Re: New Internet Draft: draft-ietf-uri-urc-trivial-00

Paul Hoffman (ietf-lists@proper.com)
Wed, 3 May 1995 11:42:21 -0700


Message-Id: <v02120c19abcd7ca2c975@[165.227.40.36]>
Date: Wed, 3 May 1995 11:42:21 -0700
To: Larry Masinter <masinter@parc.xerox.com>
From: ietf-lists@proper.com (Paul Hoffman)
Subject: Re: New Internet Draft: draft-ietf-uri-urc-trivial-00
Cc: uri@bunyip.com, "Ronald E. Daniel" <rdaniel@acl.lanl.gov>

Comments from Larry Masinter:
>Why not make this more MIME-ish? The whole
>
>====
>
>thing seems pretty klugy to me; it's enough like 'multipart' to make
>me wonder why you don't just use multipart, which has a well defined
>parsing strategy.

Our original intention was to make it easy for a naive user to create a URN
in the urc0 syntax. If they have to understand MIME, and have to type their
URN absolutely correctly, it defeats the purpose. We wouldn't mind an
additional URC syntax that was MIME-compliant; however, we wanted something
easy to type, difficult to type wrong, and easy for people and programs to
parse. Multipart MIME failed on a few of those criteria.

Comments from Alexander Dupuy:
>Just to save people some grief in the future, why not choose a delimiter that
>won't interact painfully with MIME quoted-printable encoding?  Really, almost
>any printable ASCII character other than = and - would work just as well, and
>in certain MIME-encoded situations, much better.

I agree that the "=====" may be a confusing choice for a delimiter
character given that it looks MIMEish. I chose it because it looked like a
separator and was unlikely to appear in the metainformation. I'm certainly
open to other suggestions (I had toyed with "#####" as well).

--Paul Hoffman
--Proper Publishing