OK, I give up. I can't stand grading the undergraduates' programs
anymore. Can anyone suggest good comprehensive style guides for
Common Lisp (or older versions of Lisp) as well as Prolog that I can
just pick up and refer to or is this such a religous issue I'll have
to make one up? A good C style guide I've seen included points about
documentation, format, naming, syntactic conventions, and semantic
considerations. I also vaguely remember there being a more generic
guide on programming style akin to Strunk & White's for English.
Thanks for your help ...
Karl B. Schwamb ·······@ics.uci.edu
Information and Computer Science
University of California, Irvine 92715
In article <··················@PARIS.ICS.UCI.EDU> ·······@ics.UCI.EDU ("Karl B. Schwamb") writes:
>> Can anyone suggest good comprehensive style guides for
>> Common Lisp (or older versions of Lisp)...
Good style in Common Lisp (which is lexically scoped) may be very
different from good style for older dialects of Lisp. I think this
is part of the reason that Lisp still has a reputation for being slow.
Too many people program in Common Lisp as if it were Franz Lisp or
Interlisp (e.g. they use lots of dynamic binding, call EVAL all over
the place, etc).
On a related note, a couple of years ago I saw a syllabus (sp?) for an
"ai" course at Harvard. The teacher gave out a purported Lisp style
guide. The only gem I remember was "for small operations, use
macros instead of functions". I guess the teacher wanted them
inline for speed, and wasn't considering all the problems associated
with macros. The basic use of Lisp throughout the course seemed to
be "cons up symbols and put them on the plists of other symbols.
The symbols represent words with semantics, and that's how you do
natural language." And this was two years ago. Sigh. . .
-andrew
From: Jeff Dalton
Subject: Re: Lisp and Prolog Style
Date:
Message-ID: <1483@skye.ed.ac.uk>
In article <··················@brazil.cambridge.apple.com> ····@cambridge.apple.com (Andrew L. M. Shalit) writes:
>Good style in Common Lisp (which is lexically scoped) may be very
>different from good style for older dialects of Lisp. I think this
>is part of the reason that Lisp still has a reputation for being slow.
>Too many people program in Common Lisp as if it were Franz Lisp or
>Interlisp (e.g. they use lots of dynamic binding, call EVAL all over
>the place, etc).
I found that if I programmed for the compiler in Franz, I would have to
think in terms of lexical variables, make explicit special declarations
when needed, etc. This sort of programming style was easy to adapt to
Common Lisp.
From this point of view, what Common Lisp did was to make the lexical
scoping rules apply even in the interpreter and to extend them so that
bindings had indefinite extent (ie, provide lexical closures).
I don't know of a "style guide" per se, but you might look at the
textbook "Lisp: A Gentle Introduction to Symbolic Computing". He uses
very good Common Lisp style. For the higher level issues (e.g. why
not to call EVAL all over the place, when to use and not use macros)
he gives excellent explanations.