version 0.1: September 14, 2010
July 27, 2010
Total attendance: 8
Total attendance: 15
|Since 75% of the ME EC was not present, that EC was inquorate for
Since 75% of the SE/EE EC was present, that EC was quorate for this meeting
Patrick expressed his disappointment at the low turnout on the ME EC, particularly considering that Java ME is the principal topic for today's meeting.
Patrick noted that of the three topics that most interest EC members (plans
for Java SE, plans for Java ME, and plans for JCP improvements) only two are
scheduled for discussion during this meeting. He said that Java SE would be
on the agenda for the October face-to-face meeting in Bonn, and noted that
there would not have been time to cover all three topics in a single teleconference
even if we had the right people present.
Scott Jameson asked whether this meant that the the Java SE7 JSR would not be submitted before October. Don Deutsch replied that this was likely. He promised that Oracle would socialize the JSR before filing it, and that EC members would be given a chance to provide input. Josh Bloch suggested that it would be better to file before JavaOne. Scott Jameson said that EC members should not be blind-sided. Doug Lea requested that if any announcement are planned for JavaOne a special EC teleconference be held before that. Don agreed that this was a good idea and promised to pass it on. [A pre-JavaOne briefing teleconference was in fact held later.]
Patrick welcomed the two new members: Aguinaldo Boquimpani from TQTVD on the ME EC and Roger Weber from Credit Suisse on the SE/EE EC and asked them to introduce themselves.
Due to lack of time the EC stats presentation was not discussed, merely inserted into the record. Patrick did note, however, that even though the statistics covered a period of almost 12 weeks there has been very little JSR activity to report during this period, indicating a significant slowdown in activity within the organization.
Craig Gering presented some proposals for the evolution of the Java ME platform. Extensive discussion took place during and after the presentation:
Josh Bloch pointed out that Reflection APIs are necessary for a full implementation of Annotations and argued for their inclusion in CLDC. Craig responded that they are trying to keep the footprint relatively small (no more than twice what it is now.) Josh pointed out that since it's been 4 years since the last major revision of the platform Moore's Law suggests that the footprint could be increased by more than that. John Rizzo argued that twice the current size is still too much for low-end devices. Someone pointed out that Moore's law says that for a fixed cost processing power increases, but in many markets we want to drive down the cost, so available processing power does not increase. Craig agreed that this is a difficult problem and noted that his team has had extensive discussion about it. He noted that they are open to input on how large the platform should be and what could be left out. John Rizzo suggested splitting CLDC into "legacy" and "new/evolving" versions.
Doug Lea noted that concurrency was not included in CLDC and asked whether
ME was expected to run on multi-core devices or whether we should be content
to leave this area to Android. Roger Riggs and Craig suggested that concurrency
could be reserved for CDC.
Doug asked when the distinction between ME and SE could be expected to disappear. Craig responded that this wouldn't be feasible until there's a UI model and an application model that spans all platforms. Someone responded that this may never happen since there may always be UI differences between platforms. Craig responded that in this case it might be necessary to redefine what we mean by Java ME. Roger Riggs pointed out that in many markets it's necessary just to drive cost down, and scalability is not an issue.
John Rizzo noted that some of the Phase 2 proposals are already in MIDP3 and asked what belongs in the "configuration" and what in the "profile." Craig responded that this required careful consideration, but noted that some of the new features will be needed in the TV world and hence should be in the platform. Aguinaldo Boquimpani asked whether Craig saw any need to update the JavaTV JSR. Craig responded that more discussion was needed, and that tt's not clear whether the the JavaTV JSR would need to be updated.
Doug asked whether people were being encouraged to join a "pre-Expert Group" to discuss and plan next steps. Craig responded that they intend to submit a formal JSR, but agreed that they want to socialize the proposals first. Doug noted that broad participation would be required for success, yet the political situation within the JCP was making this difficult. He noted that since some of the Phase 2 proposals are similar to what Android and RIM have already implemented it would be necessary to collaborate closely with Google, RIM, and others. John Rizzo stated that Aplix was willing to work closely with Oracle on these plans. Sean Sheedy asked whether Oracle had reached out to those ME members who were not present at this meeting. Craig responded that they intended to do so.
