URI and URN syntax

Daniel LaLiberte (liberte@ncsa.uiuc.edu)
Fri, 31 Jan 1997 11:28:42 -0600 (CST)

From: Daniel LaLiberte <liberte@ncsa.uiuc.edu>
Date: Fri, 31 Jan 1997 11:28:42 -0600 (CST)
Message-Id: <199701311728.LAA04310@void.ncsa.uiuc.edu>
To: "Ron Daniel Jr." <rdaniel@lanl.gov>
Cc: "Ryan Moats" <jayhawk@ds.internic.net>, <urn-ietf@bunyip.com>,
Subject: URI and URN syntax
In-Reply-To: <199701311547.IAA01443@acl.lanl.gov>

Ron Daniel, Jr. writes:
 > My current position on this is that we should follow 1630 where it claims
 > to speak for all URI schemes. This means we should put  % / # and ? into
 > the set of reserved characters for all URN namespaces. I think the
 > direction on = ; * and ! is less clear so I can go either way on those
 > characters.

I agree that we should follow rfc 1630, and push for clarifications
there if needed.  In fact, why not simply inherit that spec rather
than copy it into the URN syntax spec?  If a URN is-a URI (it better
be!), then URNs must inherit the URI constraints, although URNs can be
more constrained.

Tim B-L recently mentioned on the URI list that relative URIs may
contain ";name=value;name=value" pairs in his original plans, but I
don't know that this is actually followed in any practice, but the
characters are still reserved unless explicitly unreserved.  Instead,
we see "?name=value&name=value" used for a similar purpose, but the
structure after the '?' is merely a convention.

Daniel LaLiberte (liberte@ncsa.uiuc.edu)
National Center for Supercomputing Applications