Re: [AUCTeX-devel] New tex-symb.el

[Masayuki Ataka]

From: David Kastrup <address@hidden>
Subject: Re: [AUCTeX-devel] New tex-symb.el
Date: Sun, 04 Feb 2007 11:06:26 +0100

> One problem I see with this is that it has _major_ overlap with the
> input functionality of X-Symbol <URL:http://x-symbol.sf.net>.  I don't
> know whether this is intentional, but I think that the X-Symbol
> functionality in this regard was very well designed (it also offers
> tables from which to pick symbols and various other stuff).
> 
Yes.  It's intentional.

I regard tex-symb is a Poorman's X-Symbol.
# You see te{X-Symb}ol.

I had been used X-Symbol for 3 years until 2005.
X-Symbol is a very well designed tool.  Yes!
X-Symbol has two features:
 1. Input method for math symbols, and others.
 2. Replace TeX commands for symbols (like `\gg')
    by respective characters using encoding.

X-Symbol have many problems at feature 2 (replacement).

 a. User can't search the TeX command because the command is
    replaced by uninputtable character.

 b. User can't replace the TeX command in the same reason.

 c. An unexpected space is inserted just after the respective
    character.  This causes a problem with circ.sty.

      circ.sty is a package for writing an electric circuit for
      LaTeX.  In circuit environment, `\S0' means `Switch 1',
      but X-Symbol converts `\S' by the section symbol and
      insert an unexpected space butween `\S' and `0'; `\S 0'
      does not represent `Switch 1' any more.

 d. Problem when saving buffer

      Emacs 21 fail to selecting coding-system.
      The following code will help for Emacs 21 user.
      (setq x-symbol-auto-conversion-method 'slowest)

      See also comments at L.5051 in emacs/src/fileio.c.
      This fix is not included in Emacs21.
      
 e. Conflict with CC-Mode

      http://lists.gnu.org/archive/html/emacs-pretest-bug/2003-06/msg00147.html
      Fixed in Emacs22, but not in Emacs21 (IIRC).

 f. Conflict with texinfo-format-buffer

      User should turn off X-Symbol manually before M-x
      texinfo-format-buffer, because X-Symbol does not put back
      the respective chars to Texinfo commands.

      The user variable `texinfo-pre-format-hook' will help
      (patched by Christoph).  (Emacs21 does not have this hook!)

Well... I gave up using X-Symbol in 2005.
(X-Symbol stopped developing from 2003?)

I believe that replacing approach is dangerous.  Ralf's
approach (Folding) is safe and easy to edit and search.

So, I started to write a tex-symb.el in 2005;  The concept is...

  tex-symb.el + tex-fold = X-Symbol (limited version)

I think that tex-fold would become better if it uses not only
utf-8 chars for math symbols but also fonts distributed with
X-Symobl.  If X-Symbol throw away the replacement approach and
uses Folding approach (tex-fold), I don't use my tex-symb.el
anymore.  But I think it hard work.  At least 2 years ago, I
couldn't do that.

# Anyway, tex-symb + tex-fold is an alternative (safe) tool of
# X-Symbol especially for FSF Emacs 21 users.

> I think it would make sense to steal the (really well-designed) input
> method functionality from X-Symbol.  It would require a copyright
> assignment from Christoph, of course, and we should take care to
> ensure that those people who actually use X-Symbol can do so (and
> switch it on and off) without AUCTeX disrupting its information.
> 
> Christoph, any comments?
> 
---
email: address@hidden
Name:: Masayuki Ataka // (Japan)