Re: [xml-dev] What characters can go into a CDATA section and acomment?

[Peter Flynn <peter@silmaril.ie>]

On 24/03/2022 05:06, Mukul Gandhi wrote:
> Hi all,
>      I personally see, few similarities between purposes of CDATA 
> section and a comment within the context of XML documents.
>
> I seem to clearly, understand the purpose of comments. Any text within 
> comments, are only for human reading (and rarely for automated reading 
> by computers. for e.g, we have XPath comment nodes).
>
> But I'd like to know, what are the use cases of CDATA sections. Any 
> thoughts, shall be helpful to know.
THE most important use case IMNSHO is for documenting markup. I like to 
be able to write:

<programlisting language="XML"><![CDATA[
<?xml version="1.0">
<!DOCTYPE foo SYSTEM "bar.dtd">
<foo>
  <title>Some &lt;title></title>
</foo>
]]></programlisting>

and NOT

<programlisting language="XML">
&lt;?xml version="1.0">
&lt;!DOCTYPE foo SYSTEM "bar.dtd">
&lt;foo>
  &lt;title>Some &amp;lt;title>&lt;/title>
&lt;/foo>
</programlisting>

SGML had other types of Marked Section, but they did not make it into 
XML because the stricter syntax made them unnecessary, and TBH they 
didn't really get a lot of use, whereas CDATA sections are a mainstay.

CDATA sections are also used by many web developers unsure of exactly 
/what/ a user is going to input, and exactly /when/ in subsequent 
non-XML processes the markup is going to be stripped, so they use it as 
a safety-net of last resort, which often goes wrong; but that seems to 
be an acceptable price for them for passing the buck to someone else's code.

Peter