Executive Committee Meeting Minutes
Total attendance: 20 of 24 voting members
|Since 75% of the EC's 24 voting members were present, the EC was quorate for this meeting|
The EC Standing Rules state the following penalties for non-attendance at EC meetings (note that those who participate in face-to-face meetings by phone are officially counted as absent):
There were no changes in status as a result of this meeting.
Patrick reported EC personnel changes (see the PMO Presentation for details.)
Patrick presented the usual EC stats. He congratulated the EC on the 100% voting turnout for JSR 344 - this is the first time we've had a 100% JSR vote turnout for a long time. Steve Winkler explained that he had been unable to vote on several JSRs due to a personal emergency, and asked whether it might be possible to have his Yes votes recorded after the deadline. Patrick explained that this question had been considered several times and that we always came to the same conclusion - that granting exemptions would put the PMO in the difficult position of making subjective judgements about when and how late such votes might be cast. Werner Keil commented that this is why members have alternates. John Weir explained that while seeking internal comments on JSRs (which might take a considerable amount of time) he will sometimes cast an initial Abstain vote so that if input could not be obtained in time at least a vote would be cast. Heather reminded members that votes are preceded by a review period that lasts a minimum of 30 days, so members do have advance notice. She also suggested that members follow @jcp_org on Twitter to receive notifications of ballots. Members made no decision to change the policy of strict adherence to voting deadlines.
See the new Action Item tracking file.
Ed Lynch led a discussion on the Java ecosystem. He explained that the presentation addressed two topics raised during the January meeting: the imbalance of Spec Lead responsibilities between Oracle and others, and the health of the ecosystem. The data in his slides was derived from jcp.org, and covered all completed and currently-active JSRs.
Ed showed how the amount of JSR activity in the JCP peaked between 1999 and 2002 and has declined ever since. He compared two snapshots of activity: one from "the first 12 months" and the other showing all currently-active JSRs. He noted that although the ratio of JSRs led by Oracle to those led by others has not changed much during this period, because the number of active JSRs has significantly declined, the absolute number of non-Oracle Spec Leads is therefore significantly smaller. He also noted that Oracle tends to lead all JSRs involving core language and VM changes and all platform and umbrella JSRs, calling on others to lead only those JSRs involving library changes and requiring specialized domain expertise. He asked whether we should be encouraging broader participation, suggesting that this would necessarily require a higher ratio of new to maintenance JSRs since revisions of existing JSRs are typically led by originator of the JSR, which tends to reinforce Oracle's dominance.
Calinel Pasteanu pointed out that the timing of the "snapshots" was such that the considerable diversity in leadership of Java ME JSRs was not reflected. He suggested that charting the numbers for each platform separately might provide a more accurate picture. Ed responded that the overall numbers still told a compelling story.
Patrick thanked Ed for providing a "fact-based" presentation, and asked whether he would distribute his raw data in spreadsheet form so that others could slice and dice it in different ways. Ed agreed to do so. [Later: Ed provided the spreadsheet. It is linked here.]
Mike Milinkovich suggested that the number of JSRs is not necessarily a useful indicator of the amount of innovation that takes place in the Java ecosystem. Steve Harris agreed, arguing that a great deal of innovation happens outside of the JCP and asking whether we should try to bring some of it into the Process.
Mike DeNicola argued that much innovation is stifled. He noted that in the early days JCP member companies had the expectation that the platform would be "jointly owned" or at least provide joint benefits. Instead, he argued, Sun and Oracle have relegated JCP participants to "advisory" status and exert their veto-power. Consequently, member companies look elsewhere for a more level playing-field, and participate in the JCP only in maintenance-mode to protect their investment. He reiterated that innovation happens in more open forums.
Mike Milinkovich pointed out that OpenJDK is also tightly controlled by Oracle, which owns all of the resulting IP.
Ed Lynch asked what we did to drive the innovation away into other forums. Scott Stark responded that the licensing and openness discussions that we are having in the context of JSR 358 are critical to this question. Ed asked whether we should be proselytizing for new JSRs.
Gil Tene suggested that we had a vibrant and creative period between 1999 and 2004, and noted that today almost all JSRs become part of the platform. He suggested that what we are seeing as a natural consolidation. Patrick agreed.
Steve Harris suggested that we might want to "bless" some open-source projects by assigning them an Incubator status in the hope that they might eventually result in JSRs being filed. Mike Milinkovich asked why the leaders and participants in such projcts would be willing to collaborate with us and assign control to others. He noted that many successful open-source projects do not have formal specifications or conformance test suites.
Patrick suggested that we might want to follow Steve's suggestion but would need to tread lightly. We could engage in outreach to teach people of the value of formal specifications and test suites.
Steve responded that these incubator projects would continue to operate independently and as open-source projects but he hoped that they would commit over time to standardization by formalizing their Spec and TCK. Their participation would indicate publicly that they hoped to become a standard at some point in the future. There would be no coercion - instead we should provide incentives. Mike Milinkovich again asked what motivation they would have to participate in such a program.
Werner Keil pointed out that we had previously discussed the value of some kind of incubation process, and noted that this might help open-source projects prepare themselves to file JSRs and also help those participating in OpenJDK work to ensure that their work is ready for inclusion in the platform.
John Weir suggested that our engagement with Java User Groups might provide a channel for this kind of program. Patrick responded that although there is some overlap JUG leaders and the leaders of open-source projects are different groups. He asked whether we should be doing outreach to open-source project leaders in a similar manner to our successful JUG outreach program. John agreed that this would be helpful. [After the meeting Heather Vancura noted that the JCP used to do this kind of outreach several years ago by attending popular open-source conference and by emphasizing the importance of the Spec and TCK.]
Ed Lynch summarized: we agree that the number of active JSRs is not a proxy for Java innovation and that much is happening outside of the JCP. Some of us believe that this is a problem. If so, should we be doing something about this and what indicators (in addition to the number of JSRs) should we be tracking?
Patrick pointed out that we only have about 30 paid corporate JCP members, which suggests that we do indeed have an engagement problem. Don Deutsch responded that fact-based analysis is important, and stated that Oracle acknowledges that we don't have the level of engagement we would like. He added that we are doing what we can within the constraints; Oracle wants a more vibrant, participatory community but we don't necessarily know how to get there. He agreed that doing outreach to "outside" communities that are innovating in Java seems a promising approach.
Patrick wrapped up the discussion by stating that we need to do more on this and that we should continue the discussion at future meetings.
Patrick summarized the discussions the IP Working Group has been having (see the java.net project for details) and then Steve Harris summarized his proposal on standardized RI and TCK licensing. Patrick and Don explained that they had taken these proposals to Oracle Legal and Execs who were sympathetic to the suggestions. They " didn't say no" but did note that it is critical that our core compatibility principles and practices be maintained.
Gil Tene responded that he had two concerns with the CloudBees proposal. Firstly, it made the platforms "special." Steve Harris responded that the existing Process Document language already makes the platforms special. Gil responded that the proposal goes beyond the current state and takes rights away from non-Oracle participants, arguing that in Java ME others also want the opportunity to monetize and commercialize their work. His second concern is the centrality of the platform JSRs. Azul cares most about having access to the platforms.
Werner Keil interjected that some Java EE Spec Leads believe that copyright applies only to the Spec and not to the JavaDoc. Patrick promised to get a formal response to this concern (which Werner has already logged in the JSR 358 Issue Tracker.)
Steve Wolfe suggested that Gil Tene and Scott Stark bring their own proposals to the IP Working Group. They agreed to do so.