version 0.2: December 22, 2010
December 7 , 2010
Total attendance: 13
Total attendance: 11
|Since 75% of the ME EC was present, that EC was quorate for
Since 75% of the SE/EE EC was present (there being only 14 members at the moment), that EC was quorate for this meeting
Patrick reported several personnel changes and noted that with Tim Peierls' resignation there were now two seats to fill on the SE/EE EC: one ratified seat (originally held by Doug Lea, to which Hologic was nominated but not elected) and one elected seat (vacated by Tim.) He noted that a third seat would open up if Apache followed through with their threat to resign. He pointed out that this would necessitate a special election soon, but noted that the date for the election had not yet been set.
Werner Keil asked whether Tim's seat would be filled by another individual. Patrick responded that there are no quotas for individuals on the EC - anyone (corporation, institution, or individual) can nominate themselves.
Patrick welcomed Stefano Andreani to the ME EC and asked him to introduce
himself. See here for
a brief biography of him.
Patrick reported on the results of the four Java SE ballots, all of which were approved:
He noted also that a fifth new JSR was approved:
Patrick apologized for Oracle's late response to questions raised by EC members
during the ballot period and reported that Oracle is working to become
more "agile" in
answering community questions.
Sean Sheedy expressed disappointment that the ballot wasn't delayed to allow for more socialization and discussion, suggesting that it could even have been withdrawn and re-submitted later. He noted that many of those who voted "yes" expressed reservations in their comments, and suggested that these issues might be raised again when the JSRs reached the final approval stage.
Mike Milinkovich stated that unless the modularity implementation in Java
SE 8 accommodated the OSGI ecosystem Eclipse would have to vote "no" on that
JSR later in the development cycle.
Scott Jameson stated that HP did not expect to see significant changes in the licensing terms.
John Rizzo said that EC members wanted to see all possible licenses that Oracle might offer. Patrick responded that the Spec Lead had no obligation to publish them all, pointing out that Oracle has disclosed the "standard" licenses that are available to everyone. John responded that if the Spec Lead can offer special terms to a licensee then this amounts to discrimination. Patrick replied that since the disclosed licenses are available to everyone this is the exact opposite of discrimination.
Mike DeNicola seaid he hoped Oracle will take the licensing concerns under consideration, and improve them. Patrick responded that the JCP.next working group could discuss the changes members wanted to see, but cautioned that they shouldn't expect too much. John Rizzo said that EC members simply wanted Oracle to follow the rules that everyone else is following.
Josh Bloch stated that the press release from Oracle that morning, claiming "overwhelming majority" support, was exaggerated and misleading since the JSRs received only 12 votes, and many of those who voted yes expressed serious concerns in their comments. Patrick responded that Oracle executives were well aware that they needed to do a better job of communicating.
[John Rizzo requested the addition of this agenda item despite Patrick's concern that the discussion would be one-sided, since he did not have time to bring the relevant Oracle people into the discussion.]
John Rizzo stated that MIDP3 (JSR 271) had a reciprocal licensing agreement that was overwhelmingly approved by a 14-1 vote of the ME EC. Patrick noted that Sun abstained in the final vote. [Later review revealed that Sun abstained because of their serious concerns about the licensing terms, which they discussed in their comments.]
John said that the license terms for JSR 135 were identical to those of MIDP3, yet Oracle's lawyers have said that the terms are unacceptable. He argued that the JCP's Process Document doesn't mandate a legal review and that Oracle shouldn't be able to block the release of the RI and TCK. He reiterated that the license is identical to one that has been accepted by the community and asked the PMO to post the final materials for Maintenance Release. He stated that whether or not they did so, the materials are available on Aplix's website and he encouraged developes to download them. He repeated that Oracle had no right to veto the finalization of a JSR.
Mike DeNicola responded that If Oracle's lawyers are reversing decisions previously made by Sun's lawyers then implementors could not know whether existing licenses are still considered legitimate, and this is unacceptable.
