Subject: Re: Charging the Charger
From: "Stephen J. Turnbull" <stephen@xemacs.org>
Date: Tue, 12 Apr 2005 11:30:21 +0900

>>>>> "Joe" == Joe Corneli <jcorneli@math.utexas.edu> writes:

    Joe> Anyway - my question is, if what you're working on is
    Joe> designing a simulation, why try to design a socially optimal
    Joe> license?  If your simulation is good, it should tell you
    Joe> which is the socially optimal license,

Doesn't work that way.  The point of a simulation is to get a
realistic prediction of outcomes.  You then iterate over the known
licenses and compare the predictions.  However, a simulation does not
say anything about test cases you haven't tried, except that you can
do sensitivity analysis and hope that the principles of interpolation
and extrapolation apply.  Note that there are frameworks (eg, linear
models) where those principles provably apply---but "this model is
linear" is an analytical statement, it is not something you can prove
using the simulation itself.

As for making the model open source: *chuckle*.  It's invariably
possible to make "innocuous" changes to these models and get "more
accurate" results.  It takes a specialist to decide whether those
changes really are innocuous, and whether the model really is more
representative of reality post-change.  Of course, I'm also looking
forward to getting a look at the model; but you shouldn't have too
much hope for scientific contribution from the "many eyes" of the
community.  A lot of heat will be generated, but little light.

-- 
School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.