PWG - Internet Printing Project
Conference Call - May 28, 1997


Attendees:

Don Wright - Lexmark
Peter Zehler - Xerox
Roger DeBry - IBM
Scott Isaacson - Novell
Carl-Uno Manros - Xerox
Tom Hastings - Xerox
Ron Bergman - Data Products
Randy Turner - Sharp
Ira McDonald - Xerox
Angelo Caruso - Xerox
J.K. Martin - Underscore
Bob Herriot - Sun
Jim Walker - Dazel
Jeff Copeland - QMS

Agenda

1)  Revised Charter
2)  JOB MIB/IPP Harmonization


REVISED CHARTER

Carl-Uno led the discussion on the updates to the IETF Charter.  There have been few comments on the 
mailing list.

An author is needed for the LPR mapping document.  Someone with a system background, with LPR 
experience, etc. would be the ideal author.  At this time, no one volunteered.  

Should the date for prototyping move from July out?  The group's opinion is that the July date is not 
realistic and should be moved out to at least August.

The group should plan on a bake-off of some type.  The biggest concern is how to actually connect it to the 
Internet and be able to test the clients and servers in a real Internet environment rather than a more 
controlled, contrived network.

Should we split up the PROTOCOL document into two parts?  One part that is actually protocol neutral and 
the other mapping to a specific protocol?  Roger Debry suggested putting the protocol neutral items into the 
MODEL document rather than creating another document.  Roger reminded the group about the difficulty 
of keeping multiple documents synchronized.  A consensus began to form that putting the protocol neutral 
information into the MODEL document was the right thing to do.  At this point Randy Turner expressed his 
concerns that fully separating the PDU from the transport would make a less efficient solution.  
Additionally, feedback from the Memphis IETF encouraged the group to put a stake in the ground and 
specify a complete protocol rather than attempt to be so generic and effectively specify nothing concrete.  
At the end, the group generally agreed to keep the MODEL document abstract and to create a PROTOCOL 
document that including encoding as well as the specific HTTP information.

A brief discussion on print-by-reference and its importance to the NC (Network Computer) environment 
was held.  This discussion was suspended to allow the Job MIB/IPP harmonization discussion to begin.

JOB MIB and IPP Harmonization

The discussion centered on the number of job states that should be supported.  There are basically three 
states: before, during and after printing.  These three states are then broken down further to provide more 
information.  This is a total of seven states:

Pending
Pending - Stopped
Processing
Processing - Stopped
Done - Abort
Done - Canceled
Done - Completed

JobState and JobStateReasons will be mandatory.  A separate JOB conference call will be scheduled to 
work on this in more detail.  Having harmony between the JOB MIB and IPP is necessary and a strong goal 
of the two groups.

Tom Hastings will write up a new proposal and most it to the mailing lists.

PROTOCOL DOCUMENT

The discussion on how much information to include in the PROTOCOL document resumed after the JMP 
discussion.  A protocol working group meeting was scheduled for June 17th at SUN in Menlo Park starting 
at 9:30 AM.

Comments on Randy's IPP Protocol Document:

1)  Where should the concept of levels of function be included? 
2)  Should the length fields be binary values or text?  Many of the participants on the call thought fixed 
length ASCII encoded text for length is preferred.  At the end, the opinions were still mixed but won't 
be changed right now.
3)  We shouldn't call the OPERATION and attribute but rather keep it at the first of the PDU and call it 
something besides attribute.
4)  A concern was raised over the concept of "unsupported-attributes."
5)  What should happen on a GetJobs request?  Everything?  A specific list asked for by the client?  
Something else?  Resolution: The clients get everything up to a specified limit using a filter provided 
by the client.  This will be specified in the model document.
6)  Values are unsigned
7)  Names are limited as specified in the model document.

Carl-Uno reported on the security work and announced a security conference call will be held on Friday.

Bob Herriot will set up a Monday conference call to discuss the protocol document in more detail.

The conference call ended at 6:50 PM EDT