- From: Arjun Ray <aray@q2.net>
- Date: Mon, 2 Jul 2001 00:55:21 -0400 (EDT)
- To: www-talk@w3.org
On Sun, 1 Jul 2001, Ian Hickson wrote: > On Sun, 1 Jul 2001, Arjun Ray wrote: >> Other than giving the pope heartburn, what's wrong with a processing >> instruction? (That's what they were invented for, after all.) > > Pages that comply to the Appendix C guidelines can't contain them. > (They're not tag-soup-compatible, apparently.) Well, it's pretty clear that Appendix C is a crock. Even so, C1 simply says "Be aware that processing instructions are rendered on some user agents". This is an historical inadequacy of the kind that anyone stil using a browser of such vintage can't but be prepared for some things not to um, work. Besides, the xml declaration has the lexical form of a PI. To take C1 as you're suggesting would imply that this construct should *also* be avoided. Wonderful! In short, raising the spectre today of PIs "failing" is a very lame excuse, a copout. One of the major divisions pf Dr. Goldfarb's _The SGML Handbook_ is a "Structured Overview of SGML", which goes over the alphabetically ordered definitions of ISO 8879 Clause 4 in a separate, logically coherent order. Section 4.2.3 "Other Markup" (p.138ff) has this: : 4.2.3.3 Processing Instruction : : Finally, as war is to diplomacy, there is the last resort of : descriptive markup: : 4.234 processing instruction: Markup consisting of system- : specific data that controls how a document is to be processed. What is the argument not to use standardized markup for its intended purpose? Arjun
Received on Monday, 2 July 2001 00:40:02 UTC