xml-dev - RE: Why the Infoset?

[Kay Michael <Michael.Kay@icl.com>]

Personally, I don't have any problems identifying the need for the infoset:
I've seen so many people try to attach meaning to lexical distinctions that
should not carry meaning that I yearn for an authority I can point to when
telling them they're wrong.

But the problem with the Infoset as currently defined is that it has had to
make too many compromises. Creating a common abstraction with the constraint
that XML, XML Namespaces, the DOM, and XPath should all conform with it is,
I think, a requirement that has proved impossible to satisfy.

Which perhaps explains why the Infoset and Canonical XML, which are both
essentially trying to answer the same question, have come to different
conclusions. We now have the perverse situation that the core infoset of a
document does not contain enough information to generate its canonical form,
and the canonical form does not contain enough information to generate its
core infoset.

For example the two documents below have the same core infoset but different
canonical forms:

1: <x:a xmlns:x="one.uri"/>
2: <y:a xmlns:y="one.uri"/>

While these two have the same canonical form but different core infosets:

1: <x:a xmlns:x="one.uri"><x:b/></x:a>
2: <x:a xmlns:x="one.uri"><x:b xmlns:x="one.uri"/></x:a>

So in my book, the essential question "when do two XML documents convey the
same information" remains unanswered.

Mike Kay