Re: [xml-dev] Deterministic Content Models (was DTD ( From Eliotte

[Richard Tobin <richard@cogsci.ed.ac.uk>]

>However, it's not clear what kind of error this is.

It's the "is an error" kind.

You're right that even validating parsers don't *have* to report it,
but it *is* an error.

RXP reports it.  Eg for <!ELEMENT foo (A?, A?)>:

 Warning: Content model for foo is not deterministic.   At start of content
 there are multiple choices when the next element is A.

I would be interested to know whether the validating parsers that
don't report an error correctly check the content of the corresponding
elements.  They won't if they just build a finite-state machine in the
straightforward way.

-- Richard