A modest proposal (was: Comment on PEP-0238)

Quiptic quiptiq at yahoo.com.au
Thu Jul 12 12:15:26 CEST 2001


"Alex Martelli" <aleaxit at yahoo.com> writes:

> "Guido van Rossum" <guido at python.org> wrote in message
> news:cpr8vnh6ln.fsf at cj20424-a.reston1.va.home.com...
> >
> > Good point, Art.  I trust my intuition more than my reasoning in these
> > things, and often when I try to defend my intuition against
> > challenges, what comes out are unsatisfactory rationalizations.
> > Sometimes Tim's "channelings" say it better than I could have said it
> > myself...  :-)
> 
> Being a great designer (or other kind of decision-maker) and
> being a good rationalizer/evangelist/salesperson/marketeer are
> really separate skills, and pretty much uncorrelated ones to
> boot.  

Indeed. Intelligence and talent often go together, but they're not the
same.

I think language design is the one area of CS in which "taste" matters
most. The ability to see solutions that are "good", clean, simple,
elegant, clear, concise, etc apparently comes from a deeper (or at
least different) level of awareness than the ability to rationalise
after the fact.

Of course, there's always the remote possibility that a buffoon is
fated to produce an accidental masterpiece ;-)

> People in general just won't accept that, which of course suits us
> good rationalizers just fine (as we often get credited for better
> decisions than we actually make, just because we rationalize better
> than others do:-).

Sure. It's like the distinction between artists and critics in other
fields. It's reasonably easy to work from insight down to
rationalisation / realisation, but I've never yet seen it work the
other way. There is no recipe for quality; we recognise it _after_ we
see it. (I think that even applies to the person who creates it).

I've heard gifted musicians and painters sound like complete idiots
when they try to explain their work. I've also read the work of
intelligent critics who genuinely add to a work of art by their
thoughtful interpretations, even though they could never have produced
the work themselves. Plenty of that here on c.l.p. It creates a nice
synergy.

Whether Guido's explanations are satisfying or not (sometimes yes,
sometimes no), I'll always trust his judgement. I'm grateful that
Python is not mine. If it were mine, it would be one fucking ugly
mother.







More information about the Python-list mailing list