open-source software for bioinformatics (was Re: Unix vs Linux - the movie.)

James Bonfield jkb at arran.mrc-lmb.cam.ac.uk
Tue Aug 15 11:32:28 EST 2000


In <87lmxzyzn6.fsf at genehack.org> "John S. J. Anderson" <jacobs+usenet at genehack.org> writes:

> So, about 25% interface, 75% 'real code'. Is that fair?

Sounds about fair.

>    Given that:
>    Peer review is done to make sure the conclusions of papers are
>    'correct'.

To reiterate a point already made by another followup, this statement is not
true.

People don't make their crystals available to the reviewer to check validity
of a paper describing their structures. People have to take responsibility for 
their own work somewhere along the line. The reviewer simply needs to check
for obvious mistakes and make judgements on whether the result is worth
publishing. In the case of a paper on software I feel it is important for the
reviewer to be able to test the software, but not necessarily to scrutinise
the code.

>    And given that:
>    For some papers, software plays a critical role in determining the
>    conclusions of the paper.

I couldn't agree more. In which case why is it that most papers using software
(rather than publishing new software) do not list the programs they used, let
alone the versions and parameters? For a while we even had this as a
_requirement_ in the licence (it may still be, but I haven't checked), but
people still didn't bother.

This would have two obvious benefits: firstly it gives an indication of the
credibility (or lack of) to the results; secondly it helps to promote the
software to others (and so helps the authors).

I _do_ agree that the software described in a paper (that is publishing that
piece of software) should be available for reviewers to use and experiment
with - otherwise how can it be judged properly. However I do not feel the
quality of the source code itself is as important as the algorithms and their
validity.

In the open source world, obviously the software will indeed be available. In
the closed source world the software whould be available, but not necessarily
in source form.

James
--
James Bonfield (jkb at mrc-lmb.cam.ac.uk)   Tel: 01223 402499   Fax: 01223 213556
Medical Research Council - Laboratory of Molecular Biology,
Hills Road, Cambridge, CB2 2QH, England.
Also see Staden Package WWW site at http://www.mrc-lmb.cam.ac.uk/pubseq/







More information about the Bio-soft mailing list