I use it to determine the desired format. You can see this blog entry (including comments) in html, rss, rdf, and txt formats. The rss2 feed is experimental.
Other aspects of the namespace: if the path element is exclusively digits, then it is interpreted successively as a year, month, or day. If it is alphabetic without an extension, then it is treated as a (as of yet unexploited functionallity) category.
Exploiting categories is on my todo list. As is creating permalinks for comments that people have left, and to integrate in a traceback system.
Oh, and queries are supported. And all of the above can be combined, so you can get a list of all of the postings on Keith in August of 2002 in RDF format, if you so desire. Filtered by category, once I start using that function.
All of this from a simple URL.
This is more or less how I feel. Mime-types are all well and good, but you need a human-readable version of the data contained in the URL so that people and dumb-apps know what they're asking for. If I have two versions of my resume, for example (which I do) people need a way to distinguish between downloading the online version and the printable version. The .html and .doc extensions are an organized way to do that.
Ugo responded again on this topic and says that Google will cache dynamic links. GREAT! The only reason I care about fixing my URLs is for Google. So why is Google only recording my front page!?!?! See all those calendar entries on the side? Google's ignoring all of them for some reason.
I just did a look up for "Russell" - I've bopped from page 13 to page 3!!! whooohoo! I want to be the number one Russell on Google. That would rock. There doesn't seem to be any other bloggers competing for that name (that I know of) so I might be able to do it if Google starts caching the rest of my site.