At today's IPP telecon we discussed the 'multipart/report' feature that I
proposed for the 'mailto' Delivery Method. (The details of the original
proposal are repeated at the end of this email message.)
There were strong feelings on both sides. The telecon compromise was to
change the entire proposal (the 'multipart/report' feature and the
accompanying "notify-mailto-report" attribute) from required to optional.
This change requires an additional optional "notify-mailto-report-supported"
If the "notify-mailto-report-supported" attribute is not present in the
Printer or has a value of 'false', the Printer does not support the
"multipart/report" feature. If the "notify-mailto-report-supported"
attribute is 'true', the Printer supports the "multipart/report" feature.
Someone in the telecon stated that this compromise is the "ecumenical
solution". This compromise allows those who believe in this feature to
implement it. Everyone else can ignore it. Furthermore, it ensures
interoperability for vendors who decide to add this feature some time in the
We are now soliciting comments about this proposal from the DL.
From: Herriot, Robert [mailto:Robert.Herriot at pahv.xerox.com]
Sent: Monday, August 14, 2000 2:20 PM
To: Manros, Carl-Uno B; IETF-IPP
Subject: IPP>NOT mailto feature from IETF meeting (RE: IPP> ADM - The
IPP Notification I-Ds will now go the IESG)
At the recent IETF meeting Ned Freed (our AD) asked whether we could send
machine readable information with the "mailto" Delivery Method. He
suggested that "mulipart/report" could be used to send machine readable
information, and he left it to the IPP WG to decide whether such a feature
should be added. I then worked out a very simple design using Ned's ideas.
A use scanario is a browser with a plug-in which uses IMAP or POP3 to get
and display email that represents Printer Event Notifications (the plug-in
uses the Content-Type to separate the email).
So I am asking the IPP WG two questions (please respond to the DL):
1) Should we augment the "mailto" Delivery Method so that it can send
Machine Consumable information. That is, does it add value?
2) If so, is the proposed solution acceptable?
Before answering these qustions, please read a quick summary of the solution
below. The solution is very simple. The text in the spec leverages existing
features and I expect the code would too. The URLs of the changed document
are at the end of this discussion.
A new Subscription Template attribute "notify-mailto-report (boolean)" is
When its value is 'false' the Printer behaves as it did before this
attribute was defined.
When its value is 'true', the Printer sends a message body whose
Content-Type is 'multipart/report'. According to the definition or
'multipart/report'(RFC 1892), the first body part of the 'multipart/report'
contains the Human Consumable information and the second body part contains
the Machine Consumable body part. So, the first body part contains whatever
the entire message body would contain if the value of the
""notify-mailto-report" were false. The second body part contains the
message body of a Send-Notifications request for the indp Delivery Method.
The Content-Type contains two parameters: the RFC specified parameter
"report-type=application/ipp", and a new parameter
"report-content=ipp-notify". A browser plug-in can search for the latter
A Printer MUST support the "notify-mailto-report" attribute and the default
is "false". Thus there is no need for "notify-mailto-report-default" and
"notify-mailto-report-supported" attributes. Furthermore the code to support
this feature is a trivial combination of the indp and Human Consumable
Below is an example taken from the spec:
When the Printer jams, the Printer generates and sends the following email
The Machine Consumable body part below is represented in a symbolic manner
with the following characteristics:
a) Fields that specify length of the following attribute name or value are
b) Other binary data is enclosed in angle brackets with the symbolic name or
2 hex-digits per octet.
c) Commas separate fields when an angle bracket is not present to delimit
d) The '<>' mean empty octet-string
e) Comments occur between the ';' and the end of the line.
Date: 29 Aug 00 0832 PDT
From: tiger <printAdmin at abc.com>
Subject: printer: 'tiger' has stopped
To: pwilliams at abc.com
Printer tiger has stopped with a paper jam.
<0101> ; Version 1.1
<001D> ; operation Send-Notifications
<00000000> ; request-id
<operation-attributes> ; tag for operations attributes
; the 2 lines below contain a syntax type,
; an attribute name and an attribute value
<event-notification> ; tag for Event-Notification Attributes Group
; each line below contains a syntax type,
; an attribute name and an attribute value
<text>notify-text,Printer tiger has stopped with a paper jam.
<end-of-attributes> ; end of attribute tag
The full details are at