Message-Id: <9310210538.AA16568@mocha.bunyip.com>
To: Simon E Spero <ses@tipper.oit.unc.edu>
Subject: Approx. discussion status (was: Re: URN: wrapper)
In-Reply-To: Your message of Wed, 20 Oct 1993 23:51:27 -0400.
<9310210351.AA22893@tipper.oit.unc.edu>
Date: Thu, 21 Oct 1993 01:38:04 -0400
From: John Curran <jcurran@nic.near.net>
--------
] From: Simon E Spero <ses@tipper.oit.unc.edu>
] Subject: Re: URN: wrapper
] Date: Wed, 20 Oct 93 23:51:27 -0400
]
] With regards to the URL: wrapper - it seems that I may have gotten the
] wrong end of the stick from Kevin's original message. Having spoken to Jim,
] I now understand that Kevin was not suggesting changing the syntax of the URL
]itself, and merely talking about a different syntax for the plain text wrapper.
] There are absolutely no plans to make anything other than minor editorial
] changes- thus URLs should be polished off pretty quickly. ...
For URL's:
I believe that there is (very) rough consensus that: The URL does not
require a "wrapper" in common use (e.g. in-memory storage, APIs, etc.)
but that some wrapper was highly recommend when URL's appear in "simple
text" so that the reader can easily detect and determine the boundaries
of a URL in text.
There's been some discussion about wrapper syntax, with <> historically
favored although somewhat problematic for SGML users. Several other
delimiters have been suggested as sacrificial offerings, with no strong
consensus.
The use of a "url:" prefix as part of the URL proper has also been
raised by several folks are a desirable feature. There have been many
voices both for and against on this issue. Proponents are arguing for
a strong tagging so that the type of a URi can be identified, opponents
point to the compatibility problems which would be created and lack of
a real application which would benefit from the change. Rough consensus
has not been reached, although it may be close at hand.
For URN's:
There is active discussion of whether the URN _requires_ a wrapper,
as per the current spec. Consensus, if ever present, is rather light
regarding this feature.
There is general consensus that URN's are hiearchically organized,
with a publisher id (potentially a domain name) and publisher document
name contained in the URN. Potentially a namespace identifier may be
present, although this is also being discussed.
Case-sensitivity is recommended in the spec, although under discussion
on the mailing list. Character set identifiers are also under active
discussion.
There is general consensus on the use of a "urn:" prefix for URN's.
There has been much discussion about URN resolution, and attributes for
URL's and URN's. This discussion is still in the early stages and sharing
of ideas is taking place slowly but surely.
Does anyone feel that the above notes are not representative of the
discussions that have taken place recently on this list?
/John