The Ars Technica website has picked up on an article in the journal Nature about a pre-print of a paper written by two theoretical physicists from the UK on the arXiv.org e-Print archive that includes a mathematical model about why Open Source projects can inherently fix problems quicker than a Closed Source project for the same parameters.
The nub of their argument is that it is the frequent release of projects and the slew of bug reports that come back from these rapid releases that result in this effect.
Their conclusions seem to be:
- "our model shows that software projects can converge to a bug-free state even with imperfect programmers"
- "closed source projects are always slower to converge to a bug-free state than open source projects at constant parameters"
- "the quality of open source project programmers does not need to be as high as those of close source projects in order to achieve the same rate of convergence to bug-free programs."
- " the abilities of the maintainer has a much less dramatic influence" than "having better programmers"
- "ignoring bug reports on already modified code is the best option for closed source projects; this even outperforms open source at short time scales, because the programmers only work on fully buggy parts, hence the bug fixing rate is higher"
Whilst the paper has a number of assumptions (they acknowledge them and proposes further work to confirm or refute their validity) it is an interesting model that could inform developers on either side of the fence on how to speed up bug fixing. This can only be good for everyone.