Subject: Re: free lisp compilers?
From: Erik Naggum <>
Date: 1999/09/05
Newsgroups: comp.lang.lisp
Message-ID: <>

* Marco Antoniotti
| All in all this are not "language inherent" problems.  I suppose you
| could expect an improvement for CL if many more programmers were working
| at it.  I.e. it is a matter of "scale" and invesment.  I would venture
| out to say that it is reasonable to think that CL compilers have fallen
| behind C/C++ compiler technology in recent years, w.r.t. the situation
| of, let's say, 10 years ago.  Of course I have no real data to support my
| case.  It is just a hunch.

  why even say this if you can't back it up?  don't you realize how much
  this reinforces the stupid myths?  obviously, you speak about CMUCL from
  experience, but why not confine your comments to CMUCL?  sheesh.

  the biggest problem is that disproving your claims or undoing the damage
  is very hard work.  to win an argument with somebody who believes you, CL
  compilers not only have to be as fast as C/C++ compilers, but faster,
  with better machine scheduling and register usage and whatnot, otherwise
  your negative point will still hold: "CL lags behind C/C++".

  I suggest those who care work on ways to improve every single instance of
  suboptimal code they can find.  if there is a unique preferred order of
  some instructions, for instance, notify the compiler writers.  you can't
  expect them to know _everything_ about every processor, even though you
  may think that's what they get paid to do.

  (and with all the source access and everything, why not just fix it?)

  save the children: just say NO to sex with pro-lifers