only cause us trouble if we  
>  ever were to adopt it) will survive with the least scars.

Yes, since joining this list and doing some more research, I am now
convinced that CC0 or some other PD-alike would be much better for
everyone..

-- 
Opinions above are GNU-copylefted.