Date: Mon, 18 Apr 1994 16:47:07 +0200
Message-Id: <9404181447.AA05931@dxmint.cern.ch>
From: hallam@alws.cern.ch
Subject: mail *TO* ?
Hi,
mailto: protocol is giving me problems. I have a set of routines
to read mailfiles. So there is a natural tendency to want to be able to
use the same URL to read a message that was originaly used to create it.
Here I begin by inverting the hierarchy somewhat. Instead of considering
methods GET, PUT, POST etc as being HTTP ideas I make them the fundamental
operators regardless of protocol. Some protocols will only accept a limited
number of methods because only a limited number make sense.
In this context mailto: and mid: start to look fishy. The URLs refer to the
same objects, why have different protocol names? Why is the mailto: protocol
the only one that specifies a direction? Either it is implicit in which case
why bother with the extra letters or it is inaccurate in which case why
not fix it.
Surely we should unify these protocols to produce a mail: protocol. This would
allow the methods GET, PUT, POST and DELETE and allow a complete mail
interface to be created.
mailto and mid could still be kept but regarded in the same manner as NNTP,
as a lower level protocol that should be kept out of sight as far as possible.
Phill Hallam-Baker