Phoenix.pm: quoting constant hash keys survey

[Scott Walters]

Yes, and there is a language, Haskell, named after Haskell Curry, too.
Obviously it's a "functional" language, and quite a good one from what
I hear. I haven't quite gotten around to taking a look at it. 

Hrm. Pascal, C, Perl, Java, PHP, C# ... I'd have to disagree with Paul. I think
Lisp is something that is rediscovered over and over. Scheme, Perl, Ruby,
Haskell, Ocaml have all rediscovered Lisp in various levels of purity and
pragmatism. 

Allow me to introduce Scott's theory of language design. Most people buy
their language from Walmart and throw the instruction manual away with
the receipt. Any time one language is billed as "easier", people gravitate
towards it. Visual Basic was probably the single most popular language
until the announcement of C# and discontinuation of VB. Perl was popular
while it was billed as easy. PHP is billed as making web programming
very easy. I think this is an impossible ideal, due to Larry's law
of cruft conservation (the language can sweep cruft under the rug by
being pure and clean and making the programmer do more work, or the
cruft can be conserved in the language itself), but just like perpetual
motion, people won't stop trying to find the "easy" language and thinking 
you've found it.

Functional programming is useful for logic-intensive applications - 
code that tries to solve hard problems. But people who know better
don't describe function languages *or* VB as "easy". VB is very painful
when you're trying to solve hard problems with it, even moderately hard
ones, and functional programming has a learning curve. So the languages
that are billed as "easy" are billed as "easy" by the ignorant, so they're
usually bad languages. C is "easy" right now because the compiler is everywhere
and a lot of people know it, but it isn't a good language for most of what
people are trying to do now days - applications and daemons. A higher
level language would be much more appropriate.

So, if people weren't forever questing for the nonexistant "easy" language,
I think we'd be migrating slowly towards different takes on functional
languages, including bastardizations of them such as Perl.

-scott

On  0, Eden Li <Eden.Li at asu.edu> wrote:
> 
> All this discussion of new Perl features. reminds me of what Paul Graham said
> in his essay about Lisp:
> 
> Over time, the default language, embodied in a succession of popular
> languages, has gradually evolved toward Lisp.
>  - http://www.paulgraham.com/diff.html
> 
> Looking at currying and some other features heading into Perl6, I can't help
> but agree with him on that point.  Any thoughts?
>