Re: Standards for programming languages

[Thien-Thi Nguyen]

() Richard Stallman <address@hidden>
() Sat, 20 Jul 2013 09:35:10 -0400

   If you write a program in Emacs Lisp, then it runs in Emacs.
   For some purposes, that is good, but it is bad for others.

Yes, it runs "in Emacs" but it is possible to arrange so that it can be
invoked from the shell, via "emacs --script", which puts the program on
par w/ others that require an interpreter installed.  For example:

 $ cat > hello-world <<EOF
 ":"; exec ${EMACS-emacs} --script $0 "$@" 2>&1
 (message "hello world!")
 EOF
 $ chmod +x hello-world
 # install hello-world /usr/local/bin

The user does not really need to know/care that the program
/usr/local/bin/hello-world is implemented in Emacs Lisp.

In sumamry, Emacs Lisp is a fine vehicle for delivering useful programs
outside of Emacs, and i think it would boost GNU cohesion to recommend
it (or at least mention it, non-diffidently).

I submitted such a program a year (or two) back to create the GNU gnulib
modules HTML, since the previous (and current, AFAICT) implementation of
that program is very slow (the Emacs Lisp impl was approx. 6x faster).
However, the program was rejected because it was written in Emacs Lisp.
I believe that, had there been some (even minimal) support for Emacs
Lisp in standards.texi then, i could have pointed there and pushed it
through, but there wasn't, i didn't, and so that's why i'm spewing now,
to prep the field for another push in another year (or two)...

[cc trimmed]

-- 
Thien-Thi Nguyen
   GPG key: 4C807502
   (if you're human and you know it)
      read my lisp: (responsep (questions 'technical)
                               (not (via 'mailing-list)))
                     => nil

pgplKqYZqYKWJ.pgp
Description: PGP signature