Re: IMAP URLs

Tim Berners-Lee (timbl@w3.org)
Tue, 26 Nov 1996 13:24:27 -0500


Message-Id: <329B35DB.417E@w3.org>
Date: Tue, 26 Nov 1996 13:24:27 -0500
From: Tim Berners-Lee <timbl@w3.org>
To: Larry Masinter <masinter@parc.xerox.com>
Cc: Chris.Newman@innosoft.com, uri@bunyip.com
Subject: Re: IMAP URLs

Larry Masinter wrote:
[...]
>         draft-fielding-url-syntax-00.txt
> 
> to see if the (revised) description of generic URL schemes clarifies
> those things that were ambiguous from RFC 1738 & RFC 1808.

	It also seems to change some things.

> # I also believe there is a bug in RFC 1808 with respect to multiple
> # parameters.  I think the relative URL should inherit all parameters form
> # the base URL, except those that are specified in the relative URL.  The
> # current spec indicates that if *any* parameter is present in the relative
> # URL, it overrides all parameters in the base URL.  Comments?

	Philosophically the idea that a relative URL provdes a delta
	only allows you the interpreation that you should inherit
	all parameters except where specified.

	Because the ; syntax (for matrix-like spaces) was added after the 
	/ (for hierarchical-like spaces), I imagine that most client software
	parses the parameters as part of the filename. In fact, I wonder
	whether it i distinguished at all. (Try href=";c" in "/a/b" - do
	you get "/a/;c?") 

> The goal originally was to revise 1738 & 1808 to make them consistent
> with current practice. What would current web browsers do with
> relative URLs with multiple parameters?

	It may be that (as noone has been sure enough) no one has
	actually used them, of course, which would leave the option
	of specifying the philosophically clean case.

> Larry

Tim