80 likes | 98 Views
This document proposes revisions to the presence information delivery registration process, suggesting the use of Q-values for MIME types and a new element under <presence> to signal tuple removal. It explores the option of utilizing the 'xcap' package for presence document changes and discusses the cons of this approach. The document explains the need for a distinct MIME type, handling of namespaces, identification of non-unique elements, and the complexity of the proposed solution. Feedback and review from the working group are welcomed.
E N D
draft-ietf-simple-presinfo-deliv-reg-00 Mikko Lönnfors mikko.lonnfors@nokia.com IETF#57, Vienna
Status • No comments received for this draft • Would be good if some people could review it • WGLC after review?
draft-lonnfors-simple-partial-notify-02 Mikko Lönnfors mikko.lonnfors@nokia.com IETF 57, Vienna
Changes since -01 • Use of Q-values with MIME types in Accept: header to indicate preferred Content type • Use of new <removed> element under <presence> element to signal removal of tuples
Current solution • In current solution <tuple>s are considered as atomic data elements -> Updates are send in <tuple> level • Presence document level info is always included • Document format is similar to normal PIDF • Should not have any major open items
One other alternative could be ‘xcap’ package • This was proposed in interim meeting • It might be possible to use ‘xcap’ package (or something similar) to deliver changes in presence document • The event package itself is not needed. Only the MIME type used to convey changes • Utilizes Xpath functions to identify element which has changed • Might allow sending smaller changes than current solution (in a level of any XML element/attribute)
Cons in using ‘xcap’ • Will need its own MIME type because current ‘xml-change’ contains elements/attributes which cannot be used: • URIs, these are defined as HTTP URIs. Version info • Handling of namespaces. At least in cases where notification contains information from namespace which hasn’t been used in previous notifications • Identification of elements which don’t have unique identifier (in cases where element can have multiple instances) • More complex solution than existing one • PUBLISH uses <tuple> as atomic element. Not consistent with it.
Way forward • Go forward with existing solution • If some other solution could be used please propose text • WG item? • Would be good if WG could review the document