Re: when you gotta have a variable value for a symbol name

[Emanuel Berg <embe8573@student.uu.se>]

"Buchs, Kevin J." <buchs.kevin@mayo.edu> writes:

> I want to evaluate (kmacro-name-last-macro variable),
> where I want the value of "variable" passed as the
> symbol name. Despite years of trying, I don't think I
> ever really conceptually "got" the distinction
> between symbols and variables and that seems to be
> critical here.

Good question. In books on programming when they
mention Lisp in one paragraph they always say it came
out of a crazy mind. No, it came out of the push for AI
in the US in the 50s (probably with the cold war in
mind and everything). I thought that "AI" referred to
the data/code blend that can make a function rearrange
another function like any other list of data, perhaps
swapping a pair of operands around a binary operator
and reinsert the modified function in a pool of
competing functions, to see what first morphs into a
function that can solve some task... And more, the Lisp
reliance on the list data structure, that can be
nested, and searched, as so much of the AI applications
revolve around searching trees according to probability
calculations... But they (the books) also usually say
Lisp is about "symbolic manipulation". Is that
literally manipulating symbols?  Drawing on the
dynamic/lexical scope discussion the other day, symbols
would then be dynamic scope as the function would act
on the existing symbol. Variables would be the lexical
scope or the traditional programming approach of using
variables to build programs (and not the other way
around, to build programs to juggle around with
symbols). I guess almost all ordinary Elisping isn't
anything uncommon like that, just to have Emacs
automatize and do things certain ways, nothing fancy
from a theory/science point of view.

--
underground experts united