Last Modified: 2005-10-31
The goal of this BOF was to confirm whether the work on XML diff algorithms taking place in the SIMPLE WG should continue in that working group, and whether it should be designed for general use and/or reviewed generally. The SIMPLE WG has XCAP drafts that are relevant and require operations included in those provided by a generic XML patch algorithm. Netconf and WIDEX were also mentioned as having related work and/or use cases as well but probably not precisely the same requirements. Jari Urpalainen presented the architecture of a generic XML patch algorithm and demonstrated it working: - http://opensource.nokia.com/projects/xmlpatch/index.html - http://www.ietf.org/internet-drafts/ draft-urpalainen-simple-xml-patch-ops-01.txt - http://validate.openlaboratory.net/patch/ We discussed whether character data within XML elements needs to be patched as well. Note that currently Jari's solution does not allow that but only allows replacing the entire enclosing XML element, and it would not be trivial or simple to extend it to allowing character data patching. Overall, we considered whether a fully general XML patch algorithm was needed, or whether individual use cases could be addressed separately, possibly with simpler approaches. While there are general use cases (patching large XML format documents on HTTP/WebDAV servers, source control scenarios and possibly WIDEX), there was not a large call for a general solution for immediate use. Thus, this implies that there's not a need to extend Jari's mechanism immediately to support text patching or other functionality not already covered, nor is it necessary to pursue this work outside of SIMPLE. It is possible that Jari's mechanism could be extended later although versions of content formats can be difficult to manage. We discussed whether this work ought to go on in the W3C. They've been pinged on this topic and suggested looking at XUpdate, but that work is dormant and didn't meet the requirements of at least the SIMPLE use case. Since there wasn't sufficient interest in pursuing a general solution outside SIMPLE, we ended by discussing the process within SIMPLE. A separate mailing list (to allow people to follow the XML patch work without having to filter out general SIMPLE work) was suggested and rejected by the SIMPLE chairs. Ted Hardie promised help from outside SIMPLE, and offers to relay relevant stuff to people who don't want to be on the SIMPLE list. |