Re: telephone URLs, comments on draft-antti-telephony-url-04

Larry Masinter (masinter@parc.xerox.com)
Tue, 30 Jun 1998 02:17:10 PDT


From: "Larry Masinter" <masinter@parc.xerox.com>
To: "=?iso-8859-1?B?UGF0cmlrIEbkbHRzdHL2bQ==?=" <paf@swip.net>
Cc: <uri@Bunyip.Com>
Date: Tue, 30 Jun 1998 02:17:10 PDT
Message-ID: <000f01bda407$dc254bc0$15d0000d@copper-208.parc.xerox.com>
In-Reply-To: <v04011708b1be56240161@[192.168.111.25]>
Subject: RE: telephone URLs, comments on draft-antti-telephony-url-04

well, this is perhaps simple to solve.

I see that there are several reasons for using hierarchical naming,
including:

a) sometimes you can rename the parents and keep the children's
   names the same
b) using just the child name rather than the entire lineage is shorter
c) a subtree can be viewed as having multiple parents
d) with a hierarchy, you can elide/omit/default some components
   of the parentage explicitly, while naming others, and have
   that omission syntactically evident
e) synonyms for various levels of the hierarchy are possible

You're saying that (a) is important/criterial/most important/the
only important criterion. I believe (a) is important, but not the sole
justification for hierarchical naming, and that it is useful
for other reasons too.

You want to make (a) the definition of "is the namespace hierarchical"
and then decide whether hierarchical naming applies. I want to make
hierarchy a design option which can be employed whenever it is deemed
useful by the namespace designer, even if (a) doesn't hold.

Does that characterize the difference?

Larry
--
http://www.parc.xerox.com/masinter