Subject: Re: NT and Linux Implementations of Lisp
From: Erik Naggum <erik@naggum.net>
Date: 16 Oct 2000 05:27:17 +0000
Newsgroups: comp.lang.lisp
Message-ID: <3180662837430170@naggum.net>

* "David Allen" <s2mdalle@titan.vcu.edu>
| As for NT, you can also possibly get a copy of  Allegro Common Lisp
| through www.franz.com.  It's a very good lisp, but be warned - the
| trial version isn't all that fabulous, and the full version costs
| several thousand dollars AFAIK.

  If you were interested in helping anyone along, would it be a crime
  to give enough information that others might be able to come to a
  conclusion of their own instead of having to believe your conclusion
  with no clue what it is based on, how _you_ arrived at it, what you
  wanted and got and didn't get, or how you rate other things?  For
  all we know, you're one of those people who think all software
  sucks, so "isn't all that fabulous" could be the best thing you have
  ever said about any piece of software.  There's no telling with only
  s aingle data point to speculate from.

  The Trial Edition offers more than the Professional Edition does,
  but it has a reduced heap size, and the clear intent is to avoid
  abuse while allowing reasonable trial use.  Abusers are good at
  making goodwill seem wasted on the community and they make it much
  harder for those who would like try something out on a good faith
  basis be able to do so.  Therefore, there have been serious changes
  in the way the Trial Edition has worked, and with the next release,
  the program that Franz Inc has always had wherein you would present
  a case for purchasing a license and be rewarded a trial license on a
  case-by-case basis (so you could actually try to do _some_ of the
  stuff you would purchase a license for) would be extended and made
  more automatic.  Giving away access to software that otherwise is
  the mainstay of your business involves a difficult set of trade-
  offs.  I'm not sure it makes sense to complain about what you get
  for free.  More productive is trying that novel approach in this
  electronic day and age and _send_them_mail_ and communicate your
  needs and desires.

  But yeah, the pricing is horrible.  To be _honest_, though, it gets
  much worse than just the software: The programmers and developers
  who are going to use it cost _many_ times what the software costs.
  As long as you're into this programming business for its commercial
  value (and more and more IT and Internet businesses have come to
  their senses and face the need to look at the commercial value of
  their operations), you might as well realize sooner rather than
  later that a programmer who wastes a couple weeks of company time a
  year because of inferior tools costs more than one who gets a
  license for Allegro CL and wastes that much less.  But, again to be
  _really_ honest, since all programmers piss away a lot of time in a
  year, the issue is whether we can reclaim some of that time with a
  better tool and then the value of that programmer's time increases
  in ways that are hard to measure.  Good tools can often make as big
  a difference as a factor of 5 on some programmers, especially if
  they learn to appreciate it and don't try to reinvent it all the
  time, which often happens to good programmrs using inferior tools.

  However, if _your_ time is free, too, _do_ go for the free tools.

| CLISP is free as in speech and as in beer.

  It's also a great toy and much can be done with it, but as with all
  toys, don't get too attached to it if you plan to move on to a
  production systems.  In particular, don't ever time your code in
  CLISP -- the timing statistics are so skewed in favor of using only
  built-in operators and so skewed in disfavor of writing anything
  advanced of your own that CLISP users learn how to optimize their
  code by remaining at the primitive, builtin level.  I have seen good
  programmers piss away months trying to get CLISP to run fast enough
  for real use, when they could at least have moved to CMUCL and get
  some reasonable compilation, were it not for marrying themselves to
  parts of that system, and not even learning to use the -a option to
  make their code an order of magnitude more portable among modern
  Common Lisp implementations.

#:Erik
-- 
  I agree with everything you say, but I would
  attack to death your right to say it.
				-- Tom Stoppard