Find JSRs
Submit this Search


Ad Banner
 
 
 
 

JCP EC minutes: October 22-23 2015

Executive Committee Meeting Minutes
for October 22-23 2015

version 0.1: December 2 2015

Date

October 22-23 2015

Location

Face-to-face, San Francisco CA
Hosted by Twitter

Agenda

Attendance

Thursday October 22

PMO
  • Patrick Curran
  • Heather Vancura
Executive Committee
  • ARM – Paul Manfrini – present
  • Azul Systems – Matt Schuetze – present
  • Credit Suisse – Susanne Cech present
  • Eclipse – Mike Milinkovich – present
  • Ericsson – Johan Mårtensson – present
  • Freescale – Maulin Patel – present
  • Fujitsu – Mike DeNicola – present
  • Gemalto M2M – Thomas Lampart – present
  • Goldman Sachs – Donald Raab – present
  • Hazelcast – Greg Luck – present
  • HP – Scott Jameson – present
  • IBM – John Duimovich – present (plus Steve Wolfe by phone)
  • Intel – Michael Berg – present
  • Werner Keil – present
  • London Java Community – Martijn Verburg – present
  • Geir Magnusson – present
  • MicroDoc – Hendrik Hoefer – present
  • Oracle – Don Deutsch, Calinel Pasteanu, Anish Karmarkar – present
  • RedHat – Mark Little – present
  • SAP – not present
  • Software AG – Prasad Yendluri, Alex Snaps – present
  • SouJava – Bruno Souza, Otavio Santana – present
  • TOTVS – Hernan Perrone – present
  • Twitter – Tony Printezis, Ramki Ramakrishna – present
  • V2COM – Leonardo Lima – present

Total attendance: 24 of 25 voting members

Since 75% of the EC's 25 voting members were present, the EC was quorate for this session

Friday October 22

  • ARM – Paul Manfrini – present
  • Azul Systems – Matt Schuetze – present
  • Credit Suisse – Susanne Cech present
  • Eclipse – Mike Milinkovich – present
  • Ericsson – Johan Mårtensson – present
  • Freescale – Maulin Patel – present
  • Fujitsu – Mike DeNicola – present
  • Gemalto M2M – Thomas Lampart – present
  • Goldman Sachs – Donald Raab – present
  • Hazelcast – not present
  • HP – Scott Jameson – present
  • IBM – John Duimovich – present
  • Intel – Michael Berg – present
  • Werner Keil – present
  • London Java Community – Martijn Verburg – present
  • Geir Magnusson – present
  • MicroDoc – Hendrik Hoefer – present
  • Oracle – Calinel Pasteanu, Anish Karmarkar – present
  • RedHat – Mark Little – present
  • SAP – Sanjay Patil – present
  • Software AG – Prasad Yendluri – present
  • SouJava – Bruno Souza, Otavio Santana – present
  • TOTVS – Hernan Perrone – present
  • Twitter – Tony Printezis, Ramki Ramakrishna – present
  • V2COM – Leonardo Lima – present

Total attendance: 24 of 25 voting members

Since 75% of the EC's 25 voting members were present, the EC was quorate for this session

Minutes

Changes in status as a result of attendance at 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):

  • Missing two meetings in a row results in a loss of voting privileges until two consecutive meetings have been attended.
  • Missing five meetings in a row, or missing two-thirds of the meetings in any consecutive 12-month period results in loss of the EC seat.

There was no change in voting status as a result of this meeting.

Patrick reported that Scott Jameson will be retiring from the EC due to HP's split into two companies. He thanked Scott for his many years of service to the EC. Members agreed.

Action Item review

See the JIRA.

The future of java.net

In response to a late suggestion by Werner Keil Patrick asked Heather to address the question of the future of java.net and what this might mean for JCP projects for new JSRs. Heather reported that java.net will indeed be decommissioned. She pointed out that there are two separate "sections" on java.net: the social side and the "Kenai" side (source-code repositories, mailing lists, and issue trackers). She reported that the social material from java.net is being migrated to the new community.oracle.com site. While in general no new Kenai projects will be created we have an exception for JSR projects so we will not be immediately affected. The PMO is working with others inside Oracle to determine what alternate platform(s) should be recommended for future JSR projects though this is only at the "plan for a plan" stage at the moment. Heather promised to report back at a future meeting.

EC Stats

Heather presented the usual EC stats and reminded members that a number of JSR Renewal Ballots will be held starting on November 10.

JSR 364 update

Patrick provided a brief update on JSR 364, reporting that things seem to be moving again, and expressing his hope that final approval would be received soon (see the PMO presentation for details).

JSR 358 update

Patrick then provided an update on JSR 358, where he was less optimistic that progress was being made (see the PMO presentation for details).

Mike DeNicola noted that Oracle executives don't seem to be interested in this JSR and asked whether Oracle will vote to continue it during the upcoming JSR Renewal Ballot. Don Deutsch pointed out that changing legal terms is very difficult for Oracle and for others, and asked whether, if Oracle executives said no to the flat IP proposal, EC members would be willing to continue working on this JSR. Mike responded that we should discuss that when we hear what Oracle's concerns are. Mike Milinkovich pointed out that we are asking the execs to spend time on something that brings no revenue.

