This post originated from an RSS feed registered with Python Buzz by Andrew Dalke. Original Post: Where free software and peer-review software differ Feed Title: Andrew Dalke's writings Feed URL: http://www.dalkescientific.com/writings/diary/diary-rss.xml Feed Description: Writings from the software side of bioinformatics and chemical informatics, with a heaping of Python thrown in for good measure. Latest Python Buzz Posts Latest Python Buzz Posts by Andrew Dalke Latest Posts From Andrew Dalke's writings

People often say that free software is a natural extension to the scientific/academic idea of peer review. I believe that the concepts overlap, but the requirements of academic peer review and the requirements of free software are disjoint. My two main points are:

• Free software requires distribution of modified source code. In only few cases do peer-review systems have the same requirements, and so far those cases are incompatible with the GPL.
• Free software lets you (and GNU encourages you to) distribute your software for as much money as you want. To be effective, the peer review system requires some limits on how much you should charge.

Peer review doesn't mandate distribution with arbitrary changes

Free software requires that anyone who gets source code (and anyone who gets the binary must be able to get the source code for at most minimal cost) must also get the right to modify the source code and to distribute the modified code to anyone else.

The requirements for peer review do not have the same mandate. While some journals, like PLoS, allow "anyone to download, reuse, reprint, modify, distribute, and/or copy articles in PLoS journals", most journals do not. (And do note that the PLoS license "is incompatible with the GNU GPL and with the GNU FDL.")

I point specifically to the BioMedCentral Charter which says that you may distribute changed versions "provided that no substantive errors are introduced in the process." This is clearly contrary to the GNU GPL which lets you add as many errors as you want.

(The differences likely relate to "moral rights", which is a legal concept in most jurisdictions except the US.)

Don't get me wrong! I think you should distribute your published software under a GPL or BSD license. It makes things easier for everyone involved. But I haven't heard a good argument for why allowing free redistribution of modifications to anyone is essential for good scientific peer review.

In an earlier essay I pointed out the CHARMM license which lets you get the source code under reasonable and non-discriminatory license terms of under US$1,000, and where changes can be distributed freely to anyone who has a license. How much does that inhibit effective peer review? (And I want to stress "peer review" here. It's easier to point out how it inhibits developing new science, but that's not the issue here.)

If it was available at no cost but prohibited redistribution for commercial use, would that still inhibit effective peer review?

Free software lets you distribute the software at any price

The GNU project states "we encourage people who redistribute free software to charge as much as they wish or can."

What this means is that I can publish a paper in a journal and say the software is available under the GNU GPL from my web site. Only, to get access to the software you need to pay me US$25,000 or find the web site of someone else who has already bought it and is rehosting it - which means I've got my $25,000 already.

If you think that US$1,000 for access to CHARMM inhibits peer review then you'll definitely think that $25,000 for access to my project inhibits peer review. Yet the justification for my actions come direct from the GNU project!

I'm interested in what an Open Access journal might do if presented with this case. Do they require a copy of the source as part of the submission process? Have they realized that the GNU project encourages them to "charge a substantial fee and make some money" by charging for access to those programs? Is there anything in the deposition process which limits the publishers from that sort of behavior?

Of course this isn't a new problem. I'm certain that people dealing with mice or cell lines have similar problems. Or even worse, since the publisher isn't going to keep pure mice strains alive. Are there community limits to how much a knockout mouse researcher can charge for access?

I do know that there are often limits to what can be done with physical materials used in science research, as in [the] only restriction on the use of the mouse materials is that they may not be used for industry-funded contract research.. Were that software it would be not be free and it would be against the Debian Discrimination Against Fields of Endeavor guideline.

Do bear in mind that the GNU project is concerned about software and things related to software, and "works that embody useful knowledge." Not germ lines. I bring up this point to show that "peer review" works just fine with restrictions that are not compatible with "free software."

Conclusion

I like free software. I like open source software. I've written a lot of it, I've encouraged others to write more. I've helped organize conferences based on the idea. Nothing here should be read as saying that proprietary software is a better way to do science.

I also like science and research.

The sole point of this essay is to show that while there is an overlap between the needs of scientific peer review and of free software, neither is a subset of the other. They have different philosophical underpinnings and you can't simply say that one is a natural consequence or basis for the other.

Feel free to leave comments.