The link you provided doesn't seem to be working in IE. I guess this is to make up for the fact that Erik's original link only works in IE :)
Anyway as for the stuff you feel is incompletely described or missed completely like namespaces and mixed content, I'd like to point out that Xen's goal was to create a way to manipulate what people consider to be data-oriented XML as opposed to document-oriented XML (I give better descriptions of these concepts in http://www.sys-con.com/xml/article.cfm?id=727). There are known gaps in the story and some of this is stuff I'd like to see end up in the .NET Framework as opposed to remaining research work so I'm definitely actively involved in seeking solutions to these problems. Mixed content is particularly problematic, Erik and I have talked about how one would actually deal with them in a reasonable fashion to some degree but haven't sketched out anything concrete.
With Whidbey (.NET v2.0) ship work winding down and Orcas (post .NET v2.0) not yet started I'll be able to spend more time thinking about these problems and perhaps putting together a paper of my own.
Dare: namespaces are certainly relevant to data-oriented XML. I also personally believe that the line between data oriented and documented oriented is as illusionary as the line between metadata and data.
If the discussions you describe are open to outside participation, I'm interested. In any case, I'll be sure to look Erik up at the xmlconference.
P.S. I fail to see anything wrong with the URLs I provided... they seem to be properly URL encoded. Any idea as to why they confuse IE?
Dare: speaking of mixed content, have you considered posting your messages as xhtml:body or atom:content? If your application already 'groks' XML, then I wouldn't have to guess as to the formatting. Often getting it wrong, such as in the botched hypertext link in your comment.
There is no reason why you couldn't continue to include description in the entry in support of clients which only are looking for that element...