Glenn,
it is important to be aware of context when you are looking at the
relative-URL doc.
The *normal* case of a relative-URL is the one you quote from section
3.3: That the base URL is the one you used to access the document.
Indeed, I expect a lot of Web-readers to break horribly if I put
documents on the Web which use Base-URL: to establish a different
base URL and depend on this to parse relative URLs!
I think the idea of Base-URL would be more to point out to the
accessing reader that there are more "natural" ways to get to
the document than the one which is currently used.
In the MIME email context, it would be natural for instance to say:
Content-type: text/html
Base-URL: http://domen.uninett.no/~hta/public-mail-sent-by-me/328
In my <a href=325>previous message</a> I said that I didn't like relative
URLs. Now I have changed my mind.
If I ship this in the mail, the Base-URL would place the message
in a context, using a HTTP URL, which would ensure that all readers
with Web access could access the message I was referring to.
So, I think the Base-URL mechanism is good for something, but not
necessarily for making sense of intra-message references.
(the devil is in the details....so many details, so little time....)
Harald A