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

[Rufus Pollock]

2009/2/9 Neylon <cameron.neylon at stfc.ac.uk>:
[snip]

> [wanting to define terms]

This is really useful!

[snip]

> By ³license² I mean some form of contractual arrangement between the
> provider of data and the user of data, such as a copyright notice,
> declaration of database rights, or ³click wrap² agreement. As Yishay pointed
> out in comments to my last post, a Public Domain Dedication looks quite a
> lot like a license. However I don't think of it as such because a PDD
> removes the need to give or seek permissions at all, removing the need for
> any contractual arrangement between users and provider.

OK, this is maybe where confusion comes in. For open stuff by a
license I just mean a document that says:

"You're free to do all of this stuff without asking any kind of permission"

subject to, at most, requirement to attribute or share-alike.
Furthermore, I would insist that the license should impose zero burden
on users (other than perhaps attribution or share-alike) so no
'click-wrap' type stuff should be allowed. (See item 9 of
http://www.opendefinition.org/1.0/).

Declarations of e.g. copyright or db rights do not need to be in a license.

> By ³data² I mean collections of data. As Jonathon Rochkind pointed out and
> as we should all remember, a specific datum itself is in the public domain
> by its nature. A fact of nature is not subject to copyright or intellectual
> property law. Collections of data, and presentations of data, however can

Quite right. I've now posted one of the appendices I mentioned originally at:

<http://blog.okfn.org/2009/02/09/facts-and-databases/>

> acquire potential database or ³sweat of the brow² rights in certain
> jurisdictions. It might be advisable for us to use the singular and plural
> to draw the distinction. The line between a datum and data that defines when
> these rights might be acquired is a difficult one, even for lawyers to
> define, and this confusion forms part of the argument for public domain
> dedication.

Though this problem exists elsewhere (though perhaps less severely).
If I copy the structure but not the letter of someone's code am I
infringing etc etc.

>> The remaining question is: will attribution and share-alike (if
>> included at all) be in licenses or in norms?
>
> For me, share-alike only makes sense in the context of a license because it
> only has any value if you can enforce it. As Michael has said community
> enforcement of norms can only be through exclusion from the community. The
> point of share-alike is to force people who would otherwise not to use the
> same protocol/licence. Without a contractual arrangement there isn't any
> real means of forcing people who choose to act outside the community.

Yes though we might imagine this community's views having some degree
of force. For example, one might want to participate in this community
in which case share-alike requirements on a given dataset might matter
(even in norms).

> But I also wonder whether we're actually worrying about something that isn't
> very important. I think we are all agreed that non-commercial clauses are
> not acceptable for legal, practical, and philosophical reasons. But this is
> where the real battle is going to lie, not share-alike, which I don't think

*100% agree with this and your points below* (snipped for brevity!).
Non-commercial restrictions are a major issue and I think it is very
important that makes things non-open (this is not saying
non-commercial is per se bad by the way but it is saying when building
the open data commons that non-commercial restrictions is not
compatible with 'openness' of data).

> I wonder if our energies are better spent making that case and just
> shuffling share-alike under the carpet. If we generally don't think its all
> that helpful, and if I am right that most scientists won't care that much
> (which remains to be seen) is it worth getting that stressed about?

I would agree but I would be unhappy to have it excluded especially
since some major data communities are using it (e.g. Open Street Map).

Rufus

[snip]