[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: another take on hackers and painters

An interesting choice for an analogy: I've never lived in Las Vegas,
but I'm from a town not too far from it and I know that it's a
pretty deceptive city.  If you ignore the Strip (and the ring of
pawn shops that surrounds it) Las Vegas is actually a surprisingly
nice family community (it's most easily comparable to Utah, with
which it has common roots).  The impression that Las Vegas leaves on
visitors is usually quite strong (as it seems to have been in your
case) but it's rarely very accurate.  The casinos are so prominent
and so well known that most people assume that they characterize the
entire community.

This is the point where I'm supposed to triumphantly overturn your
argument using your own analogy, but I think that would be misguided
rhetoric: I think that ML's type system *does* fairly characterize
the whole language.  I'll take exception to your example, but not to
your point.

- Russ

On Tue, May 20, 2003 at 03:23:45PM -0000, Paul Graham wrote:
> That's like saying that people who criticize Las Vegas
> should annotate their criticism with how much time they've
> spent there.  I've spent about two days in Las Vegas, but 
> I  feel I understand *perfectly* what the place is about,
> and that it's not for me.  Ditto for ML.
> --pg
> --Shriram Krishnamurthi wrote:
> > I'm afraid your article, while interesting, simply reiterates tired
> > and (I think) often invalid prejudices against static typing.  Though
> > this isn't the focus of your article, it's not that far removed from
> > the focal point, and it anyway gives me a chance to vent. (-:
> > 
> > It is absolutely true that in the extreme case, the "artistic
> > programmer" needs a language without type restrictions.  This is
> > because they are effectively building a non-trivial language
> > themselves, whose type system may not mesh well (or at all) with that
> > of the native language.  Even then, it's often possible to encode the
> > second tier type system atop some universal representation in the
> > first one, but one could fairly ask, to what end?
> > 
> > On the other hand, in a language with a good static type system, I
> > think programmers run into these restrictions far less often than they
> > might think.  A sufficiently crafty programmer will recognize when
> > they are hitting the language's type restrictions, and will then be
> > able to provide a lovely critique of the type system along with the
> > application that is shackled by it.  There are surprisingly few such
> > critiques, and the ratio of critiques to impassioned pleas for freedom
> > grows vanishingly small.
> > 
> > People who criticize static typing should annotate their offering with
> > an indication of how much ML they have written.  That would make them
> > far more credible.
> > 
> > I suppose I've now tossed down the gauntlet, and you can always come
> > back with "Oh, about a 100k lines, at which point I realized static
> > typing was all bunk", and nothing would make my Scheming heart
> > happier. (-:
> > 
> > Shriram