relative URLs, a canonical interpretation?

"James (Eric) Tilton" <>
Date: Wed, 11 May 1994 01:27:11 +0200
Message-id: <Pine.3.89.9405101650.A4890-0100000@jupiter>
Precedence: bulk
From: "James (Eric) Tilton" <>
To: Multiple recipients of list <>
Subject: relative URLs, a canonical interpretation?
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
Content-Type: TEXT/PLAIN; charset=US-ASCII
Content-Type: TEXT/PLAIN; charset=US-ASCII
Mime-Version: 1.0
Mime-Version: 1.0
In putting together my "Composing Good HTML" document, one suggestion 
that was made to me was to include a discussion on relative URLs.  I 
tried doing this, but realized that I had a major problem -- namely, that 
I couldn't find any canonical reference on what is safe and what is 
unsafe.  Does such a reference exist?  What relative URLs are guaranteed 
to work (i.e., within the same directory, subdirectories, on the same 
server), and what kinds of URLs will actually trip up the browser?  I've 
run into cases where different browsers will interpret relative URLs 
differently, for instance.

Perhaps there's a canonical test set? (heh.)