[Haskell-cafe] Re: Libraries for Commercial Users

[Curt Sampson]

On 2009-10-09 20:53 -0600 (Fri), John A. De Goes wrote:

> The vast majority of applications being built today are web apps.

Right, and the vast majority of these are done in PHP, followed by Ruby
and Python. For these languages, the majority of libraries are not native,
but C libraries.

So this makes me wonder, if cross-platform support is so necessary,
why have these PHP, Ruby and Python folks not switched to Java, rather
than remain suffering doing their development on Linux boxes?

And if libraries are the issue, why would not just creating a SWIG for
GHC make all of these people move to Haskell?

On 2009-10-08 09:50 -0600 (Thu), John A. De Goes wrote:

> I don't dismiss Haskell in business. I only maintain it's a niche  
> market.

I agree it's a nich market, too. So I guess our point of disagreement is
that you believe that it's lack of libraries keeping it a nich market,
whereas I believe that even if it had libraries out the wazoo like Java
(and of similar quality) it would still be a nich market. In other
words, having a relatively small quantity of libraries is not a problem.

> There are some domains where the infrastructure in more established  
> languages is minimal, and in such cases, I think Haskell can be more  
> efficient than those languages.

Indeed. I found many domains (including web ones) where I didn't use
that infrastructure (such as it was) in Ruby, either.

Where there are libraries and frameworks that are considered
signficantly better, such as Rails over what's available in PHP, people
switching are switching for incremental improvements in what they
already do; these are not the kind of people who would ever switch to
Haskell. (Rails is beloved basically because it's the bastard spawn
of a PHP programmer and web designer; it's certainly not a framework
that offers any big difference over doing the same thing in PHP in the
traditional style it's always been done there.)

> What has moved [from the agile development moment] into mainstream is
> unfortunately connected chiefly to agile by virtue of the word itself.
> Agile means more than getting software out the door quickly, a fact
> many businesses have yet to learn.

Right. But that practically defines the mainstream: adopting
misinterpretations of truly radical changes so that they don't really
have to change. If you want Haskell to *really* go mainstream, that's
what you want. If you want libraries for Haskell that will help it go
mainstream in that sense, you want bad ones that offer no significant
improvement over the bad ones already out there for Java or whatever,
and thus provide little productivity improvement.

That just seems pointless to me, because those looking for real
improvements can't use that sort of stuff. When I need RDBMS access,
libraries like Hibernate and Active Record are useless to me, because
they force me to work in a stupid manner. For web sites, Rails is
useless because again, it deals with stuff in a stupid, unproductive
way. This is why I say libraries aren't important. True, it is nice to
have good ones that really help you be productive, but not only does
Haskell not have them, neither Java nor Ruby do, either.

cjs
-- 
Curt Sampson       <cjs at starling-software.com>        +81 90 7737 2974
           Functional programming in all senses of the word:
                   http://www.starling-software.com