Re: Metrics for OSS performance

James Howison <[email protected]> Mon, 19 Jan 2004 10:58:05 -0500
Newsgroups gmane.science.opensource.community
Message-ID <[email protected]>
On Jan 19, 2004, at 10:24 AM, Jorge wrote:

> Does anyone know of any reference with a good (valid
> and reliable) mesure of "performance of Open Source
> Software"?

This is an interesting question and AFAIK there isn't a canonical 
answer.

Kevin Crowston, Hala Annabi and I presented a paper on a related topic 
at ICIS late last year.  It is a survey of success measures which I 
think will be of use to you.  I just realized that this wasn't posted 
to opensource.mit.edu which I'll do now.

Meanwhile the project and paper is available here:

http://crowston.syr.edu/floss/

Is performance the same as success? Not exactly, but some of the issues 
(and probably measures) are the same. And this is something that I want 
to add to the Success Paper.

Measuring success is hugely dependent on the goals of of the one doing 
the measuring.  In Proprietary software development this defining power 
rested almost wholly with the Executive management, who largely relied 
on the advice of accounts.  While commercial performance is still 
relevant in FLOSS projects the cross-institutional nature of FLOSS 
means that the power to define success lies much more directly with the 
developers.

I think that any definition or measure of performance will need to take 
into account this element.

Please keep in touch because performance is of great interest to us as 
well.

Best regards,
James Howison

--

PhD Student
Information Studies
Syracuse University
PGP.sig (application/pgp-signature, 194 B) - not displayed