Re: What is a lightweight language

[Paul Graham <address@hidden>]

I'm well aware that a language has to have an app to
drive it.  See Being Popular 

  http://www.paulgraham.com/popular.html

which if anything overstates that case.  BP was the 
initial blueprint for Arc, and suggests the kind of app 
we plan to launch it with.

I don't want to say a lot about the app till there is
something out there for people to use, but we are
already writing it, even though large parts of Arc
don't even exist yet.  That's how important it is
to the project.  (Plus I think it's a good idea to
write programs in a language while it's still malleable,
so you can be sure it feels good to hack in.)

--pg

--- Lennon Day-Reynolds <lennon@day-reynolds.com> wrote:
> On Wednesday, December 12, 2001, at 05:44 PM, Paul Graham wrote:
> 
> > Mike raised this question at LL1. The answer is OS
> > connectivity and libraries, esp. string libraries.  People
> > need those to write actual programs. Common Lisp and Scheme
> > don't have them.  (Maybe some implementation thereof does,
> > but the standards don't).
> >
> > For a big project like Viaweb, esp. when you have
> > competitors snapping at your heels, it can be a win
> > to use these more powerful languages.  But for smaller
> > problems (and all problems start small) the best libraries
> > win.
> > [...]
> > Lisp has a lot of latent power though.  The hypothesis of
> > Arc is that all you have to do is add OS connectivity
> > and the usual libraries to Lisp and you have a butt-kicking
> > language.  (You may in fact get what Perl and Python become
> > as t approaches infinity, modulo syntax.)
> > [...]
> 
> This makes a lot of sense, but there *are* Lisps that have these 
> kind of practical libraries in the core...PLT Scheme, Rep, and 
> Guile all spring to mind, without really thinking hard. On the 
> other hand, when Java was first released, it had a pretty miserable 
> set of string-processing libraries, and very little in the way of 
> system integration beyond basic byte-streams and JNI. Sun kept 
> pushing it to businesses, though, and the libraries gradually 
> accumulated.
> 
> I think the "killer app" argument works here, too, even if it's 
> just an artifact of hype. To "average" programmers, Lisp is a 
> language for AI, not for modern application development. Perl's 
> high-visibility win has been CGI (despite the fact that CGI scripts 
> can be written in almost any modern language), Python has Zope and 
> great support for beginning programmers (again, not necessarily 
> unique except in perception), and Java has J2EE and XML libraries 
> (ditto).
> 
> So, if you're not going to dispose with the perceived dominance of 
> any of these languages, what is the niche that Arc (or any new push 
> to popularize Lisp) is going to fill? Can you really predict it, or 
> is it just a matter of dumb luck to be in the right place, six 
> months before (or a better-marketed) than the "also-rans"?
> 
> Lennon Day-Reynolds
> 


__________________________________________________
Do You Yahoo!?
Check out Yahoo! Shopping and Yahoo! Auctions for all of
your unique holiday gifts! Buy at http://shopping.yahoo.com
or bid at http://auctions.yahoo.com