Message-Id: <9305111953.AA12790@mocha.bunyip.com>
Date: Tue, 11 May 93 14:52:58 EDT
From: Richard W Wiggins <WIGGINS@msu.edu>
Subject: Re: Wrappers for URLs
To: uri@bunyip.com
In-Reply-To: Your message of Sat, 8 May 93 14:06:00 -0500
>.... How about let's bite the bullet on quoting spaces in URL's, as is
>currently done and as is currently deployed in existing systems, and
>move forward on URN's, so URL's can assume their more-or-less proper
>role as back-end mechanisms out of normal user sight as soon as
>possible. I'm ready to put developer time into building X/Mac/PC Mosaic
>clients that use URN's (and servers, if necessary), as soon as a spec
>is ready -- the sooner the better.
Throughout these discussions the assumption is made that URNs will
render URLs hidden, out-of-sight, process-to-process. Is this really the
case? It seems to me that we need both a handle for specific copies of a
document, and the generic ISBN-like name.
If I write a monograph or scan a photo or publish a visualization file,
and I want to announce same on a mailing list, it would be nice to have
an agreed-upon standard syntax to use. And it'd be nice if users reading
their news could click on the reference and their favorite browser would
go get the document. Let's say I don't expect my file to live very long,
to make it into various URN registries, and my readers really shouldn't
have to do a URN lookup just to grab this thing. (Rhetorical question:
will mapping from URN to URL be as fast as domain lookups, or as slow as
Archie searches?)
So it seems there is a need for humans to be able to pass these handles
around. With or without URNs, we need URLs that humans can quote and
can paste -- whether it's via two forms of URL, or one friendly one.
/Rich Wiggins, Gopher Coordinator, Michigan State U