RE: Rules Language

"Eric Burger" <eburger@brooktrout.com> Wed, 10 August 2005 09:20 UTC

Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Subject: RE: Rules Language
Date: Wed, 10 Aug 2005 05:14:00 -0400
Message-ID: <330A23D8336C0346B5C1A5BB19666647B63E12@ATLANTIS.Brooktrout.com>
Thread-Topic: Rules Language
Thread-Index: AcWdaL8jy+kXMH27TGubAvutYoyQRwAIjGWg
From: Eric Burger <eburger@brooktrout.com>
To: "The Purple Streak, Hilarie Orman" <ho@alum.mit.edu>, hofmann@bell-labs.com
Cc: ietf-openproxy@imc.org, ietf-mta-filters@imc.org
Content-Transfer-Encoding: 8bit
Sender: owner-ietf-mta-filters@mail.imc.org
Precedence: bulk

It would be really, really, really good if we could wrap ourselves
around SIEVE.

The mobile messaging community wants to be able to do OPES like stuff on
content that is streamed out of a server, either by SIP or HTTP.

At the same time, the messaging community is looking at SIEVE as a
potential solution for message-based sorting, i.e. "kill if Spam Score >
90%" or "send me a page if I get mail from my boss."

Being able to use the same language everywhere would be helpful.

Note that SIEVE, while created in the SMTP and IMAP milieu, is
specifically not tied to any transport mechanism.  One of its current
shortcomings, or benefits, depending on your point of view, is the
management of rules is currently undefined (probably will become a work
item somewhere).

I would offer that OPES should get SIEVE to do whatever is necessary to
update it for streaming.

See:
http://www.ietf.org/html.charters/sieve-charter.html

-----Original Message-----
From: owner-ietf-openproxy@mail.imc.org
[mailto:owner-ietf-openproxy@mail.imc.org] On Behalf Of The Purple
Streak, Hilarie Orman
Sent: Wednesday, August 10, 2005 12:58 AM
To: hofmann@bell-labs.com
Cc: ietf-openproxy@imc.org
Subject: Rules Language


With some trepidation, I offer

http://www.purplestreak.com/ietf-opes/draft-ietf-opes-rules-language-hop
along-00.txt

It is incomplete in several aspects, particularly the detailed grammar,
but I think that it gets at the heart of defining a stream processing
language for OPES.  I had intended to use the Sieve grammar, but
in the time I had for this, I couldn't decide if it is worth the
trouble of using the funky notation.

Comments are, of course, welcome, even more so than no comments.

Hilarie

RE: Rules Language Eric Burger