Re: XHTML, Email, Data Objects and Metadata

On 2/12/2014 9:16 AM, Adam wrote:
> HTML for E-mail Community Group,
>
> Greetings.  In addition to agreeing that e-mail would be enhanced by 
> mathematical and scientific notations, as discussed in /Math in 
> Email/ (http://lists.w3.org/Archives/Public/public-htmail/2014Feb/0033.html), 
> I would like to recommend XHTML, instead of HTML, for e-mail. XHTML is 
> extensible, includes expressiveness for data objects, microformats and 
> RDFa, facilitating features.

It's not clear to me that XHTML can do anything that HTML can't do, at 
least under HTML5.
> Scenarios and use cases include, in addition to mailing lists, mailing 
> lists as interoperable with websites and content management systems 
> such as Drupal, Joomla and Wordpress.  Scenarios include mailing list 
> software and various collaborative software which can be utilized by 
> groups, web-based software.  Scenarios include new features possible 
> for discussion groups, discussion group websites and NNTP websites.

Most of this is thinking the wrong way. The HTML of a message is the 
body of the message, which is a comparatively expensive thing to 
process. In many cases, an email client won't process a body until it is 
displayed, so any metadata that is valuable before then (in, say, a 
message list pane) shouldn't go in the body. This explains, for example, 
why the header-wrapping features of S/MIME 3.0 have seen little to no 
uptake.

There are two places to put true metadata in a message: either the 
headers or as an attachment with a well-known MIME type. Invitations are 
processed as text/calendar, for example, and that works across email 
clients. Threading is indicated by In-Reply-To, Message-ID, and 
References headers, and there is a large set of mailing list metadata in 
headers--from the headers of this message alone, I can tell you how to 
post to the mailing list, unsubscribe, identify which mailing list it 
was sent to, and even tell you where the mailing list message is archived.

-- 
Beware of bugs in the above code; I have only proved it correct, not tried it. -- Donald E. Knuth