Message-Id: <9402210242.AA07977@mocha.bunyip.com>
To: Larry Masinter <masinter@parc.xerox.com>
Subject: Re: reply to various on URNs and URLs
In-Reply-To: Your message of Fri, 18 Feb 1994 17:12:16 -0800.
<94Feb18.171222pst.2732@golden.parc.xerox.com>
Date: Sun, 20 Feb 1994 21:42:15 -0500
From: John Curran <jcurran@nic.near.net>
--------
] From: Larry Masinter <masinter@parc.xerox.com>
] Subject: reply to various on URNs and URLs
] Date: Fri, 18 Feb 1994 17:12:16 PST
] ...
] I think we should take seriously the role of `Location Independent
] File Name' (LIFN) that some people seem to have a desire for, and
] consider that, outside of the possibility that the same URN might
] identify two sets of bits that were not the same bits but considered
] the `same' by the publisher, that the requirements we've identified
] for URNs are the same as the requirements for LIFNs.
]
] LIFNs are useful for determining whether a cache is valid, as well as
] dealing with resource replication.
I've been a tad busy lately, but am now going over the past URI mail.
I agree that we need to consider the LIFN issue, but an unsure that
they are necessary at the "URN" level of hierarchy.
] I think we should consider whether we might have two kinds of URNs:
] those that act as LIFNs and those that do not. (I hate to introduce
] yet another kind of identifier in the UR... series).
I'm not sure that everyone considers a URN to be a pointer to a set
of "bits", and am concerned that most of the characteristics of LIFN's
seem to more closely resemble that of URLs. Replication and caching
at the URL level seems to be very important, and should not be a problem
aside from the subset of URL's which are highly dynamic by definition.
Do we need to include an "allowed-cache-interval" or TTL in those directory
service which return URLs?
/John