W3C home > Mailing lists > Public > uri@w3.org > August 2004

semantics of host field in http URI

From: Adam M. Costello BOGUS address, see signature <BOGUS@BOGUS.nicemice.net>
Date: Mon, 9 Aug 2004 06:11:09 +0000
To: uri@w3.org
Message-ID: <20040809061109.GB7082~@nicemice.net>

I haven't had time to keep up with the URI developments, but the recent
progress prompted me to check on the status of a pet issue.

Consider the URI http://www.w3.org/.  What is this "www.w3.org" and what
does it refer to?

The HTTP spec does not say that this field is a domain name, it says it
is a host token as defined in RFC-2396, and RFC-2396 says that a host
token is either an IP address literal or a domain name (and in this
case, we can rule out IP address literal based on the syntax).

Okay, great, the thing in the URI is without doubt a domain name, so I
can send this URI anywhere and expect it to be interpreted the same way,
because there is a single global domain name space.

Now suppose RFC-2396 gets obsoleted by draft-fielding-uri-rfc2396bis-06
(after it becomes an RFC).  Do the references from the HTTP spec to
RFC-2396 continue to refer to RFC-2396, or will they be understood to
refer to rfc2396bis?  If they refer to rfc2396bis, then we need to look
at the new definition of the host token in rfc2396bis, which goes out of
its way to state that there are other possibilities for the host token
besides IP addresses and domain names:

    In other cases, the data within the host component identifies a
    registered name that has nothing to do with an Internet host."

    A host identified by a registered name is a sequence of characters
    that is usually intended for lookup within a locally-defined host or
    service name registry, though the URI's scheme-specific semantics
    may require that a specific registry (or fixed name table) be used
    instead.

This implies that the meaning of http://www.w3.org/ might change
depending on where it is interpreted, unless the HTTP spec requires the
use of domain names, which it currently does not (because it depends on
RFC-2396 for that requirement).  Does the HTTP spec need to be updated
to explicitly require domain names?  Or is it intended to relax the
semantics of http URIs and allow http://www.w3.org/ to mean different
things in different places?

(Of course the same issue is likely to arise for other URI schemes; http
is the one I've checked.)

AMC
http://www.nicemice.net/amc/
Received on Monday, 9 August 2004 06:11:11 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:08 UTC