[Beowulf] Mailing list statistics

[Kyle Spaans]

On Fri, May 22, 2009 at 06:17:39AM -0700, Lux, James P wrote:
> Some years ago, I processed some large chunk of archives to count number of
> posts per source.. 
> 
> On 5/22/09 1:51 AM, "Bogdan Costescu" <Bogdan.Costescu at iwr.uni-heidelberg.de> wrote:
> 
> On Tue, 19 May 2009, kyron at example.com wrote:
> 
> > Is there any (semi) official statistics of the Beowulf M.L. as I am
> > curious to see how it has progressed.
> 
> I'd be interested in this kind of data too. But I would appreciate
> even more some kind of semantic analysis...

I've been a longtime lurker on the list as well, and I'm interested in this
kind of thing. I'm aware that there are probably tools for doing this
in existence but I went ahead and wrote some code to do it anyway. A couple
of months ago I was working a lot with USENET newsgroups (long story short:
I was tutoring for a CS class at my school, and they use newsgroups) and so
the program I wrote works on NNTP rather than SMTP, but I should be able to
rewrite it pretty easily, and dig stuff up from the archives / process
new threads as they appear.

The code is written in the PLT flavour of Scheme (AHA! A side effect of my
University of Waterloo "upbringing"), and you can peer at the source here:
    http://github.com/kspaans/nntp-to-dot/
and picture that my program generated from a newsgroup are here:
    http://csclub.uwaterloo.ca/~kspaans/cs136
The radial graph layout pics are the most fun (cs136-newsgroup*.png), I'm
sure we could all have some fun trying to guess who would end up at the
centre of one for beowulf.org ;-)

So if people are interested, I can take the code off of the back-burner.