[clue-talk] Stupid OCaml tricks

[Sean LeBlanc]

On 08-14 14:30, Nate Duehr wrote:
> 
> On Aug 14, 2006, at 12:55 PM, Matt Gushee wrote:
> 
> Maybe it is best summarized this way: If two programs produce the  
> same inputs and outputs and are written in two different languages...  
> Who cares?  Only the programmer.  The end-users don't care about what  
> it's written in, and never will -- other than some end-users who are  
> developers and have their "pet" languages they like to "support" by  
> using things written in that language.

I understand what you are trying to say, but I think you're leaving out the
human element. Well, and arguments like Paul Graham's "Beating the Averages"
in which he argues some languages simply *are* better than others - in his
case, it's Lisp.

The human element - if people are reinvigorated and more productive in their
language of choice, or knowing another language helps you see a problem a
bit differently in the language you have to pay the bills with, that has value.
Sure, everything could be done in assembler, and developers who have a deep
understanding of assembler on up are not going to be bozos, but at the same
time, I'm sure anyone trying to kick off a large business software in
assembler would be fired. Also, bit twiddlers can sometimes get hung up on
things that no longer matter, instead of focusing on the business rules.

Even though I find the RoR hype is getting really tired, I've heard people
say that Ruby makes programming "fun" again (Dave Thomas for one). Well,
that obviously has value to someone - this may pay off in productivity.
Which sounds great for smaller teams and projects, but I cannot understand
why so many Java folks are ready to throw years of enterprise-level stuff in
Java overboard to embrace RoR, at least for large projects.  

Okay, that's more the human element - the fun part, the novel stuff. Then
there is the Paul Graham argument - some languages are way ahead of others.
Yes, you can have leaky abstractions like Joel talks about (he's in my RSS
reader). However, people who do things like Lisp apparently think about
things differently (not a Lisp-er myself) and do things like make DSL -
domain-specific languages. Not that it couldn't be done in other languages,
I suppose, but Graham is also claiming that Lisp actually has features that
others are only now catching up to. He comes down on Java pretty hard, and
even though I get a bit irritated that someone would do that who hasn't
actually used it, I do think he has some pretty good points in his writeup. 

OTOH, I do understand the aversion to the new, shiny object. I pay the bills
with Java, and man, the churn can get boring. Sometimes, something is
genuinely interesting even if a little overhyped (to me, anyway), like, say
AJAX. Other times, it's something like SOA and we're supposed to get all
jazzed about it. Yawn. Isn't SOA just web services? And Java itself is a
decent language for paying the bills - but the hype in the early years with
all the applets - blech.

In any case, I'm all about the novelty, so I may have to take a look at
OCaml, as I've heard Matt talk about it before, but missed this
presentation. Now if I could just find a job where I could pick up Lisp and
use it everyday...I do have a basic install set up and have been poking
about, but very much in the crawling stage at this point. :)

-- 
Sean LeBlanc:seanleblanc at comcast.net  
http://sean-leblanc.blogspot.com/