December 2002 comp.lang.* stats

[Aaron K. Johnson]

In message <3E333ED7.ADEB8CDC at alcyone.com>, Erik Max Francis wrote:
> "Aaron K. Johnson" wrote:
> 
> > I agree. Plus, I'm not interested in working THAT hard to be that
> > anal-retentive about data which some would argue is still vague enough
> > to be
> > discounted.
> 
> Well, it depends on what you think the data mean.  What you're measuring
> is the number of unique posters per hierarchy over some period of time. 
> To first order, your figures are probably good for that (provided you're
> doing it right, etc.).  Other complicating factors such as spam will
> throw a wrench into the validity of the measurable.
> 
> But now taking that measurable (unique posters per hierarchy per unit
> time) and trying to apply it to something more general and far more
> indirect (like the popularity of a language) is a bi-ig step.  There is
> surely a _correlation_, but how strong that correlation is and what goes
> into it is extremely hard to judge.

My research indicates that the number of python users is precisely, to the last
decimal point, proportional to the usenet volume found by my script.

I'm right, and you're wrong for questioning me. So there (sticking tongue out).

-Aaron.