RE: Input to BP 19 http://www.w3.org/International/its/techniques/its-techniques.html#AuthCDATA

Hi Christian, all,


> I had a look at "Best Practice 19: Avoid CDATA sections when possible".

Sorry, I was still working on this BP this week-end. So you may re-check it.

But mots comments I've seen are still relevant.


> 1. I wonder if we need a BP for schema developers which corresponds to 
> this BP for authors.

I'm not sure. Developers cannot prevent the author to use CDATA.

I think we have several different issue under the "CDATA" umbrella.

1) Using CDATA notation has draw back regardless of what the content is (because then you can't use markup or NCR and that leads to
the different problems we have noted).

2) It happens that CDATA is often use to enclose data that is basically not "normal" text content: scripts, HTML chunck, XML
examples, etc. In this case CDATA is just a side effect. The real problem is that we have non-textual content: basically non-XML
data inside XML.

I think those are different issues and we may need to treat them separately, and possibly have a new BP for the second case.

In this case the BP for CDATA would not assume any specific content and deal only with CDATA drawbacks. And we we would refer to the
other BP for the "non-XML content" BP.

See the latest draft

Cheers,
-yves

Received on Monday, 10 September 2007 16:01:21 UTC