On Tue, 2008-02-26 at 14:28 -0800, David Brown wrote:
> I've mostly been using the pre-release of 1.2 of Clozure Common Lisp
> (formerly OpenMCL).  It has the same kind of feel that Allegro or Lispworks
> has as far as the basics and the core working.  With the others, I seem to
> run into a lot of On Tue, Feb 26, 2008 at 10:33:24AM -0800, Tracy R Reed 
> wrote:
> >> There aren't a lot of Common Lisp implementations, at least not that are
> >> still around.  For the most part, people seem to choose either Allegro or
> >> Lispworks.  Both are basically too expensive for an individual, hobbyist
> >> developer.
> >
> > Why would they choose Allegro or Lispworks over SBCL or CLisp or CMUCL? 
> > (any other Common Lisp implementations I'm missing? I don't know if this is 
> > a lot but it's more than a couple.)
> 
> Because they're a lot more mature.  Unless you're using Debian or Gentoo,
> it can be quite a pain to even just install/build SBCL or CMUCL.  Clisp can
> be too slow.
> 
> auxiliary stuff that just doesn't quite work right, like
> debuggers and stuff.

I haven't played with Colzure, but the last time I looked at SBCL
everything seemed to work right. Of course, I was using Gentoo
soo... ;-)

> > I hadn't thought about all of the C implementations before but you are 
> > right about there being a lot. ANSI C would seem to be a smaller standard 
> > than the CL standard isn't it? Perhaps it is easier to make conforming 
> > implementations and compatible libraries?
> 
> C is a much easier language to implement.  C++ on the other hand, probably
> not much different.  But C++ is a lot more popular, so more possible
> revenue to motivate people to work on compilers.

C++ is possibly the single largest PITA to make a standards compliant
compiler/library combination. It is inarguably harder to parse than
LISP, and while its standard library is more limited, the specification
of it is remarkably baroque and precise. Finally, the performance
expecations for C++ compilers are extraordinarily high. Optimizers are
expected to keep up with C compilers despite the extra man power needed
to implement C++. In fact, even the *commerical* interests initially
rejected a lot of the things that are in the C++ standard simply because
they felt they'd require too much effort to implement. This was overcome
by open source reference implementations being made available that
companies could use for their own implementations. One would think
similar challenges and solutions would have presented themselves to the
LISP community.

> For some reason, I think people are also turned off by the incremental
> development.

That's a new one.

>   Somehow, they associate incremental compilation with an
> interpreter, and think that it will be slow.

Do you realize that MS VC++ has been working, for years, on supporting
improving incremental compilation and linking, to the point where you
can almost do C++ development with a compile-link-test cycle as quick as
an interpreter?

--Chris

-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Reply via email to