There have been a number of thoughtful responses to my confession, the other day, about cheating on Web standards. Several folks recommended this approach:
<blockquote
cite="http://www.betaversion.org/~stefano/linotype/news/35/"
title="Stefano Mazzocchi">
...
</blockquote>
Jim White also made this intriguing proposal:
<blockquote cite="urn:name:Stefano%20Mazzocchi">
..
</blockquote>
Jim pointed me to the IANA registry of URN namespaces, noting that while 'name' is not among those registered namespaces, and the one you can find there -- RFC3043, Personal Internet Name (PIN): A URN Namespace for People and Organizations -- isn't quite right either, these are examples of valid ways to extend an attribute that takes a URI as its value.
Of course that still left the other problem:
<pre class="code" lang="python"> .. </pre>
I think that for the next phase of this experiment, I should just bite the bullet and start writing nonstandard attributes -- such as 'lang' in this case -- into another namespace. For an author, as Jim points out, there's not a lot of extra friction or overhead. It could be as little as two extra characters:
<pre class="code" e:lang="python">
..
</pre>
The 'e' would be for 'experimental' -- mapped to what URI I don't yet know. As Jim rightly points out, the burden to process these experimental attributes would fall mainly on developers of authoring and search tools, not on users. Since I've got a couple of my own XML-aware search tools running now, I'll give this a whirl and see how it goes. Thanks to everybody who commented on this matter. I will continue to be interested to hear from people with ideas about how to strike the right balance.
Former URL: http://weblog.infoworld.com/udell/2004/02/05.html#a909