From: ccoprmm@oit.gatech.edu (Michael Mealling)
Message-Id: <199310181415.AA22411@oit.oit.gatech.edu>
Subject: Summary of Re: URN wrappers and URL:
To: uri@bunyip.com
Date: Mon, 18 Oct 93 10:15:48 EDT
In-Reply-To: <9310180939.AA01330@www3.cern.ch>; from "Tim Berners-Lee" at Oct 18, 93 10:39 am
I think this is a good starting point for summarizing at least my
obviously biased viewpoint of the discussions:
It seems that we have two crowds here: those that think we can do everything
with the current URL format and those that think that we need other
UR* things floating around to handle meta-information.
The first group correctly asserts that given current network technology
that the URL format could very well handle everything. The second group
correctly asserts that within several years the then current technology
may break our format.
What I see as the first groups concern is that the second group (of
which I am a member) is adding something that will break current code
to gain nothing but a nice looking structure. I realize this is a
good concern to have I just think that the second group hasn't
communicated what they want the URL: wrapper for.
Also, some appear confused by the fact that we call the angle brackets
a wrapper as well. I don't think anyone was advocating dropping those in
favor of URL:. What I think the second group was advocating is this:
<URL:bla:string> as the total specification as far as unstructured text
goes.
Now, being part of the second group, I come to the part where I think
that the second groups arguments make more sense. I would like
someone from the first group to give there opinions on these and their
opinion on this whole summary. These are my arguments why URL: makes
sense :
1) orthogonal (This is least important since if it doesn't really gain
anything then it has no value.)
2) <URL:bla> allows robots to scan text easily for items and be able
to dicsern things that are locations and things that are names for
passing off to other servers.
3) URL: follows URN: which other URIs should follow as well. I think
this buys us a lot beyond #1 in that it makes everything a nice
template that is parsable (whether or not you know the protocol
in the URL) and can be used by other template based
network services to help integrate all network services together.
4) URL: is an identifier that shows the _purpose_ of what follows
without any net calls at all.
5) ....anyone else want to jump in here?
-MM
-- ------------------------------------------------------------------------------ Michael Mealling ! Hypermedia WWW, WAIS, and gopher will be Georgia Institute of Technology ! here soon via MIME. Your view of the Michael.Mealling@oit.gatech.edu ! internet is about to change completely!