Subject: Re: Open letter to those who believe in a right to free software
Date: Thu, 28 Oct 1999 09:27:33 -0400

Stephen Turnbull writes:
> >>>>> "Ian" == Ian Lance Taylor <> writes:
> I simply think that we can disentangle the assumptions with scientific
> implications from the purely technical ones (eg, Ben Tilly doesn't see
> a practical difference between discontinuous functions and ugly
> continuous ones, so he doesn't care if I use continuous ones; he might
> want me to use differentiability, absolute continuity, a Lipschitz
> condition, or something like that to get better behavior, which is a
> scientific assumption), and the purely scientific/technical ones from
> the ones with ethical implications,
Lipschitz, definitely Lipschitz.  I like Lipschitz. :-)

The others are all consistent with models that contain at the least
good approximations of such things as hysterisis curves or the
classical catastrophes.   If such features are present I would think it
unwise to ignore them...

Conditions like continuity say much less than most people realize.
To pick a poorly known example, the Jordan Curve theorem says
that a continuous function from the circle into the plane that has no
self-intersections (or more generally from the n-dimensional sphere
into R^(n+1)) divides it into two disjoint connected open sets
known as the interior and the exterior.  OK, that makes sense, where
is the surprise?  What people do not realize is that it says nothing
about the relative areas taken by the curve and its interior!  In fact
the image of the circle can not only have a positive area, but it can
exceed the area of the interior by any factor you desire!

> OK, OK, economics does not dominate anybody who has personal goals and
> acts according to them.  I wish RMS would admit that.  :-)
But en masse?  I think that economic factors help shape public
concepts of ethics.  So unpredictability in the ability to predict a
particular person need not translate into unpredictability of what
happens in general.  To draw an analogy to the much-abused
butterfly changing the weather example*, while we cannot predict
the weather we have much more luck in predicting climate and
various long-term patterns in behaviour.  So we do not know who
will be hit when on the East Coast by hurricanes, but we do know
that more places are likely to be hit over the next decade than
were over the last one.

This is not a hypothetical example either.  In the draft that I
circulated earlier (which I should be looking for a home for, does
anyone have suggestions?) I pointed out the example of
slavery, and I gave reason to believe that the success of Open
Source could lead to a more general acceptance of the Free
Software ideology.  Because of economic factors!

>     Ian> And that I would dispute, unless economics is
>     Ian> going to become the science of everything.
> Most sociologists, psychologists, lawyers, and business school types
> believe that economics does try to be the science of everything, and
> they wish it would cut it out and go away.

Economics is pretty mild.  I have seen linguists make the claim
that their field of study properly encompasses all others because
they study how and why people use language, and since all
other fields of study communicate using language...


* The best abuse has to be Terry Pratchett's when he asks,
"Why doesn't someone do something about all of those
butterflies that cause all the storms?"