Larry Wall's comment on python...

[Tim Lesher]

pinard at iro.umontreal.ca (=?iso-8859-1?q?Fran=E7ois?= Pinard) wrote in message news:<mailman.1031342469.3006.python-list at python.org>...
> I take Larry Wall as a nice and clever guy.  Perl does not marry so well with
> evolution, and this is probably the doom of all languages which do not enough
> foresee evolution, and overly praise backward compatibility.  

Because the language has grown by accretion, many Perl advocates tend
to draw parallels between the evolution of a language over its
lifetime and the evolution of species.  I think that this leads to
some invalid assumptions.

Evolution of species works because of natural selection.  Many random
mutations are "tried", similar to the individual displacements in
Brownian motion.  Those that prove advantageous for a given set of
circumstances persist.  That's all there is to it--no worry about
whether this mutation is a good idea overall, or whether it makes the
organism compatible with previous generations, or whether this
mutation makes the DNA sequence easier to read or harder to fit on an
80-column screen. :-)

The moment you talk about designing for language evolution, or
maintaining backwards compatibility, or pretty much any kind of
intelligent process governing the change over time, you have to drop
the comparison with natural evolution, as enticing as the comparison
might be.

Keeping a bad analogy just muddles conversation.

-- 
Tim Lesher
tim at lesher.ws