🚀 go-pugleaf

In article <JKH.90Nov26131532@bambam.pcs.com> jkh@bambam.pcs.com
(Jordan K. Hubbard) writes: [Warning -- selective quoting ahead]

   To state it more succinctly this time, why not Scheme as Emacs'
   extension language? I know E-Lisp already exists, and works, but it's
   Yet Another Dialect and I'm wondering why we still need go that way
   when a small & optimal lisp now exists.

Looking at the E-lisp manual, I get the distinct impression that the
major reason for the large size of e-lisp is the need to supply many
commands related to editing, buffer handling, key bindings et cetera
et cetera.  This will not change if you do it in a different language.

   Existing E-Lisp code could
   probably be converted fairly easily and automatically, so I don't see
   that such an upgrade would require chucking out everything that's gone
   before, either.

I wouldn't be so sure.  I believe much existing e-lisp code relies
heavily on the dynamic scope of variable bindings, so you can say
things like

   (let ((some-variable some-temporary-value))
      do a bit of work, calling functions which look at some-variable
      to decide how to behave)

and have some-variable restored to its original value afterwards.
With the lexical scoping of Scheme, you would have to do this in a
very different way.  (I believe RMS once wrote a paper about the
design of emacs, explaining all this in more detail.  Does someone
have a reference, or can shed more light on this point?)

   Finally, a number of good books on programming in
   Scheme already exist, so we wouldn't need to redocument the wheel.

Again, the e-lisp manual is mostly concerned with emacs specific
stuff, not with the base language itself.

   Since V19
   is going to be such a major departure anyway, why not go for the
   whole enchilada?

I think people think they have waited long enough for V19 as it is.
Don't let us add another year or two to the wait.  Don't misunderstand
me, I would love to see an Emacs based on Scheme myself, but I don't
think it is worth the trouble.

- Harald Hanche-Olsen <hanche@imf.unit.no>
  Division of Mathematical Sciences
  The Norwegian Institute of Technology
  N-7034 Trondheim, NORWAY