[open-science] Openness and Licensing of (Open) Data

Michael Nielsen mnielsen at perimeterinstitute.ca
Thu Feb 5 22:24:58 GMT 2009

On Thu, 5 Feb 2009, Rufus Pollock wrote:

> As your comments are going to be public on your blog soon, hope you
> won't mind me cc'ing the open-science mailing list on this.

Good idea.

>> (1) I think the article underrates the problems that may be caused by
>> licensing incompatibilities - witness all the problems this has caused in
>> the open source world, where the commons has fragmented;
> There have been occasions where license incompatibilities have been an
> issue but I am not sure I'd agree that the 'commons has fragmented' in
> open source. Do you have specific cases in mind?

See, e.g., the licensing slide at:


Software created under (for example) the Apache, Mozilla and GPL2 licenses 
is all mutually incompatible, in the sense that code from one project 
can't be combined with code from another.  Instead of a commons, we're 
effectively seeing the construction of silos --- you may take freely from 
one silo, but once you've picked it, that's the end.

Another example which comes to mind is the saga of getting Wikipedia and 
CC into sync, now happily concluded:


The obvious response someone might give is to say "Well, we should make 
sure that kind of license proliferation doesn't happen in science."  But 
that's easier said than done, especially if licensing becomes a big issue. 
My instinct is to concentrate on norms, and keep the licensing side as 
simple as possible.

>> (2) I think the article takes for granted that scientists are going to want
>> open licenses.  I don't see that this is necessarily true, certainly if
>> current norms are encoded in the license; and
> I quite agree with you so I think there's a misunderstanding! I'm not
> sure either that many scientists or other scholars will -- at least at
> the beginning. 5 years experience of talking with people on this has
> definitely taught me that :)
> The question I would ask to you here: if encoding current norms
> regarding data in licenses is a problem then encoding them in explicit
> community norms is also a problem (and if we're not explicit there are
> going to be serious problems ...).

I'm not sure what you mean by "explicit community norms".

>> (3) the article implicitly assumes that the license (not the norm) is how
>> enforcement will be handled, yet I think there is little evidence to suggest
>> that this is true in academic science, where norms are far more often the
>> remedy of choice.
> Let's be clear: most open licenses whether in software, content or
> anywhere else aren't 'enforced' (and won't be enforced) by recourse to
> the courts. Like norms licenses will be primarily enforced by social
> pressure. Saying use licenses isn't saying: let's settle all our
> disputes in courts! Much of the time a license is just a way to make
> explicit what the 'social contract'. But it has the added advantage of
> potentially being enforceable against those who might ignore the
> social norms (I give other reasons in the essay so won't rehash them
> again here).

I'll just say that I agree with what John Wilbanks said here --- the 
little you gain is not worth the cost in flexibility.

Michael Nielsen

Home: http://michaelnielsen.org/
Blog: http://michaelnielsen.org/blog
del.icio.us: http://del.icio.us/nielsen
Email: mnielsen at perimeterinstitute.ca
Phone: +1 (519) 498 1476

More information about the open-science mailing list