Minutes from PWG IPP Phone Conference 971119 1. Attending: Roger deBry Lee Farrell Tom Hastings Bob Herriot Harry Lewis Carl-Uno Manros Ira McDonald Xavier Riley Randy Turner Peter Zehler Steve Zilles The following subjects were discussed. 2. Game plan for finalizing the IPP Model & Semantics and Protocol Specification documents a. Close of Last Call comments, Tuesday, November 25 b. PWG Phone Conference, November 26: summarize the Last Call comments. c. Rest of Thanksgiving week and the beginning of 12/1 week: work on resolutions to comments in preparation for IPP LA meeting, 12/3. d. PWG IPP meeting in LA, December 3: discuss and agree on resolutions to WG last call comments. Prepare for IETF meeting in Washington DC the following week. e. IETF IPP meeting in Washington DC, December 10 or 11: present Last Call results and suggested resolutions. f. Second half of December: final editing and submission to the IESG. 3. Status of the Rationale document Steve Zilles will be able to make the agreed updates (mostly resynchronization with the latest versions of Model and Protocol documents) to this and have it sent as an Internet- Draft to the IETF secretariat before the IETF Washington DC deadline on Friday, November 21, 5:00 PM EST (2:00 PM PST). ACTION ITEM: Carl-Uno will issue a WG Last Call, when the document is available from IETF. 4. Write up of security discussion from last week Randy is still working on some text to reflect our previous discussion on security implications due to changes in the latest TLS spec. He will call Scott for help on which section (3.1.5 or 8) in the Model document to update. ACTION ITEM: Randy expects to have it out this week to the DL as part of the Last Call comments. 5. MIME type definition for application/ipp There might be a need to update our current description to allow application/ipp to be sent over ESMTP. We may need to allow the Model attributes that are transmitted as HTTP headers to be in the body when using ESMTP. ACTION ITEM: Ira will look into this further. 6. Suggestion for improved text on operation processing procedures (section 15.3 in the Model document) The contribution from Tom, Bob, and Scott on validation of attributes in operations was briefly reviewed and discussed. It was agreed not to add new response group, but instead to return unsupported Operation attributes and Job Template attributes in the same Unsupported Attributes group. The response group will be renamed in the protocol document to remove the word "job" from the name (and to agree with the Model document). As a consequence, we agreed that the names of Operation attributes and Job Template attributes shall be unique, i.e., the same name will not be used for an Operation attribute and a Job Template attribute. The same name can be used for Operation attributes and Job Description attributes when the Operation attribute is being supplied to initialize the Job Description attribute. ACTION ITEM: Tom, Bob, and Scott will make a couple of revisions and re-issue the proposal to the DL as part of the Last Call comments. 7. Discussion about length boundaries for text strings (from recent DL discussion) Everybody seemed to agree that we want to have maximum lengths for all attribute syntaxes (see section 4.1 in the Model document), including the 'uri' attribute syntax, even if HTTP has not set such limits (pointed out by Larry Masinter). We have to make it explicit what the lengths mean and whether they apply to server, client or both. We reaffirmed that IPP is intended to be implemented by low-end printers that don't spool, as well as devices and servers that do spool. Therefore, these length conformance requirements need to be carefully reviewed as part of the WG last call. We reaffirmed that the maximums for read-write attributes required a conforming IPP object to support the full maximum length without truncation. There was concern that the current maximum for the 'text' attribute syntax of 4095 octets was too large. A maximum of 1023 was suggested, but no consensus was reached. The only read-write 'text' attribute is the 'message' Operation attribute in the Cancel- Job operation. However, since this Operation attribute is OPTIONAL for an IPP object to support, a conforming IPP object SHALL ignore the attribute if it is not supported. But the IPP object SHALL accept the maximum length without truncation if the "message" attribute is supported. There was also agreement that the maximum length for read- only attributes NEED NOT be supported by conforming IPP objects. Read-only attributes are ones set by the implementation and/or the system administrator when configuring the system. The entire list is: "status- message" OPTIONALLY returned in a response, "job-state- message", "job-message-from-operator", "printer-location", "printer-info", "printer-make-and-model", and "printer-state- message". Support for all of these read-only attributes is OPTIONAL for as IPP object. However, when they are supported, we agreed that the Model document needs to agree on minimums that MUST be supported for these read-only attributes. It was suggested that the minimums should agree with the Job Monitoring MIB. ACTION ITEM: Tom to draft a proposal for the maximums for those attribute syntaxes that don't have a maximum and for minimums for all attribute syntaxes discussion on the DL as part of the last call comments. 8. Changes to the Model and Protocol documents since Boulder It was suggested that the list of changes to the Model and Protocol documents be reviewed at the LA meeting, just to re- confirm agreement on the changes. ACTION ITEM: Tom to review the changes that were in Scott's e-mail announcement of the posting of the Model document for completeness and send out the list of changes this week to help Last Call review and for the LA meeting. 9. Next Telecon, Wed, 11/26 It was agreed to run a phone conference next week from 1-3 PM PST (4-6 EDT), even though this is close to Thanksgiving, considering that the Last Call closes on Tuesday and we want to see what needs to be done in the way of preparation for the PWG IPP LA meeting. Note Takers, Carl-Uno Manros and Tom Hastings