[okfn-discuss] Licensing question from the Open Science Training Initiative pilot

Wed Jan 16 16:19:22 UTC 2013

I'll do some inquiries, but going forward, Peter's advice is strategically as well as normatively the best policy -- as long as possible, use only free and open source software in a course training young academics in the ways of open science -- teaching, practice and preaching are all intertwined.

Matlab has many foss alternatives -- 

* R (and R Studio)
  http://www.r-project.org and http://www.rstudio.com
* Perl Data Language
  http://sourceforge.net/apps/mediawiki/pdl/index.php?title=PDL_for_Matlab_users
* Scipy/Numpy
  http://www.scipy.org and www.numpy.org
* Sage
  http://www.sagemath.org

On Jan 16, 2013, at 4:15 AM, Peter Murray-Rust <pm286 at cam.ac.uk> wrote:

> On Wed, Jan 16, 2013 at 11:18 AM, <sophie.kershaw at keble.oxon.org> wrote:
> 
>> Hi everyone,
>> 
>> As many of you will know, I'm currently running the pilot scheme for my
>> Open Science Training Initiative at the Doctoral Training Centre,
>> University of Oxford.
> 
> 
> Congratulations! My greetings to the students on the course (this has crept
> up on me!). Unfortunately I am off abroad soon so can't come and look in...
> 
> 
> 
>> The students have taken to licensing fantastically well, even though the
>> concept was completely new to *all* of them this time last week. Yesterday
>> I fielded a more specific question from one of the students and was hoping
>> some of the legal eagles on this list might be able to help.
>> 
>> The student in question has developed a series of his own routines in
>> Matlab and has drawn them together in a GUI, which he has also designed. He
>> wanted to take a screenshot of this GUI and put it into his report.
>> However, he wasn't sure how this should be licensed. If parts of the
>> Mathworks Matlab interface are visible as part of his screenshot, would
>> this preclude him putting say, a CC-BY license on the resulting image?
>> 
>> I don't think there is a generic answer - you will have to read the Matlab
> contract (which I assume the university has signed). It'd unlikely, though
> possible, that this is specific to the licensee so you *may* have to
> contact Oxford U purchasing.
> 
> Software Companies often set clauses requiring or preventing this sort of
> thing. For example Gaussian Inc. (whose program solves Schroedinger's
> equation) FORBID the publication of any results. This is because people
> might compare the results with other programs which solve the same problem.
> (BTW solving S's eqn is a billion+ industry).
> 
> OTOH I have encountered companies which require the publication of company
> info. If you use the program you are contractually required to tell the
> world and sometimes cite a particularly publication.
> 
> This is a compelling argument for the value of F/OSS software where all
> such permissions are granted by (I think) all conformant OSI licences. Is
> it possible to do the same thing in R? This area - how to compute Science
> Openly and reproducibly is what I am working on at present. I can see the
> possibility of tools which can do some of what Matlab can do and do it
> semantically and Openly.
> 

--
Puneet Kishor