Ask HN: What is a good license for university based open-source software? | Hacker News Reader

Ask HN: What is a good license for university based open-source software?

5 pointsHoliTrea7 comments

To make a long story short:

After reading an ever growing number of articles about "research should be open source" I finally decided to become active and try to convince my boss to publish some of the software framework I am developing during my PhD. It's nothing fancy but I like the idea of sharing knowledge and it might help other PhD student.

The only problem I have to solve is: My boss wants to have some form or license which states: "If you use this software to produce publications, then you must cite our paper describing this software".

I tried to find something like this but did not come up with anything useful. Has anybody here been in a similar situation? Can I add such an addition to a BSD style license? Or is it enough when I add it to the README?

View on HN

There is a bit of a problem using a "must cite our paper" clause. It's very similar to the "obnoxious" advertising clause of the original BSD license, described at http://www.gnu.org/philosophy/bsd.html .

In short, it would mean that people who incorporate your software into their work must also include that citation requirement. Put enough of those together and you might end up with a dozen or more required citations - even if the only thing I copied from your code was a routine for, say, doing matrix multiplies!

Also, what constitutes a "publication"? Is it in a peer-reviewed journal only, or do non-peer reviewed publications also count? Conference papers? Blog posts? If someone does not cite the paper then is the license withdrawn for that person? For how long? Do they restore the license by downloading a new copy of the code?

I'll use the Biopython practice as an example of how another project does things. The README says "If you use Biopython in work contributing to a scientific publication, we ask that you cite our application note (below) or one of the module specific publications (listed on our website): ..."

In similar situations I've either used a "BSD with attribution" style license (i.e. http://fedoraproject.org/wiki/Licensing/BSD_with_Attribution) or simply included a "How To Cite" section in the README. For example, ImageJ is popular in research and the docs include a citation recommendation:

    Abramoff, M.D., Magalhaes, P.J., Ram, S.J. "Image Processing with ImageJ".
    Biophotonics International, volume 11, issue 7, pp. 36-42, 2004.

Another license I've seen used in academia is Matt Might's CRAPL (http://matt.might.net/articles/crapl/), which is in some ways very well-suited to the academic publishing world but is pretty tongue-in-cheek.

In my experience (in physics and materials science) most academics want to cite your software if it was useful. Make it clear and obvious, and you'll probably get citations.

Just make it easy to cite - often the most difficult part is finding the appropriate citation. For example, how would you cite bash?

And consider the University of Illinois/NCSA license http://otm.illinois.edu/uiuc_openSource - according to gnu.org, it is a simple, permissive non-copyleft free software license, compatible with the GNU GPL.

FWIW, I've avoided licensing "must cite" resources several times because the terms were worded problematicaly (e.g. they encoded invalid assumptions, etc.). But I've never had a problem with "how to cite" - in fact, it makes integrating the resource that much easier.

Use the BSD or MIT license, and just add the line your boss wants to that.

Do such modified licenses still qualify as FOSS (for those who care about such formal qualification)?

It's conceptually similar to the advertising clause of the original BSD license, so I believe the FSF viewpoint is that it's free but "obnoxious" and not GPL compatible.

Reply on news.ycombinator.com