Date: Fri, 30 Jul 93 22:04:44 MET DST
From: timbl@nxoc01.cern.ch (Tim Berners-Lee)
Message-Id: <9307302004.AA10977@ nxoc01.cern.ch >
To: masinter@parc.xerox.com
Subject: Re: URLs for non-Internet networks?
Larry suggests a URL for mail server access. This has been suggested
a few times. Problems include...
1. Sometimes the address, sometimes the subject, and sometimes
th ebody are used to encode the request in various combinations;
2. It is very slow compare to "Real OnLine" resources and
so people would be led into clicking on things which
they wouldn't have bothered with had they known; (minor
problem) and basically everyone with mail server will think
they have done their job of putting things on line
when in fact they could do much better;
3. When the mail comes back, it is difficult to filter it
out and extract it. If all mail servers guaranteed
to make use of a "In Reply to" field to tie
the response to the request, then an application
could in principle be hooked in so that when the
document had been once retrieved it was cached and then accessible
as part of the web. But it is really hairy. There is
no well-defined machine-usable protocol.
Apart from that, it is a great idea. Do we want to encourage it?
As for making MIME external=body and URL systems
isomorphic, that may be a reasonable aim..
or allow a URL to be specified in a MIME external body part.
Tim BL