Web Services PMC Report for June 2008
New PMC members - Rich Scheuerle, Jeff Barrett, Bill Nagy, Jarek Gawor, Amila Suriarachchi
We had several new releases, including Axis2 1.4, Axiom 1.2.6, Rampart 1.4, Axis2-c 1.4.0, and WSS4J 1.5.4
Two discussions in the Web Services PMC were of particular note to the board since our last report. First, in the midst of voting on the new PMC members listed above, we got into a conversation about the status of our PMC and the fact that we have a bunch of PMC members who have been inactive for a loooong time. As a result, I sent out a "probe" message asking for the PMC membership to please respond if still interested in being a part of the PMC. Most of the members responded, but a number did not. It was our plan to "emeritize" the folks who didn't respond, but I haven't yet followed through with this. I believe my current plan is to send personal mails to those who haven't responded and confirm that it's ok for us to remove them from the PMC for now. Of course if they wish to return at any time they'll be welcomed with open arms. Just wanted to run this past the board for comment.
The second interesting discussion revolves around some new code that WSO2 has written that has recently been contributed to the Sandesha2 (WS-ReliableMessaging implementation) subproject. The new code is an alternate RM implementation built from scratch known as "Mercury". Mercury is built with an entirely different, state-machine-based core model, and solves some existing problems with Sandesha2. WSO2 has already released a version of this code, and needs to continue maintaining it, while at the same time we'd very much like to see the developer cycles and the codebases from both efforts merge into one if at all possible. After some back-and-forthing, it seems like the current plan is to a) get the Mercury code available to everyone for reference (done), and b) start work on a new implementation, with participation from all interested parties, that will carry forward the good architectural ideas from both Sandesha2 and Mercury without necessarily relying on either codebase explicitly. This will likely be done as a "revolutionary branch" of Sandesha2, and might become Sandesha3 or might be a newly named WS subproject.
Deepal's new book on Axis2 was published.
The team voted to move all future development of Axis2 to Java 1.5, dropping the requirement for 1.4 compatibility. This had been discussed several times previously but we're finally biting the bullet and making the switch.
Preparing for a Sandesha2 1.4 release built on Axis2 1.4 and Rampart 1.4 containing many stability and performance fixes in a branch. The trunk has now moved on to match Axis2 trunk in being Java 5 based.
In April, Woden released milestone 8 in support of Axis2 1.4. Woden has since seen new contributors come on board to tackle updating the Woden WSDL 1.1 -> 2.0 converter (Jeff Maury) and the Woden OM model (Vikas Gujjar). Woden's validation framework has seen some work recently. There has therefore also been some increased activity in the implementation of the validation logic for the WSDL 2.0 assertions. Woden was also recently mentioned in the developerWorks article Describe REST Web services with WSDL 2.0 by Lawrence Mandel.
No notable happenings since the last board report.
Rampart 1.4 was released on Jun 12 with lots of bug fixes and some new features. New features includes WS Security 1.1 features (Encrypted headers , Encrypted SHA1 references ), WS Security Policy 1.2 support, WS Trust features ( Renew binding,Validation binding ) ,New configuration options ( Ability to plug in Custom policy validators, Ability to define certAlias different from username for signature , Ability turn off mili second precision ) and ability to secure messages faults. Rampart 1.4 is tested thoroughly for interoparability with other WS Stacks.
Apache Rampart also got a Google Summer of Code project this summer. http://code.google.com/soc/2008/asf/appinfo.html?csaid=FF8E81CA884F19CA
The jUDDI v2 code is close to being done. We need to go through a round of sanity checking before we can do the final 2.0 release. In the mean time coding for v3 has started. jUDDI version 3 is slated to support UDDI v3.
The bug list got shorted, and we are getting closer to the final 1.0 release.