Rael Dornfest <rael@...> wrote in
news:4808B97E-78CC-11D6-85C9-000393089D30@...:

> My point exactly. Less work on the part of the producer to do
> any escaping; they can just drop their HTML into a CDATA and
> know that it'll all come out nicely in the end.

Well, I really hate the idea of embedding pseudo-markup in RSS
feeds. For *me* RSS stands for 'Really Simple Syndication'. It
should be simple, lightweight syndication format, which can be
viewed with all kinds of devices, from high-end desktop computers,
to mobile phones and on Web pages.

But now people are instead using CDATA sections, double escaping,
and all forms of oddities which really makes no sense in an XML
environment. To write a RSS viewer, you mustn't just support the
very simple markup format XML is, but instead implement a complete
HTML Web browser, with support for tagsoup markup, and implement
at least two parsers (one for XML, and one for the pseudo-markup).
And this means you can't write RSS feeds *about* Web development,
as som viewers will display &lt;em&gt; as emphasis text, even if
you were just *talking* about the 'em' element. It's really ugly.

XML *has* a well though-out mechanism for mixing various XML
markup languages. It's called namespaces. And it would work
perfectly for RSS. With it, an RSS viewer mostly only needs to
support a normal XML parser (which of there are tons freely
available). If it doesn't support any XHTML elements, it can just
treat them as unknown elements, and render the text in them.
This will work perfectly. And *if* it supports XHTML, or at
least a subset of XHTML, such as the text module,
<URL: http://www.w3.org/TR/xhtml-
modularization/abstract_modules.html#s_textmodule >
it can render these elements. (Even mobile phones or other simple
display device can render these, e.g. <em>foo</em> can be rendered
as *foo*.

--
Karl Ove Hufthammer