Roger Riggs pointed out that the change logs for JSR 135 gave no indication that the licensing terms had been radically changed from the previous release and suggested that the Process Document ought to require such notice. John Rizzo responded that Aplix sent the materials early to the PMO for review, and that they assumed the EC wouldn't have a problem with the license terms since they are identical to a previously-approved license.
Roberto Chinnici presented the Java EE team's JSR plans
focussing on JPA
2.1 and JAX-RS 2.0.
Josh Bloch asked whether Roberto had said that he believed multiple open-source implementations are a good thing, and whether this applies to all platforms. Roberto responded that he couldn't speak for the other platforms but that this had worked well for Java EE.
Roger Riggs asked whether the client-side of JAX-RS be implemented
on other platforms. Roberto responded that it probably could, though it was
not architected that way now.
Jason Gartner asked how much demand there is from the developer community for an EE7 release, particularly since SE7 is only "half a release." Roberto responded that they are getting good feedback and that there is strong demand for JMS, Servlets, and Web-sockets. He argued that a platform release is a natural step and preferable to piecemeal changes. Jason responded that these proposals would make a large platform larger at a time when people are looking for more agility. He argued that there has been no significant adoption of EE6 yet and wondered whether Roberto was moving too fast in proposing EE7. Roberto responded that adoption is always slow and argued that if they waited for significant adoption of the previous release before starting work on the next there would be a 5 or 6 year release cycle. Jason suggested that it might be better to focus on the individual JSRs and not on the platform. Roberto replied that it's reasonable to have alternating small and big releases. EE7, being based on SE7, would be small while EE8 - including modularity - would be big.
Werner Keil asked whether there would be any modularity changes in EE7, perhaps based on OSGI. Roberto responded that modularity will be introduced in Java SE 8, while Java EE7 will be based on SE7. He did acknowledge that they might make some changes in anticipation of EE7 apps running on EE8.
Patrick invited Roberto to come back during the January meeting to tell the ECs more about his plans.
Patrick reported that the first meeting of the JCP.next Working Group will be held during the week beginning December 13. He promised to poll the members for a suitable time. It was also agreed that a face-to-face meeting of the Working Group would be held on the Wednesday afternoon after the full face-to-face EC meeting in January.
Patrick discussed an attempt at "ballot stuffing" that occurred during the October 2010 elections (see the PMO presentation.) He reported that a company that was running for an elected seat encouraged its employees to join the JCP as individuals so they could vote for their employer. The attempt was unsuccessful, partly because the applications were not processed in time, but he warned that someone might try this again. He noted that there is nothing that the PMO can do to prevent such manipulation but promised to shine the light of publicity on any offenders. He suggested that EC members consider a modification to the Process Document to prevent this kind of behavior.
2011 meeting schedule
EC members discussed the 2011 meeting schedule (see the PMO presentation,) agreeing in principle to face-to-face meetings in Korea in the Spring and at AT&T in Atlanta in the Fall. Patrick agreed to talk to the Korean EC members (neither of whom was present at this meeting due to the time-difference) to discuss hosting an EC meeting, and to poll EC members for the best times for face-to-face meetings. He also asked members to let him know whether any of the suggested conference call dates would be inconvenient for them.
EC members briefly discussed how they might engage with JCP members during the year. Patrick suggested a "meet the EC" meeting during JavaOne in October, and the possibility of meeting with local JUG members at some point during our face-to-face meetings.
Jason Gartner suggested that the JavaOne conferences held outside of the US would be a good opportunity to meet with JCP members. Patrick noted that the plans for 2011 seemed to be limited to Russia and India, and that there was not a significant JCP presence in either of these countries. He agreed that the ECs should watch the JavaOne plans closely, and should seize the opportunity to "piggy-back" on JavaOne meetings wherever this made sense. (The next time these events are held in Brazil and China, for example.)
Jason also suggested that the September face-to-face meeting include a session during which Oracle would preview its JavaOne announcements to the ECs, obviating the need for a special meeting like the one that was held prior to JavaOne 2010.