CDATA

Michael Kay wrote:
> I always suspect that software that blindly wraps everything in CDATA
> sections regardless of its content is probably not bothering to check
> whether the text contains "]]>".

Without reducing the culpability of these lazy programmers, I'll say 
that CDATA is pretty broken and should be fixed if there is ever an XML 
2.0. It is CDATA itself that works 99.9% of the time. XML should have a 
readability-preserving escape mechanism that works 100% of the time. The 
most logical thing is to let the application choose its own delimiter.

  Paul Prescod