[
Lists Home |
Date Index |
Thread Index
]
Elliotte Rusty Harold wrote:
> At 4:15 PM +0000 10/31/02, Richard Tobin wrote:
>
> >It also observes that Infosets may be created by other means, but it
> >doesn't really attempt to constrain them much.
>
> Which is the problem. Naturally a real XML document is well-formed.
> The problem is that there can be synthetic infosets that cannot
> possibly be serialized as well-formed XML documents, and other specs
> are being defined in terms of this most general infoset instead of
> the much more restrictive and interoperable case of XML itself.
What well-formedness errors are possible in a synthetic Infoset?
The only ones I can spot are:
+ duplicated attribute names on an element
+ processing instructions whose [content] property contains
a PIC delimiter (?>)
+ comments containing a COM delimiter (--)
+ illegal characters in element and attribute names
(this one is questionable)
+ countless namespace pathologies
Are there any others?
I really don't believe the first four items are worth getting
worked up over. The last one (namespace pathologies) is the only
serious problem.
--Joe English
jenglish@flightlab.com
|