Message-Id: <9405170006.AA0034@notebook.aus.xanadu.com>
Date: Tue, 17 May 94 11:59:00 EST
From: avatar@notebook.aus.xanadu.com (Andrew Pam)
To: mitra@pandora.sf.ca.us
Subject: Re: URN Requirements
> The real question is whether something is named with a URL or a URN. A URN
> is independant of the retrieval method, so it would only work in Xanadu if
> the object has some existance outside of Xanadu, i.e. if a file was copied to
> a disk and shipped to someone else would it be recognisable as the same
> object.
> If Xanadu has (or rather will have :-) such an object then it could be named
> with a URN, if not then its really a URL, and that can be handled by
> allocating "xanadu:" as a prefix for URLs and whatever additional syntax the
> xanadu folks want.
Excellent point. While in Xanadu there is no such distinction (all identifiers
are golbally unique and globally valid like URNs, but are also used to retrieve
information like URLs), this is only true inside the Xanadu docuverse.
Therefore referring to them with a "xanadu:" URL should work, and Xanadu links
to information not in the Xanadu docuverse can be achieved with regular URIs,
presumably much as with WWW.
> My understanding of Xanadu is that it makes extensive use of quoting of
> portions of documents ( a weakness, IMHO of the WWW) this needs handling
> by specifying a URN (or URL) and a fragment, standardisation of fragments was
> bumped off the agenda of the URI group at the DC IETF, and I still think we
> will need to tackle it sometime.
Well, in Xanadu any (possibly discontiguous) span of bytes is a valid endset
for a link, so effectively Xanadu URLs would have to contain a "fragment list",
if you will.
Share and enjoy,
*** AVATAR ***
Andrew Pam avatar@notebook.aus.xanadu.com
Manager, Serious Cybernetics avatar@jolt.mpx.com.au
Coordinator, Xanadu Australia <http://www.aus.xanadu.com/>
P.O. Box 409, Canterbury VIC 3126 Australia gopher gopher.aus.xanadu.com