From: eostrom@gac.edu (Erik Ostrom)
Message-Id: <9310161627.AA05901@gac.edu>
Subject: Re: The URN: wrapper and URLs...
To: ses@tipper.oit.unc.edu (Simon E Spero)
Date: Sat, 16 Oct 1993 11:27:16 (CDT)
In-Reply-To: <9310152352.AA10709@tipper.oit.unc.edu> from "Simon E Spero" at Oct 15, 93 07:52:03 pm
> Breaking existing code for cosmetic reasons is truly the ISO way.
Just for review, could someone explain to me what existing code we're
talking about breaking (that wouldn't be broken by the inclusion of
URNs anyway)?
I thought we were talking about a way of picking up and distinguishing
between URLs and URNs when it's not obvious from context--this would
be kind of a metasyntax, rather than anything to do with the syntax of
URLs (or URNs) themselves.
By "picking up" I mean something like "My web-aware newsreader should
be able to scan through a plaintext Usenet article, spot strings like
URL:http://jhm.ccs.neu.edu:7043/, and mark those as links to be
followed." You probably don't want to flag everything beginning with
some letters and then a colon; you also probably don't want your news
scanner to have to know about every kind of URL scheme (updating as
new ones are added). Encapsulation issue, I guess.
Other people have given reasons for wanting to know whether something
is a URL or a URN, which is of course what "distinguishing" means.
So I guess I'm confused about the "existing code" argument. I know
some people have written such things, and I don't have a strong desire
to make them work more, but I don't think they've got a wide installed
base. Most of the argument I've seen has been from Web people, so I
presume we're talking about some problem with Web software; but I'm
not sure which part.
I have some ideas, and some responses to them, but before I start
arguing I want a clear summary of what exactly we're arguing about
(both wrt "why/where do we need a URL: wrapper" and wrt "what code
will this break"). I suspect the discussion in general would benefit
from a restatement of the problem.