RE: Emacs as word processor / Text Properties

[Drew Adams <drew.adams@oracle.com>]

> Internal data structures being XML is a small step -- but only a
> very small step. Angle brackets do not in themselves semantics
> make:-)

Angle brackets are not an inherent part of XML, except its serialized
form.  Think XML data, not XML as marked-up text.  No one represents
XML internally using text.  Not an angle bracket in sight.

And internal data structure being XML or Lisp sexps (which I also
suggested) would be a GIANT step.  A definition of well formedness
and being able to validate against schemas (e.g., XML schemas or DTDs)
makes a huge difference.

> For seeing what I mean, just take a look at the xml files in any
> of the modern MS Office file formats.

Not a reference, sorry - not representative.  I've already commented
on this.  MS Office's use of XML is not up to par.  But even it might
improve with time...  It has to interact with the rest of the world.
At a minimum, it has to grok clean XML and sooner or later will need
to be able to output clean XML for others (or they will filter to
get that).

> Most word-processors after a while lose the distinction between
> layout style and content, and what results in the file format is
> a messy bag of bits.

Wrong.  The tendency is in the other direction, at least if you
include high-end tools that people use to create doc.  Of course,
if you just count the number of executables out there in the wild,
then MS Word might beat all others combined (just a guess).  Still,
it is not the measure I use, and it certainly is not a goal that
anyone starting today would aim for.

In the context of Emacs, which would be to a large extent starting
from scratch in this area, there is zero reason to emulate something
like MS Word, in terms either of its internal data representations
or its *.doc files.  (That does not mean that Emacs could not import
or export *.doc files.  I mean only that it is far saner to build on
something like XML.)