Re: [ee4j-pmc] PMC Minutes

[Wayne Beaton <wayne.beaton@xxxxxxxxxxxxxxxxxxxxxx>]

The fact is that there are rules. All projects must follow the Eclipse Development Process and conform to the Eclipse IP Policy (and engage in the IP Due Diligence Process).

It is the responsibility of the PMC to establish requirements, coordinate plans, etc. The PMC absolutely can set themes and priorities for releases, determine how projects fit together, etc. (the PMC should set themes and priorities in consultation with the project teams and build concensus, but ultimately has the power to impose). The PMC absolutely plays a role in making sure that all of the projects play well together. As others have suggested, for this to work, anarchy isn't an option. 

The Specification Committee will decide how specifications get developed. Community input is, of course, welcome, but there are factors beyond simple democracy at work. There are, for example, legal responsibilities that need to be taken into consideration, and--ultimately--these specification aren't going to be of much use if the companies that build implementations based on the specs can't or won't implement them. The specification process needs to consider these (and other) factors.

It may look like open source projects (in general, not just at the Eclipse Foundation) are run by the committers in a bottom-up sort of way. And this is, to a very large extent true. However, there are many influences and controls. The spec committee will decide what specifications get accepted as official specifications; the projects providing the specifications, APIs, reference implementations, and TCKs need to work together; the PMC must approve releases, new committers, and new project leads; the companies that implement the specifications need to actually want to implement the specifications; and consumers need to actually be able to deploy applications using the technology. 

Of course, all of the stakeholders are totally interrelated (there is a lot of overlap amoung project teams, the PMC, the spec committee, etc.). There's no us-vs-them, here. We're all just us.

Requiring that all project leads be elected isn't a priority, AFAIK.

Wayne

On Thu, May 17, 2018 at 11:57 AM, Markus KARG <markus@xxxxxxxxxxxxxxx> wrote:

David,

 

I need to disagree with virtually all you told me so far. Starting with "Crocodile tastes like Chicken." ;-)

 

Joking aside, I simply have a different vision how to set up EE4J's future, and I do not see which legal or scale issues stop us from the following:

 

- There is no such thing like "the future direction of EE4J" as all EE4J subprojects are independent entities.

- Whether or not EE4J subprojects like to participate in a common policy is up to their committers.

- If the EE4J subproject committers like to define a common policy, it is up to them to set up a common discussion platform with the help of the EF and the PMC.

- Send an email to all JCP EGs (the addresses are public) and ask the EG members to contribute to EE4J projects. No need to ask the JCP for addresses, so no break of GDPR.

- Send an email to all JCP EGs (the addresses are public) and ask the EG members to become Jakarta WG members. No need to ask the JCP for addresses, so no break of GDPR.

- Send an email to all EE4J project mailing lists and ask the committers whatever you wanted to ask the current project leads.

- Send an email to all EE4J project mailing lists and ask the committers to vote for new committers immediately to get frequent contributors upgraded ASAP.

- Send an email to all EE4J project mailing lists and ask the committers to vote for new project leads, replacing the initial (non-elected) ones.

- Prune all EE4J projects from inactive committers and non-elected project leads.

- Do not come up with anything before asked by a committer. I do not see a real need to discuss a lot of issues the PMC actually discusses. Be more relaxed. EE4J will be a success once more committers get active, they will come up with ideas, and the PMC can then simply implement those ideas instead of upfront trying to set a direction that might be totally wrong. The committers will one day have an idea how to replace the JCP EGs and until then simply the committers do their work: What they come up with IS the specification.

 

That way you will get a much more democratic basis in EE4J. The priority of getting EE4J set up democratic is of higher priority than simply having some progress in an arbitrary direction.

 

-Markus

 

 

From: ee4j-pmc-bounces@xxxxxxxxxxx [mailto:ee4j-pmc-bounces@eclipse.org] On Behalf Of David Blevins
Sent: Mittwoch, 16. Mai 2018 00:31
To: EE4J PMC Discussions
Subject: Re: [ee4j-pmc] PMC Minutes

 

On May 15, 2018, at 2:02 PM, Markus KARG <markus@xxxxxxxxxxxxxxx> wrote:

 

I need to say that asking only the project leads will not really bear a bottom-up result

 

The intent is bottom up, truly.  Or as best as we can get with our non-existent specification process, incomplete projects, no clear definition of what replaces an Expert Group or how it should work, and still no legal rights to the javax namespace and all the complications that brings.  That "best" would be pretty low quality and non-permanent at this point in time, but there is a strong desire to find some way to get people engaged despite the very obvious obstacles.

 

The intent is to take a baby steps forward because if we don't the result will be that everything will be waiting on everything.  I.e. the goal is anything but a deadlock, which is what we have.  It isn't clear how to untangle all of this so any correspondence should be read with the perspective all steps are temporary and we're pulling at strings in a knot.

 

The general desire:

 

 - Start getting some technical direction

 - Have that come from the projects

 - Specification Committee would just dictate process (JCP-like) not technical direction

 

Some of the obstacles:

 

 - We can't legally take contributions yet, because we don't legally own javax.* in any form nor do we have patent rights to any of the specifications

 - We already know the standard Eclipse Contributor Agreement doesn't provide enough rights to support a standards process, so any ideas we discuss now in the projects could be legally questionable in the future.

 - We've only had two Specification Committee meetings (2 hours total) and are gaining consensus that group shouldn't be setting technical direction and it should come from whatever replaces Expert Groups.  More time needed for that to get more firm and it's clear as you say the groups we do have are barely enough to seed the future "Expert Group" regardless.

 - There've been over 250 participants in the various EE-related Expert Groups in the last 18 years, which presents primarily a scale issue and more recently a legal issue.  The scale issue is Eclipse has never onboarded so many at once and is already overwhelmed.  The second is the JCP/Oracle has no legal rights to give the contact information for 250 people to Eclipse given the new GDPR laws.

 

The results make things hard.  This email is as imperfect as the situation, so full of holes and I haven't listed everyone's thoughts on how to overcome the obstacles.  There are many, but instead I'll just encourage people here to write some.  It definitely won't hurt.

 

On the technical direction, ideally that's something that can be done in the coming weeks (not days) and to the best we can reflect people's thoughts.  Ideally we can be comfortable with the unclear rules/lines and use this as an exercise to start inching towards some clear rules/lines.  I think at minimum something like JSR description is what we need.

 

If someone wants to grab that and take a stab at filtering out the parts that wouldn't apply or filing in anything new we might want to ask of future specifications, that would be very helpful.

 

Goal is to have a basic template and hand it to all the spec projects and say "fill this in to the best of your ability."  I should probably start a new thread, but ideally we'd have the template ideas by next Tuesday.  Not spend too long on what it should be, then give a couple weeks to the projects to fill it in.  That won't be perfect, but I suspect it would be a considerable help in the efforts to move the specification process definition forward.

 

-David

 

 

 

 

 

 

 

 

 

 

 

_______________________________________________
ee4j-pmc mailing list
ee4j-pmc@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/ee4j-pmc

--

Wayne Beaton

Director of Open Source Projects

The Eclipse Foundation