[okfn-discuss] Modding CC-licenses for data

Mon Jun 25 18:42:25 UTC 2007

Benj. Mako Hill wrote:
> <quote who="Rufus Pollock" date="Mon, Jun 25, 2007 at 09:07:36AM +0100">
>> However, in general, I'm still not sure why you wouldn't want to use
>> an open license as opposed to a waiver. In particular one might want
>> to impose attribution or sharealike type provisions.
> 
> I was under the impressions that it was two fold. One was a set of
> practical barriers to sharing information that doing so would entail.
> This, of course, could be offset by types or problems that you mention.

You've summarized the situation admirably. As you say there are two 
distinct reasons to not use a license (or rather, to have a waiver that 
is the equivalent of something like the BSD/MIT license).

Regarding the first, as I understood it when John explained the matter, 
the main concerns are about complexity and uncertainty regarding the 
meaning of sharealike provisions (and the scope generally of the 
underlying 'database' rights). In particular, there are worries that the 
db right -- and the associated provision --- could be interpreted very 
broadly so that e.g. if you used a particular database in developing a 
drug with the database available under sharealike license then that 
would require that the drug be 'open'.

I see this point but I'm not sure how important it is. After all, 
similar arguments could be applied to the GPL in the area of software -- 
if everyone just used a 'liberal' license such as BSD/MIT this would 
greatly reduce complexity (all those discussions about just how 
'recursive' the GPL is e.g. does the fact that flex is GPL mean its 
output is GPL'd ...). Is data really that much more complex or is just 
we're only just starting to grappling with the issues?

> The second, and more important, was a feeling that the DB rights were of
> dubious legality  or long-term enforcability (since they are very
> juristiction specific, at the moment) and that arguing for control or IP
> in DBs is, in the process, arguing for stronger IP. We have to be *very*
> careful when we do this because we stand to lose a whole lot more than
> what we gain.
> 
> It's a tricky situation to know when and where this happens but I found
> John Wilbank's description of the situation reasonably convincing. I
> don't think anyone has a problem with your desire to have copyleft
> provisions. They have a problem with arguing for the property necessary
> to make it stick. We *really* shouldn't be on that side of that argument
> at this stage of the fight.

This indeed is the real nub of the issue. Personally, I'm still not sure 
how the trade-off works out. It seems possible, at least in the 
commercial world, to restrict access to databases fairly effectively 
simply using contract + access control. Of course, one cannot do this 
for all dbs -- a phone directory being the obvious example of a database 
that by its nature must be made available without such restrictions -- 
but I think it is true in many cases.

If that is so then the gain from not having formal 'db' rights is maybe 
not as large as one might imagine -- when I asked Bernt Hugenholtz why 
he thought there was no 'db right' in the US his answer was that most of 
the interest groups who would benefit from it could already achieve what 
they wanted with contract.

Whatever the case it certainly true that while 'db' rights exist (and 
they do exist in some form or other in many places currently) people 
need to put some form of *explicit* 'license', be it a simple 'BSD' 
style waiver/disclaimer of rights or something more elaborate.

~rufus