Wednesday, February 22, 2012

The Case for Open Computer Programs

A few moments ago Nature made public a paper that I've written with two co-authors: Professor Darrel Ince and Professor Les Hatton entitled The Case for Open Computer Programs. The paper argues for the routine release of source code for software used in generating results presented in scientific papers.

Here's the abstract:
Scientific communication has always relied on evidence that cannot be included in publications. But the rise of computational science has added a vast new layer of too often inaccessible data sets and computer programs. Research journals are now coping with the issue of reproducibility of computationally derived results, and while it is generally accepted that all data should be made available upon request, the current landscape of regulations regarding the openness of computer programs remains muddled. Here we argue that, with some exceptions, anything less than release of the actual source programs used is intolerable for any scientific results that depend on computation. The vagaries of executing programs in different hardware and software environments ensures that exact reproducibility will always remain uncertain, but withholding code increases the chances that efforts to reproduce results will fail, possibly due to ambiguities or misunderstandings in non-explicit program descriptions, rather than to scientific errors.
I'm not the lead author on this paper, Darrel Ince is, so it's best to send any questions to him. The preferred means is by commenting directly on the paper via the Nature web site so that comments and responses are public.

No comments: