Thinking About SGML

With the XML craze in full frenzy, SGML is being pushed into the background even more, for all but diehards. This is a pity, because SGML has a number of very useful features that didn't make the cut for the first (and to date only) version of XML.

Even so, SGML isn't without faults. There's room for improvement.

Data Attributes: Extending the Basic Syntax

Taking Annex K.4.4 of the WebSGML TC to its limit.

LPD Before DTD in the Prolog

Getting rid of a nasty difficulty in ISO 8879

Generalizing IDLINK

Addressing context by other attributes.

How useful is a NOTATION Declared Value?

Not much, IMHO!

XMap: A Mechanism for Mapping Names

A way to combine XML Namespaces and SGML Architectural Forms

The Namespace Bogosity

A broadside on XML namespaces, which I have always considered worse than useless.

XML Namespaces Demystified

A Usenet post explaining the basics of mapping names, and how XML Namespaces got this hopelessly wrong.

XML Namespaces: A non-solution to a non-problem

The first part of a write-up (in May 2003) of the issues with XML namespaces. Some of these were brought out in a long thread on the xml-dev list.