Date: Tue, 08 Nov 1994 14:36:37 -0700 (PDT)
From: Ned Freed <NED@INNOSOFT.COM>
Subject: Re: x-uri draft
In-Reply-To: Your message dated "Tue, 08 Nov 1994 17:30:13 +0000"
To: Martin Hamilton <martin@mrrl.lut.ac.uk>
Message-Id: <01HJ8NT25IJI9VV6TV@INNOSOFT.COM>
> I was thinking in terms of the X-URI: potentially becoming URI: if
> there were any interest in it beyond experimental use. Do I have the
> protocol wrong here ?
First of all, there is no requirement that informally defined headers begin
with X-. The only rule is that X- header are ALWAYS informal.
As such, any specification that is likely to be considered for standardization
status should not use X- headers.
> | Second, the mechanism you've described isn't general enough. You now have
> | personal contact information in the header. But who is it for? The person or
> | persons listed on the From: line? The sender? The reply-to address? How about
> | contact information for recipients, so one recipient can find out about other
> | recipients?
> It would be the author of the message - the URI is analagous to the
> sender's signature. Perhaps it should be Author-URI: to make the
> semantics explicit ?
What happens when there is more than one author? RFC822 allows for this.
I would think that from-uri would be a better name choice, assuming you want
to use a separate header.
> | I would much rather see something along the lines of the RFC1522 approach to
> | handling non-ASCII character sets used to provide these services. A URI in th
> | e
> | personal name field associated with my address would be just right for my
> | contact information, regardless of whether I'm the message originator or
> | recipient.
> Interesting... It would be nice to be able to include the URIs with
> little or no mangling, so that users could copy and paste them if
> necessary. Do you think that would that be feasible with an RFC 1522
> type approach ?
I don't see why not. I believe the quoting conventions already available
for use in URLs can be used to advantage to deal with most of these issues.
One really interesting issue is how to handle long URLs. Some kind of folding
mechanism is needed, even if you use a separate header.
Ned