🚀 go-pugleaf

In article <1991Jun18.042950.4473@netcom.COM> jls@netcom.COM (Jim Showalter) writes:

   Sorry. I retract my statement that when it comes to compilers you
   get what you pay for, at least as it concerns the particular case
   of GNU. I've received nothing but praise (usually
   stated quite angrily) via e-mail for the GNU family. I find this,
   by the way, to be totally bizarre, since it seemingly flies in
   the face of 6,000 years of free-market economic theory. Why something
   free would be quite bug free and would have great support in contrast
   to something that costs money is completely beyond me...

Free software is a public good.  Public goods are not supplied by
private parties except under certain conditions.  One of these is when
the private party encompasses the entire public (as in a lake
improvement association where there is no public access to the lake).
Of course, then it's not a public good, it just looks like one.

Another case is when the private benefit is larger than the private
cost.  E.g. a lake improvement association where there is a small
amount of public access.  The Free Software Foundation derives more
private benefit from the distribution of free software than it costs
them.

   Why something free would be quite bug free and would have great
   support in contrast to something that costs money is completely
   beyond me--were this true in general one would expect cheap cars to
   be better than expensive ones, and free cars to be best of all.
   Anybody care to enlighten me?

The incremental cost of producing a car is fairly high.  A lot of
labor and raw materials goes into each car.  The incremental cost of
producing software is very low.  Shareware distribution companies make
a good bit of money copying floppies for $5 each.

There are three costs of software: writing it (and getting it right),
changing it to meet different needs, and getting it to work under differing
circumstances.  The first of these is very low, and it is the only one that
the FSF feels responsible for.

   >It's been my experience that the vendors of compilers *never* share their
   >bug list with customers.

   The compiler vendor I used to work for not only published known
   bugs but known workarounds as well. If this is not common practice,
   it jolly well ought to be.

You can make a reasonable case for my experience being incomplete.  :)

Let me make a different and weaker claim, that the vendors of
compilers *marketed to nonprofessional programmers never* share their
bug list with customers.  They are under the impression that a large
part of their market would shun their product if they knew it had bugs.