Re: Licsensing scientific data
On Fri, Aug 12, 2011 at 2:01 PM, Egon Willighagen
<egon.willighagen@gmail.com> wrote:
> Hi Andreas,
>
> On Sun, Jul 31, 2011 at 10:04 AM, Andreas Tille <andreas@an3as.eu> wrote:
>> For one moment I was thinking that it is a correct thing to
>> release scientific data under a license which does not allow changing
>> the data because - hey, I do not want even myself change the result of
>> my measurements so why should anybody else do so.
>
> Format changing is one reason why being allowed to change data is
> important. Fixing data is another, and transcription error are
> abundant.
>
>> However there was some interesting argument that some data might be
>> in some unusable format and simply needed reformating to be used by
>> the usual programs which handle this kind of data.
>
> The format changing is an important one indeed. Making data available
> in Resource Description Framework (RDF) formats already requires that.
>
>> So what the license should actually reflect is that we do not want
>> anybody to change the real content like changing / adding / removing
>> numbers or something like this. There is no need to forbid changes
>> which are relevant for say the MD5 sum of the file but just changing the
>> spacing.
>
> I am not sure why you should disallow change at all, and would
> strongly argue against it. Not being able to mix, change, build on
> other work, does not sound very useful to science to me.
>
> Can you please elaborate on your specific example, and why changing
> that data would be evil?
>
> Egon
I think one can argue that UniProt using CC BY-ND 3.0 forbids common
tasks like converting a UniProt XML record into GenBank plain text
format, loading it into a BioSQL database, or even making a multiple
sequence alignment of several proteins from UniProt. Those are all very
reasonable and everyday tasks, but they are derivative forms of the
original UniProt data, and therefore in breach of the license. This is a
ludicrous situation, so I guess anyone which has noticed this probably
just turns a blind eye to it.
I believe that UniProt should follow the revised advice of the Creative
Commons, and not use a CC licenses:
http://sciencecommons.org/resources/faq/databases/
Ironically the UniProt license page http://www.uniprot.org/help/license
links to the above advice - well it actually links to the old advice at
http://sciencecommons.org/data/dbfaq which redirects to the current
advice.
Peter
Reply to: