If you thought that discussions on perl5-porters were hopelessly arcane and abstract, try reading xml-dev. :-)
I lurk on xml-dev, and occasionally a very long winded thread pops up, like how Microsoft abuses the notion of "internet standards" to lock out everything except IE to see their nonstandard MSN pages. Sometimes those threads are interesting. Many times, they seem to go on forever without discussing anything practical.
Of course, real work happens on xml-dev. That's where SAX was developed (both versions), and that's where interesting new work like Regular Fragmentations is announced. Recently, there was a discussion of the concept of an ID within a document. Most people are familiar with <a name="foo"> being linkable via #foo in a URL.
However, that behavior is well defined in HTML, but not so well defined in XML. The way IDs and document fragments are defined is through a DTD, yet a series of XML specifications (XPath, XPointer, XSLT, XML Schema, etc.) rely on the fact that there is *some* way to identify where IDs are found in a document, even though that process is optional. I've come across this shortcoming myself a series of times, and tried to explain the behavior when teaching XML.
Last week (or the week before), this issue came up on xml-dev, and started a long-winded discussion that I ignored as hopelessly abstract. Today, Leigh summarized it in a very readable column, along with the depths of the problem and the workable solutions being proposed.
Thank you, Leigh Dodds.