Don reviewed the proposed content of JSR 358: flat IP, no patent changes, and open-source. He reported that the lawyers are tentative on the last item. He pointed out that the legal agreements are very complicated and that we are asking the lawyers to agree to do things differently. Being naturally conservative, they are inclined to say "no change". Mike DeNicola responded that killing the flat IP proposal might mean that EC members would decide that it's not worth the effort to proceed.

Elections and JavaOne updates

Heather provided final update on this year's elections and on the PMO's plans for JavaOne. See the PMO presentation for details.

Locations for 2016 face-to-face meetings

We agreed that we will meet in Rio in January (see the next topic).

On the assumption that we will meet in Rio in January several members said that they would prefer Ottawa to Germany for our summer meeting. Patrick agreed to initiate a Doodle poll to gather members' preferences for location and dates. Mike Milinkovich reminded us that when he originally suggested Ottawa he contacted both IBM and Oracle and both agreed that they could host us there. If we decide on Ottawa we will need to follow up with them.

Plans for the Rio meeting in January

We first discussed dates for the January EC meeting in Rio to be hosted by TOTVS. Bruno Souza noted that some EC members might like to attend the Campus Party Brasil conference in Sao Paulo from January 26-31. He therefore suggested, and members agreed, that we should hold the EC meeting on Thursday Jan 21 and Friday January 22. He and Otavio Santana presented a suggested schedule (see their presentation for details) for these days.

We then discussed concerns from EC members (particularly US citizens) about potential difficulties in obtaining visas. We agreed that any members who were unable to attend because of visa problems would not be penalized. David Britto said that TOTVS would provide invitation letters to those who need them.

SouJava update

Bruno Souza and Otavio Santana presented a report on SouJava's involvement the "20 days of Java" tour and other JUG, open-source, and Java events in Brazil during 2015. See their presentation for details.

Oracle and Java

Mark Little gave a presentation that expressed community concerns about Oracle's commitment to Java EE in the light of the news that they had laid-off most of their Java Evangelists. Patrick responded that while it was true that Java Evangelists had been laid off, the only ones who remained are actually working on Java EE. Bruno Souza pointed out that it's understandable that Oracle will from time to time adjust its level of investment. SouJava would like to have an open discussion about how they might help the community to adjust to these changes.

Don Deutsch pointed out that JavaOne will provide EC members with the best information about Oracle's plans for Java. He noted that when Oracle took over Java five years ago it was in very poor shape. Since then Oracle has made a very significant in Java, progressing all three platforms and opening up the JCP. He noted that "now you're saying you're concerned that we will stop doing what you feared earlier that we would not do!" He said that Oracle has grown the community and increased Java User Group participation. He noted that Oracle has consistently expressed concerns that others have not invested in Java to the same extent. He suggested that we revisit this question at the next meeting after JavaOne.

Greg Luck pointed out that the Info World article was based on an unpublished letter, and suggested that the Evangelist issue is a "storm in a teacup". Martijn Verburg agreed.

Mark Little responded that he agreed that Oracle has done more for Java in the last five years than Sun had done in the previous five, and he also agreed that Oracle has the right to adjust their standards-involvement as they see fit. However, he reported that many customers are worried. He said that RedHat is willing to step up and pick up some of the slack.

Anish Karmarkar pointed out that several Java EE JSRs are in the process of publishing Early Draft Reviews, suggesting that work is continuing.

Werner Keil asked how many Java EE JSRs are led by Oracle. Patrick responded that almost all of them are, noting that the Spec Lead role is now more dominated by Oracle than at any time in the past (by Sun or Oracle).

Mike Milinkovich suggested that one reason for the lack of others' involvement might be that standards are no longer cool - open source is. Bruno Souza responded that the companies represented on the EC should think about how they can do more. Mike responded that if companies aren't leading JSRs this indicates that they have no business incentives to do so. He suggested that Oracle invest some time in writing up an explanation of why it's good business to lead a JSR. Geir Magnusson said that Apache first got engaged with Sun around Java EE, and suggested that perhaps Java EE is not so relevant any more.

Mark Little asked what the impact would be on the EC if Java EE is indeed being "End-of-Lifed". What reason would there be for the JCP to continue? The EC has very little influence on OpenJDK (being simply asked to rubber-stamp decisions made there).

John Duimovich noted that IBM has participated significantly in Java EE spec development, and has made a significant investment there.

Mike Milinkovich asked what Oracle is doing to make Java (not Oracle products) relevant in cloud environments. Martijn Verburg noted that many are migrating to node.js which is easier to deploy. Mike agreed that node.js is disrupting Java EE. Martijn noted that JSON binding is arriving very late. Calinel Pasteanu suggested that Java EE development is slower than it might otherwise be because of the lack of participation from others. Mark noted that none of the technologies that are being considered as alternatives to Java EE are standards-driven.

Patrick reported that from his and the PMO's perspective it is "business as usual" - nobody has asked them to stop doing what they have been doing, or to do anything different.

