I am just starting out

K. Fritz Ruehr fruehr@willamette.edu
Mon, 26 May 2003 14:50:50 -0700 (PDT)


Sarah Thompson (sarah@telergy.com) wrote:

>
> I hate to have to say this for all kinds of reasons, but what the FP
world *really*
> needs is for something called 'Microsoft Visual Haskell' to ship as
standard with
> some new future version of Visual Studio, and be supported by a big,
expensive
> marketing effort targeted specifically at corporate managers. Particular
attention
> would have to be given to interoperability with other languages, so
managers could
> be convinced that they had a way out.
>
> ... [various industry-attitude "horror stories" snipped out] ...
>

I appreciate Sarah's dose of reality here: I've seen plenty of the sort of
thinking she describes during the (short, infrequent) periods I've crept out
of my ivory tower and into the real world. Of course with the .NET support
of Haskell and the F# effort at Microsoft, we may ultimately get what she
describes (wasn't there even an internal MS Visual Haskell at some point?).

But before we accept "what the FP world *really* needs", I think we ought to
recognize the assumptions behind this claim, i.e., the statement should
perhaps read:

	"IF we want Haskell (or FP more generally) to be widely accepted in
	 the business world, THEN what the FP world *really* needs is ..."

And, personally, I'm not sure I *want* to see FP be widely accepted in the
business world. I suppose this is some sort of heresy, given the seemingly
universal desire by all language adherents to see "their" languages grow
to wider and wider acceptance, but I rather like the close-knit feel of
the community as it is, and I cherish the set of design principles that
community shares.

Do we really want every snake-oil salesman around hawking Haskell seminars
at $2K per weekend to stuffed auditoriums of business people, showing them
how to write imperative code in a monad to do things that might be better
done with map or filter? Could we accept an ISO standard that elevated
unsafePerformIO and such-like to a new level of acceptability?

I think it was John Launchbury who made a case at the Pittsburgh Haskell
Workshop for letting Haskell stake out a certain extreme on the language
design spectrum, one which supports formal reasoning through functional
purity; I think this is a noble purpose the language serves. We even
seem to be having some influence on the mainstream of language development:
the recently-announced enhancements to Java read like a subversive attempt
to inch that language toward FP style (and, hmmm, there is Martin Odersky's
name on some of the Expert Group listings :) ); see
<http://java.sun.com/features/2003/05/bloch_qa.html>.

Of course, it would be nice to have more "warm bodies" to devote to projects
like standard GUI libraries, browser plug-ins, IDEs, language inter-
operability, etc. (no sleight intended to current efforts). But I think this
is a question of getting more of the *right* people involved, rather than
spreading the word to the business masses. Personally, I try to help in this
effort by teaching my (liberal arts undergrad) students about FP and its
sensibilities and then steering the best of them to compatible graduate
schools. It may not be much, but I think quality is more important than
quantity (besides, with recent FP class sizes of 6 and 9, I really can't
contribute much quantity anyway! People like Manuel Chakravarty can probably
afford to do both, see <http://www.cse.unsw.edu.au/~chak/teaching.html>).

So, call me an elitist and bring the torches to my (ivory) tower door,
but  I have to question whether *broad* acceptance is what FP (or Haskell)
really needs.

  --  Fritz Ruehr