Re: 3 basic XML questions on: non-determism, reserved symbols,and xml:lang

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

> Can someone
> please clarify what it means that non-deterministic content models are
> not allowed in XML?

It's "an error" (rather than a validity or well-formedness error),
which means that parsers are not required to detect it.  So the
parsers you mention are not wrong, just not as good as they might be.

> [2] It was also my understanding that the following symbols are
> "reserved", and may not be used in an instance document's data:

>     <  >  &  "  '

> Instead, the "escaped" version must be used:

>   &lt;  &gt;  &amp;  &quot;  &apos;

> However, my experience is that XML Parsers do not require >, ", or ' to
> be escaped.  Can someone please clarify what are the "reserved"
> symbols?  When do I need to escape >, ", and '?

You always need to escape &, because it would otherwise be interpreted
as an entity reference.  You always need to escape <, because in content
it would be interpreted as a start tag and in an attribute it's just
illegal (for historical reasons, I think).  Quotes need only be
escaped in contexts where they would otherwise be misinterpreted
(i.e. inside strings quoted with the same kind of quote).  > never
needs to be escaped; &gt; is just for symmetry.

> However, I have found that
> whenever xml:lang is used on an element it must be declared in the DTD. 

Yes, if you want the document to be valid.  If you don't care about
validation - you're happy with well-formedness - you don't have to
declare it.

-- Richard