Python use growing or shrinking

Erik Max Francis max at
Fri Jan 24 12:07:00 CET 2003

Terry Hancock wrote:

> Yeah, maybe I should've mentioned that -- one might try to think of
> these
> numbers as being percentages of "all software development pages", but
> of
> course, they're not. It's just a percentage of the pages listed.  It
> might
> make more sense to scale relative to one representative member. But
> which
> one?

It's a good question.  Perhaps the arithmetic mean, geometric mean, or
even the median of all of them.  Perhaps even the results of the top
ranker, and then everything is represented as a percentile of the front
(again, not percentage of the total, but percentage of the front
runner).  The latter has the benefit of being totally impervious to the
addition of new elements (unless, of course, when added they rank first

It doesn't much matter, so long as it doesn't lead people to a pie chart
fallacy (everything's taking pieces of a finite pie, when that isn't
objectively the case), and the figures are relatively stable with
respect to the raw data.

Mind you, I never meant to suggest that you were trying to impose that
"pie chart fallacy" I mentioned above.  I only mentioned it because I
had thought of a similar representation of data in a similar
circumstance, and specifically decided against it for the very reason
that it made it look like everybody was taking pieces of a pie that
wasn't there :-).

 Erik Max Francis / max at /
 __ San Jose, CA, USA / 37 20 N 121 53 W / &tSftDotIotE
/  \ I am a gentlemen:  I live by robbing the poor.
\__/ George Bernard Shaw
    Kepler's laws /
 A proof of Kepler's laws.

More information about the Python-list mailing list