IPP> IPPGET Issues 1-7

IPP> IPPGET Issues 1-7

McDonald, Ira IMcDonald at crt.xerox.com
Mon Aug 6 13:16:44 EDT 2001


Hi Michael and Marty,

I agree that on at least two recent IPP Fax telecons we DID agree
that the actual Job itself (with all of its readable attributes)
MUST persist until some brief period of time (15 seconds??) AFTER
the last relevant Job-level Event life expires (after the last time
that the Event could have been harvested with polling or wait-mode
Get-Notifications response).

IPP/1.1 badly underspecified the requirement for a Job retention 
period to support reliable job monitoring and job accounting
applications.  The PWG Job Monitoring MIB (RFC 2707) specifically
REQUIRES that all Jobs MUST persist for at least 15 seconds
after completion and has a default Job persistence of 60 seconds.
It's literally impossible to specify zero (0) seconds Job
persistence in the PWG Job Mon MIB (the integer range on the
object syntax declaration prevents it).

Cheers,
- Ira McDonald


-----Original Message-----
From: Michael Sweet [mailto:mike at easysw.com]
Sent: Saturday, August 04, 2001 7:25 AM
To: Marty Joel
Cc: ipp at pwg.org
Subject: Re: IPP> IPPGET Issues 1-7


Marty Joel wrote:
> ...
> MS> Yes, this should be the expected behavior, with the subscription
>     and job disappearing only after all events have expired.
> 
> <MJ> The subscription object must persist, not necessarily the job
>      object.  An implementation could store per-job subscription
>      objects that are not linked to job objects, and which are 
>      deleted at their expiration time, even if the job object has
>      already been deleted because its job retention and job history
>      phases have elapsed.

This prevents you from gathering the job-media-sheets-completed
attribute after a notification that a job is complete, for example.
I *thought* that we wanted a client to be able to query a job object
after receiving a notification, and that job object should only
disappear after all events referencing the job object expire.

> ...
> <MJ> Chinking isn't required for multipart, and I wonder if they
>      are even compatible.

They are, but of course you'll need HTTP software that knows how
to do it right (with the embedded content headers chunked like the
rest of the data...

-- 
______________________________________________________________________
Michael Sweet, Easy Software Products                  mike at easysw.com
Printing Software for UNIX                       http://www.easysw.com



More information about the Ipp mailing list