Re: URL decisions in Seattle, & changes

dupuy@smarts.com (Alexander Dupuy)
Errors-To: listmaster@www0.cern.ch
Date: Thu, 31 Mar 1994 03:59:55 --100
Message-id: <9403310157.AA28621@brainy.smarts.com>
Errors-To: listmaster@www0.cern.ch
Reply-To: dupuy@smarts.com
Originator: www-talk@info.cern.ch
Sender: www-talk@www0.cern.ch
Precedence: bulk
From: dupuy@smarts.com (Alexander Dupuy)
To: Multiple recipients of list <www-talk@www0.cern.ch>
Subject: Re: URL decisions in Seattle, & changes
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
Content-Length: 1758
Content-Length: 1758

Dan Connolly wrote:

[many things which I agree with, but which are probably considered closed
issues now that everything has been resolved "in the flesh" at Seattle]

In particular, I feel the following sums up the problem with the decisions:

> This completely destroys the concept of a comprehensive URI syntax.
> The grammar must include all the special parsing features of each
> scheme, and it must be explicitly amended with each new scheme.

However, there are a few points which don't seem to have been completely
closed out:

> In future schemes, will '/' and '%2F' mean the same thing or different
> things? I gather that the answer is "it depends." This rules out the
> idea of having one algorithm for reducing a URI to canonical form.  So
> the question of whether
> 
> 	x-my-scheme:abc/def
> and
> 	x-my-scheme:abc%2fdef
> 
> can occupy the same slot in a cache isn't specified. Bummer.

I thought that one of the requirements for URLs was that it be possible to
compare them for equality.  It seems to me that a single algorithm for
canonicalization, regardless of scheme, is essential for providing this
functionality.

If everyone wants to roll over for the gopher requirement of using '/' and
'%2F' interchangeably, then that implies that other schemes may not interpret
them differently.  In particular, if a scheme like http wants to distinguish
between hierarchical and non-hierarchical uses of /, it will have to use some
scheme-specific quoting mechanism (do I hear %5C?) to mark non-hierarchical
uses of /.

I guess this makes life fractionally more difficult for Macintosh WWW servers
that want to use relative URLs, but anybody who can't hack it should just run
Gopher instead, where they won't ever need to quote their /s.

@alex