[Haskell-cafe] Why Not Haskell?

[Udo Stenzel]

Jason Dagit wrote:
> On 8/4/06, Donn Cave <donn at drizzle.com> wrote:
> >6.  Instability - available for 15 years, you say, but does the Haskell
> >    of 15 years ago support today's programs?  Does standard Haskell
> >    even support today's programs?

Uh, this one's wrong.  Does C++ of 15 years ago support today's programs?
No.  C++ of 10 years ago probably does, but the compiler will crash.
Similar for C, similar for Perl, and the question cannot even be asked
for Java.  So no, that's *not* the reason for low acceptance.

The right question would be, does Haskell of today support the programs
from 15 years ago?  Mostly it does, with minor changes.  The bitrot
isn't worse than in other languages.

> 15. OO is now tried and true in industry.  I would say it's far from
> optimal but people do know they can build large applications (say
> ~100k lines of C++).  So naturally shifting to a new paradigm will
> meet resistance.

Closer to reality is:  People know that *some* 100 kLoC OO programs in
C++ have not yet crumbled under their own weight.  However, most have.
Is that track record worth imitating?  Well, "decision makers" seem to
think so...  Anyway, *some* 10 kLoC Haskell programs are also still
standing upright (no, I didn't forget a zero).

 
> Yes, trying to meet deadlines with untrusted tools is scary business.

...and so is trying to meet deadlines with trusted tools.  Paraphrasing
Paul Graham: if you do what everyone does, you get average results.  The
average result is blowing your budget by a factor of 2 (or 3), missing
your deadline (twice) and delivering a broken product (if anything).

> Not everyone likes to gamble with their jobs :)

Yet everyone does...


Udo.
-- 
The most happy marriage I can imagine to myself would be the union of a
deaf man to a blind woman.  -- Samuel Taylor Coleridge
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://www.haskell.org//pipermail/haskell-cafe/attachments/20060804/d161652d/attachment.bin