Re: Paul Graham's Arc is released today... what is the long term impact?

are wrote:
On 31 Jan, 11:32, "j.oke" <java....@xxxxxxxxx> wrote:
It should suffice to redirect any question etc. to:

http://arclanguage.org/forum

Well, here's one question that does not fit neatly into
arclanguage.org/forum: really how different is Arc from CL? Having
read the Arc tutorial, I gather that:

1. Arc has but a single namespace;
2. I presume Arc has nothing like CLOS (and PG has said that Arc would
not be particularly OO);
3. Names of built-in functions in Arc have been made more consistent
and generally shorter.
4. Arc has more syntactic sugar, e.g., [... _ ...] for one-argument
functions.

While I can understand that item 1 would arouse strong feelings in
many Lispers, I'd have thought that the others would not raise
passions. The amount of debate about Arc seems to suggest, however,
that there are some more fundamental differences that I'm missing.
What are those?

The amount of debate is just an indication that people like to debate. Paul Graham is a special case because he made a lot of money with Lisp, so people assume that he is always right and has to be taken especially seriously. That's a typical assumption in the US-induced widespread neoliberal worldview of our times.

It's too early to tell how Arc will turn out. However, the Arc tutorial isn't very impressive, and doesn't even remotely live up to the hype Paul created about Arc in the past. So far, it seems like a very thin syntactic layer on top of an R3RS-era Scheme. The hardest part in implementing Arc seems to have been to ensure that nil and #f are the same. Duh.

Maybe there are more interesting things going on in Arc and we will learn about them in the next few years. But for a supposedly hundred-year language, there is still a long way to go. My guess that it will rather suffer the same fate as all languages that attempted to be much simpler than the others at first - the long-term result will probably be much more complex than if complexity had been anticipated from the start.

Of course, I could be wrong. Who knows. But Clojure, for example, has much more interesting ideas. However, Rich Hickey isn't as famous as Paul Graham, so it won't gain the same traction as Arc. That's pretty sad...

Second question: would it not have been (or still be) relatively easy
to build Arc on top of CL, just the way Qi has been?

So far it seems that it would be outright trivial to implement Arc on top of CL. For example, you could start from Blake McBride's Lisp1, or PseudoScheme, or any of the other existing hacks to get a Lisp-1 running on top of Common Lisp. Then just add a few hacks using reader macros, and define the Arc functions and macros. I don't see how much harder than it should be.

Qi is much less trivial to implement. The reason is that Qi provides an actually interesting and sufficiently different Lisp dialect, and not just syntactic sugar over known concepts.


Pascal

--
1st European Lisp Symposium (ELS'08)
http://prog.vub.ac.be/~pcostanza/els08/

My website: http://p-cos.net
Common Lisp Document Repository: http://cdr.eurolisp.org
Closer to MOP & ContextL: http://common-lisp.net/project/closer/
.

Relevant Pages

  • Re: An Arc Tangent
    ... > isn't a Lisp. ... In the context of the current topic, I suspect that SNL people would ... > in Arc or were likely to download it. ... these people would be playing language god themselves if it weren't for ...
    (comp.lang.lisp)
  • Re: Game Over: OReilly Does Lisp
    ... There are no technical advantages in Arc over existing Lisp dialects. ... Other new Lisp dialects are more worth to be considered because they actually provide new ideas. ... I've never written an program, never more than glanced over about it, but I have a hunch that it won't be a very successful language. ...
    (comp.lang.lisp)
  • Re: Game Over: OReilly Does Lisp
    ... this group is doing everything it can to find reasons not to like Arc. ... Then I went home to my lisp project, saw how these things could make life easier, and hacked a few macros so that I can write (with-uniqs res body) and body) and the like. ... Arc's UNIQ is similar to GENSYM in Common Lisp. ...
    (comp.lang.lisp)
  • Re: An Arc Tangent
    ... > language or dialect, and it sounded like minor variations on existing ... > to Lisp; but I didn't want to get into a debate about whether ... syntax and few about semantics) are discouraged with the party line ... > await Arc with eager anticipation, ...
    (comp.lang.lisp)
  • Re: Where to find good lisp critiques?
    ... he can create a language that doesn't suck. ... Arc using conventional methods. ... to see how many people think his new lisp sucks, ... > are lots of barriers to entry, and this is a problem scheme ...
    (comp.lang.lisp)