From xemacs-m  Fri Jan 10 01:49:33 1997
Received: from macon.informatik.uni-tuebingen.de (macon2.Informatik.Uni-Tuebingen.De [134.2.13.2])
          by xemacs.cs.uiuc.edu (8.8.4/8.8.4) with SMTP
	  id BAA22795 for <xemacs-beta@xemacs.org>; Fri, 10 Jan 1997 01:49:31 -0600 (CST)
Received: from modas.Informatik.Uni-Tuebingen.De by macon.informatik.uni-tuebingen.de (AIX 4.1/UCB 5.64/4.03)
          id AA14392; Fri, 10 Jan 1997 08:49:26 +0100
Received: by modas.informatik.uni-tuebingen.de (AIX 4.1/UCB 5.64/4.03)
          id AA18808; Fri, 10 Jan 1997 08:49:24 +0100
Sender: sperber@informatik.uni-tuebingen.de
To: xemacs-beta@xemacs.org
Subject: Re: lexically scroped lisp
References: <QQbxtu24785.199701091732@crystal.WonderWorks.COM> <kigenfupx07.fsf@jagor.srce.hr>
Mime-Version: 1.0 (generated by tm-edit 7.95)
Content-Type: text/plain; charset=US-ASCII
From: sperber@informatik.uni-tuebingen.de (Michael Sperber [Mr. Preprocessor])
Date: 10 Jan 1997 08:49:24 +0100
In-Reply-To: Hrvoje Niksic's message of 09 Jan 1997 18:40:56 +0100
Message-Id: <y9lzpyi7ywr.fsf@modas.informatik.uni-tuebingen.de>
Lines: 31
X-Mailer: Red Gnus v0.76/XEmacs 20.0

>>>>> "Hrvoje" == Hrvoje Niksic <hniksic@srce.hr> writes:

Hrvoje> Kyle Jones (kyle_jones@wonderworks.com) wrote:
>> For years I've heard how much greater, cooler and (most often)
>> faster Emacs-Lisp would be if it were lexically scoped.  Has
>> anyone given any thought toward putting a lexically scoped Lisp
>> byte-compiler and byte-code interpreter into XEmacs?  This would
>> not replace the existing compiler and interpret but would be in
>> addition to them.  Is there anything handy out there we could use
>> as a starting point?

Hrvoje> Yes, thought was given to that.  Have you read
Hrvoje> <URL:http://www.xemacs.org/beta/futures/lexical.html>?  It looks as a
Hrvoje> good starting point.

If we're doing that, we should again give serious thought to replacing
the Lisp engine entirely by something more modern.  Scheme is
definitely the right direction.  I have serious doubts that Guile is.
(I've posted more detailed stuff about the whys when this discussion
first came up.)  There are other ways of preserving Elisp
compatibility in modern Scheme implementations (in, say, Scheme 48).

Hrvoje> As for speed, Erik Naggum has posted an assessment about an average
Hrvoje> speedup of 2x (twice faster Elisp), based on various functions with
Hrvoje> speedups from none to 10x.

I'm pretty sure Elisp is by far the slowest "serious" Lisp dialect out
there.  Getting something that is uniformly (and significantly faster)
shouldn't be difficult.

Cheers =8-} Mike