Greg Luck asked whether Oracle might file some JSRs around microservices. Patrick responded that he didn't know, but perhaps there would be some news at JavaOne. Werner Keil wondered whether Java EE 8 might include a profile for microservices. Greg Luck noted that Spring and Spring Boot are extremely popular.

We agreed to revisit this topic in a future meeting after JavaOne.

Twitter presentation

Tony Printezis and Ramki Ramakrishna gave a presentation on the use of Java and the JVM at Twitter. Being new to the EC they introduced themselves, explaining that they both used to work at Sun.

They said that it is difficult to contribute fixes back to OpenJDK unless you knew the right people. Martijn Verburg commented that many people would like to participate more in OpenJDK but are unable to do so by the difficulty of accessing Oracle's infrastructure (such as tests). The LJC has been trying for five years now to open these up. Tony noted that Oracle's IT and security teams are probably also an obstacle. Mike Milinkovich argued that OpenJDK is an open-source project that is constructed in such a way that it discourages outside contributions. Don Raub said that since the JCP is slow and OpenJDK is slow - why not just do it and release it publicly?

Tony responded that they want to contribute back to OpenJDK so that they don't have to maintain a separate fork forever.

We then adjourned for the day.

MicroDoc's IoT strategy

Hendrik Hoefer began Friday's session with a presentation on MicroDoc's Internet of Things strategy (see the presentation for details). Members expressed great interest, some commenting that this provided them with insights into technologies with which they were previously unfamiliar.

Eclipse IoT projects

Mike Milinkovich followed this with a presentation outlining the various IoT projects at Eclipse. Once again EC members responded with interest.

OpenJDK

Following up from an earlier meeting, Gil Tene led a discussion on OpenJDK (see his presentation for details).

Patrick took the opportunity to remind members that JSR 358 adopted OpenJDK's TCK license - the Oracle Community TCK License Agreement (OCTLA) - as a proposed model for all JSRs.

Members expressed concerns that the EC was being asked to "rubber-stamp" work done in OpenJDK. They pointed out that although OpenJDK's Java SE 9 work is almost feature-complete there has still been no umbrella JSR filed. It was pointed out that Java SE 7's Early Draft Review was published in March 2011 and the Final Release was in July 2011; obviously such an extremely short time-scale would make it impossible for JCP deliberations to affect the spec.

It was also suggested that infrastructure issues also made it difficult for outsiders to participate fully (see Martijn Verburg's earlier remarks during the Twitter discussion).

Someone pointed out that there are very few external committers to OpenJDK. (NOTE: see https://www.openhub.net/p/openjdk for hard data.) Oracle does most of the work and consequently exercises significant control.

Gil Tene suggested that contributions ought to imply significant obligations (for example, ongoing maintenance, porting to all platforms, and development of TCK tests).

Picking up the discussion again later Patrick pointed out that the EC has repeatedly argued that implementing in open-source before standardizing is a good development model to follow.

John Duimovich argued that the goal of the standardization process is to enable multiple implementations. This implies that specs are available early in the development cycle so that people can experiment with them. The OpenJDK model effectively means that we are standardizing a particular implementation and in the process may be incorporating into the standard design choices that are inimical to multiple implementations. Mike Milinkovich agreed, arguing that developing in open-source first is appropriate, but that deriving the spec from a single implementation is not. John agreed, pointing out that "development first" ought to imply multiple early implementations - we don't get that with OpenJDK.

Gil Tene pointed out that JDK 9 is supposed to be feature-complete next month but we don't even have a draft of a specification. John responded that spec issues are discussed on separate per-project mailing lists.

Mike Milinkovich noted that third party tool vendors such as Eclipse and independent implementers of the platform cannot even get started until a JSR is filed and until we have a reasonably complete spec. He also pointed out that often the code (as opposed to the spec)  is not even developed collaboratively. Instead, it is developed behind closed doors, and then "just appears".

Martijn Verburg said that the London Java Community plans to vote no on the Java SE 9 JSR because the JCP process is being bypassed. Bruno Souza responded that SouJava would also vote no for the same reason.

After considerable discussion about how to address these concerns Patrick pointed out that the EC is not the appropriate forum. Instead, the matter should be raised with the OpenJDK Governing Board - particularly since two EC members present (John Duimovich and Mike Milinkovich) are currently serving on that body. We agreed to form a small Working Group to discuss these concerns in more detail, and to formulate specific suggestions to be taken to the Governing Board. An AI has been formed to track this.

JSR 354: Money and Currency API

Anatole Tresch presented in interesting retrospective on his role as Spec Lead for JSR 354 (see his presentation for details). Members thanked him for all of his hard work and congratulated him on successfully completing this important JSR.

JSR 375 - Java EE Security API

Alex Kosowski gave a Spec Lead presentation on his JSR. He expressed concerns that as a new Spec Lead he might not be doing everything correctly but EC members were impressed by his enthusiasm and his thoroughness, and reassured him that he was doing well, particularly in operating transparently. They suggested that he invite external "security experts" to review the Spec. Alex agreed to do so.

Adjourn

Patrick thanked Twitter for their hospitality, and Heather for the work she did in organizing the meeting, and we then adjourned.