RE: [xml-dev] XML is text-only ... why?

[Len Bullard <len.bullard@uai.com>]

Some parts of that are a little misleading possibly because the XML goals
were set by a self-selected group documenting its own perspectives.

1.  What you are describing was termed 'tag sprinkling'.  That was a point
of comp-text-sgml design threads:  document design vs tagging documents.
Both practices were evident depending on the starting point and the politics
of a project.   HTML's design (a GenCode) came late in the game although
problems in that design dominated XML and were grandfathered into it.
Possibly the lack of ubiquitous design tools for document design based on
markup led to a perception that 'tag sprinkling' was the right way or the
only way.   Any of us who were working on both SGML and relational databases
or even object-oriented programming saw that this was a flawed approach even
if sometimes the only viable one.

2.  I can't think of any case where anyone seriously suggested one could
'rip out the markup and get the original'.  The first generation of
gencoders were typesetters and they plainly knew that was impossible.  The
next generation were technical writers for the most part who had been
submitting their manuscripts to data entry clerks and they plainly knew that
was impossible.  This myth spawned when the comp-sci types began to explore
markup.  They did not tend to be document-savvy people and those that were
did not know a lot about markup.  Out of all of these was a very small
cross-section of experts who plainly knew that was impossible.

To get a flavor, it helps to understand that in the battle between what
deRose and Durand called the 'holy scrollers' and the 'card sharks', the
holy scrollers were victorious in Web 1.0.  In Web 2.0, that victory is in
doubt.  Thus comments that the OOXML vs ODF bitter butter battle is a war
over dodo technology.  The sharks are back and are winning this time.

2.  The idea of markup for structured data was around when ISO 8879 was a
draft.  Papers published on such were already being published before the web
was born.  The infrastructure wasn't up to it and getting it accepted in
comp-sci circles was a difficult undertaking.  The more frequent complaint
and I quote fairly precisely, "any language that allows mixed content like
that is a monstrosity and should never be used with a database".

The emergence of XML as a dominant data container came as a result of cheap
processing cycles and memory, not because of sudden realizations of its
utility for non-page-based applications.  The uptake of that idea on the web
was quick and that gives the appearance of invention where it is only
broader acceptance.

len


From: richard@inf.ed.ac.uk [mailto:richard@inf.ed.ac.uk] 

Digression:

This historical perspective on XML is often forgotten.  Many aspects
of XML are ones inherited from the idea of a markup language as taking
an existing text document and wrapping markup around it.  There was a
widespread assumption that you could rip out the markup and get "the
original".  This explains why, for example, HTML titles are content
while hypertext references are attributes: the titles are part of the
text, but the URLs aren't.  It also explains the default template rules
in XSLT.

The use of SGML and XML for arbitrary structured data is a later idea.
This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender. This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail.