Subject: Re: A Philosophical Diversion
From: Erik Naggum <erik@naggum.no>
Date: 1998/10/09
Newsgroups: comp.lang.lisp
Message-ID: <3116914266492853@naggum.no>

* trashcan@david-steuber.com (David Steuber "The Interloper")
| That isn't easy to do.  He knows the internals of C++ and Unix to an
| extent I would not have thought humanly possible.  He also is very good
| at architectural design and analysis.  This guy is brilliant.  However,
| he is not infallible.  I have been able to code some algorithms with
| slightly better efficiency :-).  OTOH, he has written compilers and
| implemented portions of Unix.  That is far and away more than I have
| done.  He also knows a lot more about features of Emacs than everyone
| else at the company.

  just because he has optimized himself for one mind-set doesn't make him
  even able to claim anything about any other mind-set for which he has
  deliberately _pessimized_ himself.  the odd thing about optimization in
  one direction is that it costs much more than it could ever gain on the
  whole, but that _one_ direction of optimization has been perceived as
  valuable enough to absorb the costs.  (I'm talking about what humans do,
  not machines.)  the result is extremely hard to optimize for something
  else, which was obviously possible _before_ going to the far end of the
  one optimization direction that was chosen.

  (I used to be a C/Unix guru, and I still get called upon to give courses
  in C and presentations on the history of Unix and Unix internals and
  such, but I had gotten bored, and C++ appeared to be the future at the
  time, so I _had_ to find something else.  I had programmed toys in Lisp
  on and off since late seventies, so CMUCL became the savior of my sanity,
  and I spent three years studying and working with it before I felt I was
  sufficiently competent to write Common Lisp applications for commercial
  clients.  so I made living writing articles for newspapers and computer
  magazines, instead.  I don't know anybody who doesn't think this was a
  high prize to pay to change course, but life is too long to want to be an
  expert in C++.)

#:Erik