Subject: Re: on the relations of traditional CS theory to modern programming practice
From: (Rob Warnock)
Date: Mon, 22 Oct 2007 00:52:33 -0500
Newsgroups: comp.lang.lisp,comp.lang.functional
Message-ID: <>
Matthias Buelow  <> wrote:
| People with a strong theoretical background (McCarthy, Ken Thompson,
| Donald Knuth) have created a lot better software than the
| buzzword-infested junk we often see today.
| I think if you've broken in your brain on Turing machines or other
| minimal abstractions, so to say, you have a different approach at
| problem solving than just glueing together illmatched, badly thought out
| buzzwordy libraries, which is what most programming is about today.

As much as I respect a strong theoretical background, the most
significant part of your statement might not be TMs per se, but
the bit about "or other minimal abstractions". I have generally
found that people who had backgrounds which included significant
assembly-language programming in restricted environments [whether
it was small memory, small instruction sets, and/or real-time
constraints (e.g., line-rate packet switching with PDP-8's!]
tended to also write much better code when using high-level
languages, large libraries, and the other accoutrements of
today's nearly-unlimited resources. That is, a certain tendency
towards mimimalism, together with experience in how to get real
work done in a mimimalist environment, is a useful trait even
(or even especially?!?) when the environment is not especially


Rob Warnock			<>
627 26th Avenue			<URL:>
San Mateo, CA 94403		(650)572-2607