Subject: Re: Open letter to those who believe in a right to free software
From: "Stephen J. Turnbull" <turnbull@sk.tsukuba.ac.jp>
Date: Wed, 10 Nov 1999 10:04:29 +0900 (JST)

>>>>> "rms" == Richard Stallman <rms@gnu.org> writes:

	SJT> (Note that Richard is wrong about his small subpopulation
	SJT> hypothesis; it is quite possible for the behavior of small
	SJT> subpopulations to affect the whole in economic models.

    rms> What I said is that you cannot measure the small
    rms> subpopulation through a statistical survey of behavior.  When

But this is just plain 99.44% false.  Use a stratified sample.  The
0.56% chance that this would not work comes in if the small
subpopulation were in fact a secret cult planning to sarin gas the New
York subway.  But it's not, it's loud, noisy, and easily recognizable
in "places" like slashdot and gnu.misc.discuss.

    rms> you were asked how you would determine people's preferences,
    rms> you said "Just measure them."  As if to imply that that is
    rms> trivial.

It is trivial, for the theory part: use a parametric representation
that includes the appropriate possibilities.  If the theory part and
other knowledge (such as the voice of a GNU wandering in the
wilderness) indicate that an important component is very small in the
overall population, we use a stratified sample.

Your concern is technically 100% unjustified.  My judgement in
modeling premises may be deficient, but now that you've noted this
possibility, I will be sure to account for it using appropriate
technical means.

    rms> My main concern is that discussion of such models tends to
    rms> lead people to accept the models' premises as more than just
    rms> approximately true.

"Talking about birth control leads teenagers to have sex."  I think
not.  People accept the premises that they like, then seize on models
that "justify" them (by assuming them, ante hoc, ergo propter hoc,
sheesh!).

Do you really like playing the part of the Pope to my Galileo?

-- 
University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
__________________________________________________________________________
__________________________________________________________________________
What are those two straight lines for?  "Free software rules."