RE: [xml-dev] If XML is too hard for a programmer, perhaps he'd b e better off as a crossing guard

["Joshua Allen" <joshuaa@microsoft.com>]

> doesn't let me scream "IGNORE THAT CDATA SECTION MARKER AND PARSE THE
> DAMN CONTENT NORMALLY!"

The two are not interchangeable, I am sure you know.  Markup inside a
CDATA section is completely different from markup inline with the
document.


> I'm in the classically stupid position where the export is generating:
> 
> <root>
> ...
> <repeatingNode1><![CDATA[This is in <b>bold</b>, or at least it should
> be.]]></repeatingNode1>
> <repeatingNode2>content</repeatingNode2>
> ....
> </root>

What exactly is stupid about that?  Presumably the application that
generates and consumes that data expects a *text* node, and not xsd:any.
Are you saying that the export was dumb to demand text, or that the
application really wanted xsd:any and simply screwed up?

Even more importantly, do you *really* want your <b></b> tags to be
hanging out with no namespace?  What will you do when your "markup"
contains something like "<p><br>"?

I get confused when I see people who *insist* on treating HTML as if it
is "markup" rather than text, and then get predictably upset in the
myriad instances where this causes unnecessary pain.

> More creatively, there may also be times where the use of CDATA
sections
> is appropriate, so simply nuking all of them isn't the right answer

Yeah, exactly -- use CDATA (or escaped XML) when you want a text node.
That is actually a whole lot of cases.