Lists Home |
Date Index |
> -----Original Message-----
> From: Dare Obasanjo [mailto:firstname.lastname@example.org]
> Sent: Sunday, December 09, 2001 11:30 PM
> To: Eric van der Vlist; xml-dev
> Subject: Re: [xml-dev] Formal definition of PSVI (was [xml-dev] Where
> have the element types gone?)
> Sorry about the segue but your mention of the PSVI reminded
> of something that was of much consternation to me a few months ago. There
> to be no place where the concept of the Post Schema Validation Infoset is
> formally described,instead references to it are littered all over the
> Web and the W3C website as if the meaning of the term is so
> that anyone whosees the term should be instantly aware of what it means.
Yup. This is a bit of a chronic problem in the W3C (and other bodies that
do part of their business in public, part in private for all I know).
Certain unarticulated assumptions are discussed internally or taken for
granted because of previous work until they become part of the fabric of the
organization's being. The "attributes have no defined order" meme is a
fairly trivial example. The "PSVI as the foundation for the next generation
of XML" is a more serious one. We saw last summer in the great namespace URI
debate what happens when someone innocently falls afoul of a revealed truth
that never quite got written down in an authoritative manner. For that
matter, there's now a "don't touch the namespace URI question with a 10 foot
pole" meme that is also not written down anywhere <grin>.
So *is* the PSVI the foundation for the next generation of XML? I
personally will believe it when I see a coherent explanation of what it is
and how to build specs upon it, and when some useful implementations of
those specs get built. I'm not holding my breath.