Subject: Re: CLOS: read only slots?
From: Erik Naggum <erik@naggum.no>
Date: 1999/09/09
Newsgroups: comp.lang.lisp
Message-ID: <3145859967030620@naggum.no>

* Harley Davis
| I believe that future work on Lisp should address this concern and make
| it possible to have a simple, predictable efficiency model if the goal is
| to enhance acceptability among new users of the language.

  there is an important difference between reducing inacceptability to new
  users and enhancing acceptability, and I'd say the two are not on the
  same axis: you can do the latter without succeeding in the former.

  simple, predictable efficiency models have the very obvious drawback that
  they aren't useful in a complex world.  that's why C programs (C has a
  simple, predictable efficiency model in my view -- don't know about
  yours) are frequently efficient only at the lowest and local levels and
  dramatically inefficient globally.  Common Lisp programs are often
  globally efficient and inefficient at the local level, because of this.
  that's why you can frequently improve the performance of a Common Lisp
  program greatly by tweaking a few key functions discovered by profiling.

  however, I'd settle for a predictable efficiency model.  it doesn't have
  to be simple.

#:Erik
-- 
  it's election time in Norway.  explains everything, doesn't it?