Minutes from PWG IPP Phone Conference 980325
============================================

Notes taken by Carl-Uno Manros and Tom Hastings.

Attendees:

Ron Bergman    	Dataproducts		
Tom Hastings  	Xerox	
Carl-Uno Manros	Xerox	
Ira Mcdonald  	High North	
Randy Turner  	Sharp	
Don Wright    	Lexmark	
Jim Walker		DAZEL
Peter Zehler	Xerox
Harry Lewis		IBM
Jay Martin		Underscore

Agenda:

Carl-Uno Manros led the meeting. For today's agenda topics, he said that
he would like to revisit the IPP notification and the host-to-device home 
work assignments and DL discussions.

Status of IPP in the IESG and IPP LA Agenda:

Carl-Uno reported that contacts with Keith Moore make it clear that we will 
not see any feedback from him or the IESG in time for IETF41 next week,
and probably not before the PWG IPP meeting in Portland either. The
IETF41 IPP agenda is therefore limited to the discussion of IPP 
notifications and mapping of directory attributes to SLP and LDAP.

Carl-Uno then made a quick review of other sessions in IETF41 that
might be of interest to IPP WG members coming to LA.

Notifications:

Then the discussion of IPP notifications started. Carl-Uno suggested 
that we already have a number of inputs that could be used to work
out a solution for the simple kind of user notifications intended
in the IETF IPP work item. It was clear from the following discussion
that the PWG needs to view notifications in a wider scenario that
takes into account host-to-device, administrator etc. requirements,
but that this work could be done in parallel to developing a simple
notification solution for IPP 1.0. The notification specification is
expected to take the form of a standards track IETF RFC, which 
specifies additional attributes to what is in the Model document. 
Although the notification specification will be optional to 
implement, we should aim for it to become a standards track document.

Discussions were held around how much flexibility the user needs to 
have in specifying what should be returned in a notification. Some
suggested that standardized packages should be enough, but there 
were also requirements to have full flexibility to specify the exact 
attributes (including private attribute extensions) to be returned. 
Unless the solution gets too complex, it looks like we should expect 
a few standard packages to be specified and supported, but that 
implementations could optionally support attribute level specification.

There was also discussion about the format of the notification.
Everybody seemed to agree that it should be in the form of a MIME 
type, which is identical or very similar to the application/ipp
Mime type and should look more or less identical to the 
List-Job-Attribute response format. There also seemed to be agreement 
that the life cycle of a notification request is the same as for the 
print job.

As part of this discussion, Tom Hastings introduced the DL draft on 
how a directory attribute syntax could be defined. Bob Herriot and
Roger deBry had worked with Tom on this, but had not yet seen the 
current version, when had been written by Tom. There was opposition from 
Jim Walker and others about the parallel attribute value approach 
taken in the proposal and Tom undertook to work out a full proposal 
based on the original idea of introducing a new syntax for comparison.

Carl-Uno when summing up the discussion about notifications asked
for volunteers to write up a proposal based on the earlier text 
in the model document, the current requirements document, Tom's
revised proposal for the directory attribute, and today's 
discussion. Tom Hastings and Harry Lewis will try to have a draft
ready for discussion in Portland (April 8-9) with Jim Walker's review,
if there is time.

PWG Host-to-Device protocol discussion:

The next agenda item was the alternative approaches to arrive
at a suitable solution for a host-to-device protocol (which is 
not intended as an IETF standard).

Don Wright introduced his PWG draft called "IPP to IEEE 1284.1 
Mapping". This describes how IEEE 1284.1, a.k.a. TIP/SI, could be used
to transfer IPP attributes and document data from a print server
to a print device. A few minor extensions would be needed to 
the IEEE TIP/SI standard itself to support the proposal. In addition,
the PWG would develop a PWG standard for a IPP Logical Unit (LU) 
that works with TIP/SI that accepts IPP attributes directly.  
Areas for further study would be internationalization and security, 
plus an explicit mapping of TIP/SI to TCP/IP. Don wanted to have 
further discussion of the document in the Portland IPP meeting 
before investing more work in refining the draft.

Randy Turner then introduced his proposal on how to map IPP
directly on TCP/IP, which he claimed would solve most of the 
problems that have been stated against using IPP as a 
host-to-device protocol. The following discussion indicated 
that people wanted the two channel approach that is used in
TIP/SI and CPAP (one for bi-directional control information 
and one for data) to be added to Randy's proposal. Randy undertook 
to have a revised version of the proposal ready for discussion in 
Portland. It will introduce a second pull-only data channel 
and rewrite some of the proposal for a new URI scheme.

Meeting adjourned at 12:30 noon.

There will not be any phone conferences for the next two weeks 
due to the IETF41 and PWG IPP meetings. The next teleconference 
will be held on April 15 at 10:00am PST.