Re: semantics of host field in http URI

> The combination of RFC-2616 and RFC-2396 guarantees that
> http://www.w3.org./ has one globally unique meaning

No, it doesn't.  2396 says it is an Internet host name from DNS.
Whether or not it actually is an Internet host name from DNS is
unknown as far as the protocol is concerned, and certainly not
guaranteed.  There was a lot of wishful thinking stuck in some of
the prior specifications that turned out to be irrelevant to the
protocols being defined.

> , but the combination
> of RFC-2616 and rfc2396bis allows it to mean different things in
> different places (because in some places www.w3.org. might not be a
> domain name at all, according to rfc2396bis).  Is that relaxation
> intended, or does the HTTP spec need to be updated to require domain
> names?

Yes and no.  There is no relaxation because it is a fact of life no
matter what the specification said.  There is no need to update HTTP
because HTTP never assumes that it is a DNS name, regardless of what
might be implied by a careful reading of two soon-to-be-obsolete specs.

Adam, I understand this is a pet concern of yours, but you have yet
to explain why it matters.  We expect 2396 to change when we revise it.
We expect other protocols to take into account the revised 2396 when
they are revised.  In the mean time, we expect implementers to read
both specifications and be able to understand that changes to the
specification that make the specification better match how it is
implemented in the wild are a good thing.  All that matters to us
is that an implementation of 2616+2396bis is better than trying to
implement 2616+2396, and yet both will continue to interoperate
with existing implementations because the changes that were made
from 2396 to 2396bis reflect the way existing implementations
have already adapted to the changing Internet.

....Roy

Received on Sunday, 29 August 2004 03:18:30 UTC