Sean asked whether FX-mobile was competing for attention or resources with Java ME at Oracle. Craig responded that there are two separate resource pools and no significant competition for resources. He reiterated that Oracle is very serious about making progress with Java ME.
Heather VanCura presented the JCP's
plans for JavaOne, and also provided
some general background
information on the event.
Josh Bloch commented that the exhibition space would be much smaller than at previous JavaOne events and wondered how this would affect the feel of the conference. Members expressed concerns that JavaOne might be overwhelmed by the Open World event. The PMO promised to pass this feedback to the event's organizers.
Patrick noted that the projected attendance for JavaOne compares well with last year, and that the projections for the Brazil and China events were quite high. He agreed that there will be a different feel this year, but suggested that we wait and see how it goes.
Patrick lead a discussion on the proposals for a process improvement (jcp.next) JSR. Since this had been previously discussed, and since there was little time remaining in this meeting, he proposed revisiting the topics in reverse order, to ensure that all were adequatly covered. He suggested trying to identify the topics for which members had concerns, and those which could be considered non-controversial.
Standardized Licenses (Slide 9)
Josh Bloch noted that licensing is a difficult matter for the Spec Lead and for those who implement technologies. He noted that most standards organizations have simple, uniform licenses whereas in the JCP there are multiple different licenses, which makes life difficult for implementors. He expressed Google's preference for using the Apache license for all deliverables.
Tim Peierls noted that obviously there are concerns about this topic.
Wayne Carr suggested that it would be helpful to clarify for each suggested change whether it would require a JSPA change, a Process Document change, or whether it could simply be implemented as a policy change without requiring a modification to either of those documents. He suggested that a simple recommendation to use a standard license would fall into the latter category. However, after some discussion members agreed that since the JSPA specifies that Spec Leads have the right to choose their own licenses, a mandate to use standard licenses would require a JSPA change.
Patrick agreed to classify the proposals in this way.
Collaboration (Slide 8)
Aguinaldo offered a practical example of collaboration: Study Group 9 of the ITU-T wants to normatively reference some JCP specs and in order to do so they will need to establish a formal relationship with the JCP and to classify it as a formal standards organization. Patrick responded that he had been approached by them, and noted that Oracle's lawyers would need to help with this. Scott Jameson and Don Deutsch pointed out that this is a complex matter with IPR implications. Patrick agreed to consult with them on this matter.
Addressing the bullet Confidentiality requirements are a barrier and must be eliminated, Patrick clarified that this refers to confidentiality language in the JSPA rather than to confidentiality in EC discussions. Wayne pointed out that the ECs have already agreed not to support JSRs that invoke confidentiality but members agreed that the language in the JSPA should be clarified to prohibit confidentiality in Expert Group activities.
Compatibility (Slide 7)
Josh Bloch and Jason Gartner both pointed out that no progress could be made in this area unless all implementors (including Apache) have free access to the TCKs. Additionally, Josh argued that requiring implementors to pay to prove compatibility is unreasonable.
Wayne noted that the term Secrecy rules on this slide is misleading since there's nothing in the JSPA or the Process Document that prohibits the free distribution of TCKs or the publication of test results.
John Rizzo said that Aplix believed they were not allowed to publish test results or a list of those who have passed the TCKs. Patrick suggested that this prohibition may have been specified in the TCK license they were using.
At this point the meeting ran out of time. Patrick suggested that members send additional feedback on the proposals by email. Josh suggested that it would be helpful if Patrick provided precise language for the proposed changes.
During a brief discussion about the upcoming face-to-face meeting Scott
Jameson suggested that if substantive discussions on SE7, ME.next and JCP.next
were to be held during the October meeting it should be extended to two or
even three days. [It was later decided to meet for
two full days rather than the usual one and a half days.]