🚀 go-pugleaf

Mark Up Systems vs. WYSI(A)WYG

Arguing about the relative merits of markup vs. WYSI(A)WYG is like
arguing about the relative merits of microprocessors vs. graphical
interfaces.  Markup implements text structure in some way; WYSI(A)WYG
is a way of handling the process of entering text and mark up.  All
text processing systems use markup in one form or another; even the
lowliest text editor inserts line terminators (with Function Key
Enter!).  

There is no reason why a sophisticated markup system like LaTeX/TeX 
can't be fronted by a WYSI(A)WYG editor, and there is no reason why a
WYSI(A)WYG word processor can't be underlain by a markup system as
sophisticated as LaTeX/TeX.  It would take a powerful system with a
good GUI to pull off true WYSIWYG, though, and that's why the current
best approaches to such a thing, e.g., Framemaker or Arbor Text's
Publisher, run on Unix work stations.  

Naturally, on the character-based screen of a DOS system sans GUI there
is no way that you're going to get "true" WYSIWYG.  You're really
working with a continuously refreshed dvi2ASCII window - call it "what
you see is an iconic representation of what you'll get." In spite of
its limitations, this is vastly superior to wrestling with a mixture of
markup and unformatted text.  I cannot conceive how seeing one's text
maintained in some nice approximation of the final result could	be
distracting.  Is it distracting to see one's code indented suitably
while composing it?  Is it less distracting to be continually
reorganizing the text "by hand" into some readable form as you enter
it?

It seems, therefore, that so-called WYSIWYG is really just a friendly
editing environment.  Although current ideas of such an environment
seem to be based on obtaining near perfect representation of the final
result, while retaining access to the logical structure, it isn't clear
that representational exactitude is essential.  At any rate it is almost
never achieved, and most people are still satisfied with the results.  

The real opposition between LaTeX/TeX and most "word processors" is
that the former have paid special attention to detailed control of
formatting (TeX) or to abstracting the formatting process in favor of
the structuring process (LaTeX and SGML), i.e., to perfection of markup
systems, while the latter have devoted their attention to perfecting
the user interface.  The ideal tool would have the advantages of both
approaches!  

Current "WYSIWYG" word processors are indeed system specific.  This is
partly system specificity in the sense that a dvi2X filter is X
specific, or that the keyboard of your terminal is system specific. 
Since the editing view is provided by something like that filter, it is
naturally device dependent.  Actually, most DOS word processors are
amazingly flexible in terms of what hardware they will work with,
because DOS is amazingly flexible as to what I/O devices it will
support.  This kind of system specificity is inherent in the I/O
process.  

There is, however, a real problem in word processors with
system-specificity in terms of proprietary markup systems.  Each word
processor uses its own, or, at least, there are not many cases of
shared systems.  And, of course, these proprietary systems are not as
powerful as the system embedded in, e.g., LaTeX/TeX, though some of the
best of them are much better than LaTeX/TeX users seem to realize;
automatic numbering of logical units is here today, for example, not
something vaguely hoped for.  It would be nice if there were only one
or two universal, non-proprietary markup systems in use, so that the
difference between word processors was merely a matter of the details
of the interface between the author and the markup system, or what
details of the markup system were supported in the editor, etc.  Maybe
someday.  

By the way, several DOS word processors/publishers have Unix versions
(or vice versa)!  

This is all personal opinion, and does not reflect in any way the opinions
or policies of my employer.