HTML Unleashed. SGML and the HTML DTD: Should HTML Continue Developing as an SGML Application? | WebReference

HTML Unleashed. SGML and the HTML DTD: Should HTML Continue Developing as an SGML Application?


HTML Unleashed: SGML and the HTML DTD

Should HTML Continue Developing as an SGML Application?


It is true that one can create an HTML browser without it being an SGML parser (or even its proper subset).  Equally true is that the overwhelming majority of HTML users are quite comfortable without the least notion of DTD intricacies.  As you've seen in this chapter, a DTD is not very helpful in respect to the meaningful aspects of HTML, being limited to its formal syntax only.  So what is the value that SGML adds to HTML development? Is it perhaps time to leave the SGML heritage behind?

Quite a lot can be said to advocate SGML's importance.  To begin with, SGML is an authoritative international standard that makes an SGML-supported argument especially strong in the modern world torn apart by browser wars and incompatible HTML extensions.  After all, SGML has been proving its usefulness during more than a decade while HTML is much younger and, alas, significantly less stable.

SGML has a great potential outside of the HTML arena.  A number of important SGML applications have been and are being created for various documentation projects around the world.  The SGML user community is strong and influential.  HTML development can only profit by drawing from the mainstream of SGML philosophy and practice.  Examples of promising SGML-inspired developments that may someday change the HTML world include Document Style Semantics and Specification Language (DSSSL), a versatile style language for use with SGML, and Extensible Markup Language (XML), a subset of SGML designed for use over the Internet.  (See Chapter 38, "The Emergence of XML".)

The importance of a clear and unambiguous syntax specification should not be undervalued, either.  Not only does a DTD, if present for a specific HTML flavor, give ultimate answers to many syntax-related questions, it also enables automatic checking of HTML documents in a robust and reliable way.  I'd say that an HTML version without a DTD is like a language without a dictionary: Not everyone speaking the language needs to consult the dictionary, but those who really influence its advancement will hardly do without a good reference book.

The final, and probably the most important, argument is that it's SGML, not HTML, that was designed to ensure document portability and easy transformability.  One of the main SGML missions is to guarantee that the content we create is accessible to everyone in spite of incompatible proprietary technologies.  From this viewpoint, the HTML of nowadays with its flavors and browser feuds can hardly be named a deserved heir.  However, by sticking to the SGML roots of the language, we can still considerably facilitate automatic handling of HTML files and using them in a diversity of environments, with the final effect of improving the longevity of our information.


Created: Jun. 15, 1997
Revised: Jun. 16, 1997