Lists Home |
Date Index |
At 12:41 PM -0400 4/12/04, Michael Champion wrote:
>I must be seriously missing something here. Of course, if we mean
>the verbiage in the Infoset spec about "element information items" ,
>you're right, but I don't think anyone has ever argued that the
>Infoset spec is of much interest to anyone besides specwriters. If,
>on the other hand, we use "infoset" to mean the labeled tree data
>model, all DOM, XSLT, and XQuery-based applications use the "Infoset
>abstraction" and only consume the XML syntax after a parser has done
>its magic to turn it into nodes, events, objects, or whatever.
That's a big stretch, and not accurate. Remember the Infoset
postdates a lot of specs including DOM, XPath, XSLT, SAX1, SAX2, and
others. DOM and SAX at least can handle documents that do not have
infosets, and I think XPath/XSLT can too, now that John's reminded me
about the issue of relative namespace URIs.
XML is not nearly as fuzzy as you want to believe, Michael. You're
trying to rewrite history. We're not talking about how many angels
can dance on the head of a pin. We've got specifications for all this
stuff, not always perfectly specified, but specified well enough that
it's obvious some statements just aren't true.
Elliotte Rusty Harold
Effective XML (Addison-Wesley, 2003)