httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@liege.ICS.UCI.EDU>
Subject Re: Terminology: URI and URL
Date Tue, 03 Sep 1996 00:29:27 GMT
> The terminology of URI, URLs etc seems to be not 100% clear.  Ok, the URI
> Working Group gave a definition.  In principle: ``URIs are the superclass of
> URLs and URNs''. And they gave some examples of what should be called URLs
> and what else are URNs. Hmmm...

Yes.

> But in the Web community they are not used in this way: Often
> http://www.anyhost.dom/foo/bar/quux/ is called a URI and its /foo/bar/quux/
> part is called a URL. Ok, both labels are correct according to the

No, both are URLs (being locations) and both are URIs (since all URLs
are URIs, by definition).  The latter is called a URL relative to the
server root, or an absolute path URL. You can see such references in
the Relative URL spec, RFC 1808.

Use of URL and URI is not consistent because of politics -- there was a
period of about a year in which the URI WG managed to break all usage
of WWW addresses (URLs) by adding stupid requirements to their syntax.
So, TimBL said to hell with that and changed WWW addresses to the name
URI.  Unfortunately, the NCSA documentation always used URLs, so now we
are left in total confusion about what to name things.  In the midst of
that I wrote RFC 1808, which talks about relative URLs only (instead
of relative URIs) simply because that was all we could get consensus
on within that WG.

Later, when writing the HTTP spec, I decided that it would be silly
to restrict HTTP to URLs simply because the URN people hadn't found
a clue.  So, I defined all of the HTTP stuff as using URIs instead
of URLs.

In theory, this whole mess will straighten itself out when all the
specs start moving to draft standards. I am supposed to write a revision
of RFC 1808 at some point, but I'm really quite sick of the whole
process right now.

.....Roy

Mime
View raw message