Re: revised mailto spec

Larry Masinter (masinter@parc.xerox.com)
Mon, 9 Dec 1996 18:14:49 PST


To: timbl@w3.org
Cc: liberte@ncsa.uiuc.edu, uri@bunyip.com, jwz@netscape.com, connolly@w3.org,
In-Reply-To: <32ACC075.5E15@w3.org> (message from Tim Berners-Lee on Mon, 9
Subject: Re: revised mailto spec
From: Larry Masinter <masinter@parc.xerox.com>
Message-Id: <96Dec9.181449pst."2694"@golden.parc.xerox.com>
Date: Mon, 9 Dec 1996 18:14:49 PST

Tim, I'm sorry, I shouldn't have snapped at you. In fact, the draft
doesn't say what I thought it said, or what I intended it to say. I
propose to revise the 'mailto' URL specification to clarify it.

It is my _intent_ to keep 'mailto' with the same semantics as it has:
it is a mail address, and does not imply automatic processing of mail
sending, and does not denote 'the resource returned by sending the
message'. The goal is just to extend 'mailto' to allow the
specification of additional headers and a default body in any mail
sent to the address.

I agree that "a mail address is a sort of object", but this doesn't
mean that one might have other objects that have a URL that consist
not only of a mail address but also a default subject, body, or other
message headers.

The goal of extending 'mailto' is not to blur the distinction between
'a mail address' and 'a resource retrieved through mail', but rather
to extend the capability of addressing 'a mailbox' to some additional
extended capabilities. It is true that there is a link relationship
between 
	mailto:foo@bar.com
and
       mailto:foo@bar.com?subject=current-issue
but it's also true that there is a link relationship between
       http://bar.com/foo
and
       http://bar.com/foo?subject=current-issue
or
       http://bar.com/foo#current-issue
-- it's just that there's a widespread desire (and some implementation
experience) to make this extension.

Larry