Simon: I didn’t want to name names, but Beast is a good example of nicely developed and maintained software.

I understand your point of rewards and agree with it. And because of this we need to change a little how scientific software development is perceived as a career advancement. There should be a way to make someone’s career advance without the need of a peer-reviewed publication (not in all cases of course) if he/she had spent years developing a great application that would benefit someone else’s work.

As Open Science is trying to show the impact of blogs and other media in the advancement of Science, we should change the way scientific software is viewed and developed.

There’s a real under-appreciation of software in science. Good software, is crucial to our work, especially in a field like bioinformatics/phylogenetics. BUT there’s no reward for doing it. A programmer can spend months coding, debugging, dealing with feature requests, dealing with people who want to use the software, etc. I think the BEAST developers do an outstanding job at answering questions and trying to help people (some of whom are well out of their depth using the software) , and I HATE to think how much time they spend on that a week.

At least, as things are now, the programmer can get a paper out of it, and get some citations from it. This is a small reward, and I think, totally deserved. If you’ve created something that helps someone do their job, then you deserve some recompensation. Academics are paid in reputation (citations, publications), surely this is very little to ask?

If the best reward that a programmer can get from spending 6 months developing some applicatio is a sense of satisfaction, then what’s the point? He/she’s wasted 6 months that they could be publishing papers and advancing their careers. I think you’d see a huge decrease in the number of apps produced, and I think you’d see a huge decrease in the number of apps coming out as free (as-in beer).

(Aside: I say this as someone who’s just finishing writing an application note .

–Simon

1. The PI (a non-programmer) wants software.
2. They give the task to the “lab bioinformatics person” (LBP) – usually a student or postdoc who can program, but probably has no formal training in software development.
3. The LBP hacks something together that more or less works.
4. The PI decides that enough time has passed and rushes the work to publication.
5. The LBP tries to point out that their software is far from complete, buggy and undocumented but to the PI, methods are just a means to results. They apparently have results, so they are happy.

In other words: it’s the usual publish or perish story, academia doesn’t take software development seriously and apps are just rushed out so we can all get our name on another paper.