Re: [AUCTeX-devel] Further TikZ patches

[Uwe Brauer]

   > Hi Uwe,
   > 2016-04-23 19:01 GMT+02:00 Uwe Brauer <address@hidden>:

   > We already have hundreds of lines of compatibility code for XEmacs,
   > and some of them also applies to GNU Emacs 21.


   > Out of curiosity, based on your experience, are there AUCTeX users
   > running XEmacs?  Back in 2005, these lines were added to the AUCTeX
   > FAQ:


[...]


   > After 11 years, the situation didn't improve much.  It's a bit awkward
   > to develop a code targeting a platform that you have the feeling
   > doesn't use at all your program.  Apart from you (thanks anyway for
   > your bug spotting work), we don't hear literally from anyone using
   > XEmacs.  You may say that we don't hear that much also from GNU Emacs
   > 22-23 users, but usually we don't need much compatibility code to make
   > AUCTeX work on those Emacsens, so developing for those platforms
   > doesn't require extra work.

I understand completely. I think, as (still official) xemacs pkg
manager, I heard about one or two comments/complains in Xemacs beta in
the last 11 years. The only other persons I know using Xemacs+auctex are
two collegues of mine in my department, and one changed to GNU (aquamacs
on a MaC) when I announced I would give it a try and when Steve Turnball
asked about the future of Xemacs in October.

Ok may it is time for my rant.

I kept  Xemacs for the last decades, besides some particular
features, because of x-symbol, for me the killer application.


I switch partially (90%) to GNU Emacs, because

    - it had BIDI support 

    - better support for most packages I cared (auctex, gnus, orgmode).

    - a package system from which I say after month of using it (sorry
      Steve): also it came late compared to Xemacs it seems better in
      terms of usability and it seems also easier to add new package via
      ELPA MELPA or Marmelade.

      - pretty symbol does almost everything x-symbol does. [1]

I use Xemacs however in two occasions:

    -  when I encounter old latex files, which contain mixtures to
       express no ASCII symbols like 'o \'o latin-1 or utf8 in the same
       file. For some reason Xemacs deals better with that constructs
       than GNU Emacs. Maybe Xemacs is picky about utf8. I don't know the
       reason.

    -  if I have files which a lot of graphics via includegraphics.
       Thanks to x-symbol they are instantly displayed. I know I can use
       preview-latex but this sometimes fails and is slower than the
       x-symbol solution. I asked some time ago about the implementation
       of such a feature but Tassilo told me that it would be a lot of
       work, and the functionality would not be so different from
       latex-preview. I understand completely his decision, besides I was
       the only guy who asked for it.


And one last word. I still cannot understand how a technical
superior [2] solution, 20!!! years ago now stands at the edge of
extinction and even worse it features did not made it into GNU Emacs,
but were re implemented and in most cases differently. One might call
this a sign of freedom, I call it a sign of waste of human creativity
and effort.

Uwe 


Footnotes: 
[1]  they are completely different implementations both with advantages
     and disadvantages.
|             | pretty-sym                     |   | x-symb                  |  
 |
|-------------+--------------------------------+---+-------------------------+---|
| filling     | not effected (overlays)        | + | affected                | 
- |
| velocity    | fast                           | + | sometimes slow          | 
- |
| corruption  | no                             | + | yes                     | 
- |
| aesthetic   | ok                             | - | nicer fonts             | 
+ |
| intuitive   | kill char does not kill symbol | - | kill char kills symbols | 
+ |
| sup+sub ind | buggy                          | - | perfect                 | 
+ |

[2]  x-symbol is from 1996. The package system at least from 2000