Prev: COMPILE-FILE tertiary value, failure-p
Next: Some thoughts on extensible sequences
From: ccc31807 on 2 Aug 2010 11:18
On Aug 1, 11:00 am, D Herring <dherr...(a)at.tentpost.dot.com> wrote:

> FYI, Nick had to drop the project due to poor health.  A few chapters
> are available online.http://lisp-book.org/contents/

I have those eight chapters, and have read two. Over the years I've
purchased a large number of O'Reilly books. Many haven't really been
helpful, but the ones that are helpful are really good. I have several
that I've had to re-buy over the years because of wear and tear from
(ab)normal usage.

My general impression is that the O'Reilly books are by professionals
for professionals, with minimal hand holding. The only CL book that I
have seen that approaches this is Keene's book on CLOS. I'm sorry for
Levine's health problems, and I'm sorry that the project is on hold
(but he has indicated that the project isn't totally dead -- O'Reilly
still will publish the book if someone will take it over.) I thought
that sponsorship by O'Reilly was pretty much a guarantee of high
quality.

I would like to make this observation on Lisp and the popularity of
languages.

Developers, managers, and large companies all say that they want to
use the 'best tool' for the job. In my job, I use a lot of Perl -- I'm
not a programmer or a developer but a database guy and Perl is ideally
suited for the data munging tasks I do every day. I have contacts with
both IT and non-IT managers at large corporations in my area, and they
use a lot of C# and Java because those technologies make managing
large projects and large staffs easy. I have a friend that owns a
company that writes embedded systems using C, and he will only use C
because (as he says) he has to stuff complete applications into only a
few K of memory and C is the only high level language that will do
this.

The reason that CL isn't wide spread is because it is perceived as not
being the 'best tool' for jobs. This perception has a basis in
reality, regardless of the merits of CL as a language. Unfortunately,
being the 'best tool' requires things other than pure programming
power.

All things being equal, the most powerful language should rise to the
top, but all things aren't equal. Ultimately, languages live in an
environment where survival depends on many different qualities. The
apex predators on the North American continent, like the saber tooth
tiger and the short faced bear, didn't survive - not because they
weren't the top predators, but because the environment favored things
other than sheer killing power.

If CL is to survive and prosper, it needs to evolve to the point that
it can prosper in its environment, i.e., earn the perception that it
is the 'best tool' for some kinds of jobs. If it doesn't do this,
other languages will take its place, perhaps F#, or Clojure, or Go, or
maybe even Perl (which already has many Lisp-like features.)

CC.
From: Thomas A. Russ on 2 Aug 2010 12:00
"Captain Obvious" <udodenko(a)users.sourceforge.net> writes:

> ??>>> What's about PAIP?
>
> ??>> It's an excellent book.
>
> I concur.
>
> TKP> Indeed it is, but it does a lot more than discuss CL per se: it gives
> TKP> an intro to AI. While that is interesting in its own right, it might
> TKP> be a distraction for some people who just want to learn CL and don't
>
> One can read only the first part...

I think this depends a bit on who the student is. I've tried that with
some of our graduate students and the lisp introduction was a bit too
condensed for them. (That was a shame, since I liked the quick,
no-nonsense approach.) I suppose that without much experience outside
the C/Java world, one might need a bit more explanation of the lisp
model of programming than one would get from PAIP.

--
Thomas A. Russ, USC/Information Sciences Institute
First  |  Prev  | 
Pages: 1 2 3
Prev: COMPILE-FILE tertiary value, failure-p
Next: Some thoughts on extensible sequences