The Importance of Terminology's Quality

Robert Maas, m6d04a5.3.calrobert at spamgourmet.com
Thu Jun 5 07:39:22 CEST 2008


> From: dkco... at panix.com (David Combs)
> Lisp is *so* early a language (1960?), preceeded mainly only by
> Fortran (1957?)?, and for sure the far-and-away the first as a
> platform for *so many* concepts of computer-science, eg lexical vs
> dynamic ("special") variables, passing *unnamed* functions as
> args ... maybe is still the only one in which program and data
> have the same representation -- that it'd seem logical to use it's
> terminology in all languages.

Yeah, but why did you cross-post to so many newsgroups? Are you
trying to run a flame war between advocates of the various
languages? (Same accusation to the OP moreso!)

> From C is the very nice distinction between "formal" and "actual" args.

I think Lisp already had that nearly 50 years ago. Function
definition (lambda expression) has formal args, EVAL recursively
calls EVAL on sub-forms to create actual args and calls APPLY on
them and whatever function is named in the CAR position of the form.
Whether anybody bothered to use that specific jargon, or it was
just so obvious it didn't need jargon, I don't know.

> And from algol-60, own and local -- own sure beats "static"!

Yeah. But now that you mention it and I think about it, what's
really meant is "private persistent". Global variables are public
persistent. Local variables and formal args to functions are
private transient (they go away as soon as the function returns).
but OWN variables are private to the function but stay around
"forever" just like globals do, so that side effects on the OWN
variables that occurred during one call can persist to affect the
next call. Lexical closures in Common Lisp go one step further,
allowing private persistent variables to be shared between several
functions. All those functions share access to the private variable
which they co-OWN. Another way in which OWN or lexical-closure
variables aren't like what the word "own" means in ordinary
language is that it's possible to transfer ownership by selling or
giving something to somebody else, but not with OWN variables or
lexical-closure variables. So even though I like the word OWN
better than the word STATIC for this meaning, I'm not totally
comfortable with that jargon. But "persistent private" is a
mouthful compared to "OWN", and I doubt anyone can find a word of
appx. 3 characters that conveys the intended meaning so we're
probably stuck with "OWN" as the best short term.



More information about the Python-list mailing list