Nibutes from San Diego

Folks,

below the minutes from our OPES WG meeting in San Diego. Thanks to 
Hilarie for preparing them!!

Thanks,
   Markus

=============================================

OPES WG Meeting, August 3, 2004; San Diego, California; IETF 60
---------------------------------------------------------------

Markus Hofmann introduced the new co-chair, Tony Hansen.

The agenda was accepted without modification, and Markus Hofmann moved
on to the first topic, which was the document status.  There are six
documents in the RFC editor queue.  Two others are awaiting AD followup.

Hilarie Orman asked that all authors take time to review the last
round of modifications, mostly grammatical, and give their OK to the
RFC editor.

Markus Hofmann presented the new charter. The two items of work are an
OCP profile that handles SMTP messages, and the rules language, which
is to be based on the "P" language.

Hilarie Orman and others noted that there are other SMTP initiatives
within the IETF which might be related to OPES.  MARID, MASS,
LEMONADE, and FAX were mentioned.

Hilarie Orman noted that "endpoint" for a store-and-forward protocol
is not as well-defined as it is for client-server protocols and asked
that care be taken in looking at the IAB guidance and OPES 
architectural documents to make sure that SMTP gets appropriate treatment.

She also asked that the charter be worded to note that email integrity
and privacy were especially appropriate concerns for OCP over SMTP.

Abbie Barbir presented a discussion of OCP/SMTP use cases.  He noted
that one option would be to use MIME body parts to communicate
protocol and metadata information between the OPES processor and
the callout server.  He suggested that such parts could be added
to the message, for delivery to the recipient, in order to communicate
tracing information.  The sender could use this method to communicate
bypass information to the OPES processor.

Abbie Barbir then presented an overview of the P language.  Markus
Hofmann noted that P was for invoking services, not for writing the
executable part of services.  Discussion ensued about how to check
that P parameters matched the expectations of the service API.  Abbie
expressed preference for WSDL for describing the services and their
side effects.  Hilarie Orman expressed a preference for having P
support regular expressions in the lefthand side of rules, and for
having P support sequences of services.  Markus Hofmann noted that
error handling might be difficult.