Frankly, I think those who believe they are arguing from "first
principles of SGML" are misled into seeing their own prejudices about
what SGML is or is not. I include here the preface to "The SGML
Handbook" by Charles F. Goldfarb, which I think summarizes the
situation here, too. HTML is an application of SGML, and what it
becomes is what we make of it.
> There is a wonderful poem by John G. Saxe called "The Blind Men And
> The Elephant" that begins:
> It was six men of Indostan
> To learning much inclined
> Who went to see the elephant
> (Though al1 of them were blind),
> That each by observation
> Might satisfy his mind.
> I don't remember the words, but I recall that each of the blind men
> approached a different part of the beast. One stroked the tail and
> decided the elephant was like a rope. Another touched its side and
> concluded that elephants were like walls.
> One blind man grabbed the trunk and determined it was another form of
> snake. The fourth put his arms around one of the elephant's legs and
> declared elephants to be like trees.
> The fifth man, I think, felt the elephant's ear and decided it was a
> poorly tailored leather jacket or some such, and I don't remember
> what the sixth did at all. In any event, they all wound up arguing
> about the true nature of the elephant, with each being partly right
> and all of them being wrong.
> Saxe's poem was intended as a dig at theologians, who necessarily
> argue about something that none of them can see. However, it can
> apply just as well to some of today's high tech community. Faced
> with something new and unusual, these experts and specialists,
> though sighted, sometimes fail to see any more than what they have
> been trained to expect.
> Which brings us to the subject of this book. The Standard
> Generalized Markup Language has in many respects been a victim of
> this modern form of blindness. There has been a dismaying tendency
> to characterize SGML solely in terms of the aspect with which one
> happens to make contact:
> - It is a tagging language.
> - It handles logical structures.
> - It is a file linking and addressing scheme.
> - It is a data base language for text.
> - It is a foundation for multimedia and hypertext.
> - It is a syntax for text processing style sheets.
> - It allows coded text to be reused in ways not anticipated by the
> - It is a document representation language for any architecture.
> - It is a notation for any kind of structure.
> - It is a metalanguage for defining document types.
> - It represents hierarchies.
> - It is an extensible document description language.
> - It is a standard for communication among different hardware
> platforms and software applications.
> SGML is and does all of these things, but the whole is much more
> than the sum of the parts. Moreover, it is only by understanding the
> whole that one can make the best use of SGML.
> I think that a high tech version of Saxe's poem would have a
> dramatic ending: The elephant eventually gets irritated by all the
> poking and fumbling and runs off, trampling the six blind men in the
> The moral, of course, is that if you are going to mess around with
> something powerful that you do not fully understand - even something
> benign - you had better do it with your eyes open.
> I hope THE SGML HANDBOOK will be the eye-opener that lets you see
> the elephant.