Re: [Gnu-arch-users] OT: Lisp

[Tom Lord]

    > From: Joshua Haberman <address@hidden>

    >>> I have no desire to start a language war (honest!) but I would
    >>> like to understand what Lisp offers to the landscape of
    >>> programming languages.

    >> This topic recently came up on a lisp mailing list I'm on.

    > What list?  If it has public archives I'd like to read that thread.

I don't think it has an archive.  It's a regional list.


    >> I replied
    >> as follows, which most people seemed to think was a pretty good
    >> answer: [....]

    > That is an interesting perspective.  However it seems only to
    > address lisp-the-community as opposed to lisp-the-language,
    > except for the implied idea that lisp-the-language is flexible
    > enough to allow this kind of language experimentation.

    > It also leaves me with the impression that lisp is primarily the
    > domain of language tinkering as opposed to being a productive
    > tool, especially if the language is constantly morphing.  This
    > cannot be your assessment though, or you would not be planning
    > to use it for itla.  Why does lisp's strength as a language
    > prototyping tool/community make it the right choice for
    > implementing a program like itla?

Who said anything about "language prototyping"?

Ok, you are missing some history.  Lisp has been, as I said, forked
and hacked, and recombined -- going back somewhere between 40 and 50
years now.

In _some_ of those cases, the work going on was just "language
research".   In _many_ cases, the work was "making some application
work better" or "making some community of users working on
applications more effective."

By the time you pick up Scheme or Common Lisp or even Emacs lisp, the
language you're working on has both developed and inherited a lot of
experience and deep thinking.

Another way to say it is: you should bother to learn some mainstream
lisp because I can point to a gaggle of people whose hard work
informed the design of that lisp, any one of whom is at least as good
as one Larry Wall or Guido van Rossom.  In a fistfight, my guys
wouldn't even have to try very hard.   It's no one detail -- it's many
woven together.  (Yes, yes, many people have contributed to Perl and
Python, too -- but you take my gist, I hope.)

The thing is (and switch to comp.lang.lisp if you want to join the
fun) people have non-terminating debates on issues like: "resolved:
python is just as good as lisp".  And in these debates someone says
"well, common lisp has incredible macros, or scheme has real closures
and call/cc, or whatever" and then someone says "well, what's it good
for" and then someone says "well, for example, X" and then someone
says "Yeah, but I can do X in python this other way" and then someone
says "yeah, but now try to do X' -- in lisp that's just the following
trivial change" and then someone says "ok, sure, i can't do that as
conveniently but I certainly _can_ do X' and, anyway, how often do you
want do do X'? [as often as you find it both useful _and_ easy -t]"
.... and this goes on for months and months and years and years for no
good reason whatsoever.

The good lisps are a lot of good stuff, put together really well. The
popular scripting languages tend to be crude by comparison, though
highly specialized for a relatively small set of tasks.

Learn lisp.   It's good for your brain.

I'll leave you with something someone sent me recently, trimming out
the attribution:

  Tom> I like to make the joke that Python is (as the name suggests) a
  Tom> humorous production: the punch line is that it eventually turns
  Tom> into MIT Scheme.

  > Actually, this property has been quite benificial to myself. I
  > somehow tuned out Lisp/Scheme during my freshman year because i
  > thought it was ugly and nobody i was in contact with at the time
  > could tell me why it was interesting. Shortly after i started
  > using Python extensively, and got that in my blood and experianced
  > why doing things in certain ways is much more powerful and
  > productive.

  > Much more recently between some threads on [XXXX], and reading
  > too much on [XXXX], i've been able to make the inductive
  > leap to seeing that Python is just Lisp watered down a lot for
  > those who can't buy into the truth on first sight.
  
  > So joke, sure, but its really quite a useful joke: it's possibly
  > saved me from falling for all the pathetic java/xml/c#/hype
  > propaganda.

Go learn lisp and stop sparking language flames :-)

-t