Difference between revisions of "Planning Council/November 02 2016"

Latest revision as of 17:03, 2 November 2016

Logistics

Members and Attendees

Note: "Inactive" refers to Strategic Members or PMCs we have not heard from for a while and have been unable to convince to participate. Those members can become active again at any time. Contact David Williams if questions.

Note: feel free to correct any errors/omissions in above attendance record.
Y = Yes, attended
N = No, did not
R = regrets sent ahead of time
D = delegated
X = not expected

Announcements

•  ?

Previous meeting minutes

• Review previous meeting minutes if you'd like. That is, review them before the meeting, but if questions or issues with previous minutes, this would be a good time to bring them up.

Neon maintenance

• Are we done with these Neon.1 issues?

• Was "Neon.1 help" ever deployed? See bug 500938.

- ACTION_ITEM: David to discuss with Fred to determine state and make sure "what we know" is deployed.

- ACTION is "done": updated bug, sent email. We'll see.

- Then focus needs to be on bug 499411 to make things easier for all future releases.

• [Wayne?] Create New and Noteworthy for Neon.1 (bug 500939)

- I assume we can "declare victory" but are "we" (Wayne?) prepared to do this for all future update releases (and primary releases)? Do we "need a documented process"?

- ACTION_ITEM: Wayne to "close out" the current bug and document there (or, link to a wiki page) on what the process is moving forward. I bet it will involve PMI. :)

• Any issues for Neon.2? Any new projects joining? I assume some will have new features? (But I do not know what they are. Does anyone?)

- no issues known for Neon.2

- ACTION_ITEM: Wayne to tweak PMI so that projects can document their new features for update releases. The "documentation" is primarily to improve communication to the community on "why they should care" about a particular release but there could be other consumers such as translators or tutorial writers may need react. (This is required, since "update releases" are no longer "maintenance releases".)

Oxygen Planning

Java 9 Coordination [Dani]

• After discussing what to do about Java 9 date changing, at a previous meeting, we also discussed the need for "Java 9 coordination". Dani volunteered to be the "Java 9 Coordinator". Dani, is the following a fair summary of that work item:

1) Find out which other projects plan to participate in a likely "July Java 9 update" (that is who is plans to "support Java 9 during development time"?)

2) Educate projects on how and WHEN to test their code 'running' on Java 9. Ideally, as projects test on Java 9 there will be a synergy where projects will educate each other on what was discovered and what others should look for (perhaps on a wiki, somewhere).

- Dani's statement during meeting was consistent with the above, but he may be more specific once he "begins the work" (anticipated after Neon.2).

- I have updated our Oxygen Plan document to now include a "July Update Release".

Stop using Release Name?

• There has been a lot of discussion about "giving up release name" and using "date" instead. See bug 493490.

- Further discussion in a previous meeting lead to the idea that one thing that is missing is what DO we call the thing we are releasing. "Eclipse Neon" seems too vague and definitely sounds like a different thing than "Eclipse Mars" (even if you add "release". Some quick suggestions were "Eclipse IDE - Neon Version" or similar (with dates, probably).

- What's wrong with "Package"? [dw: suggestion during 11/2 meeting was"IDE" as a suffix of package name, or "IDE's" when needed generically, would be a good choice, which I think all agreed with.]

• - Main ACTION ITEM is that we owe the community some official response on bug 493490 about what our official PC plan or response is. Is there agreement the following is our official position? If so, I will comment in bug 493490.

Much discussion again! One of the best points I heard (from Alexander) was that "keeping the code name" allows projects to easily associate their version of with the yearly Simultaneous Release (or update release). But, as agreed, the name by itself does not have any meaning. While not strong agreement from everyone, it seemed to me the tendency was to favor adding "date" to code name to those places where end-users might be reading it and be confused (or, if not confused simply "get nothing from it") -- such as splash screen, about box, and the www.eclipse.org/downloads related pages. I asked for members to open bugs for specific cases, but in the meantime have tweaked our PC statement below:

Summary of our position: decided that we can not "solve" this issue for Oxygen, but we can make incremental improvement. After that incremental improvement, we may have a better idea of the (or another) core problem to fix. As things are now, there is little chance of getting agreement on what the problem is or how to fix it. Currently, it is sort of like saying that "the problem is that things are confusing" -- too broad to fix all at once. We do not want to abandon the "code name" (at least for now) based on what we know. But there is agreement that *by itself* the codename has no meaning for casual users and is confusing because it is used in contexts that make it appear to have meaning.

Therefore, it is recommended where specific areas are confusing or meaningless that bugs be opened with a suggestion on how to improve by adding more information (typically the "releases date"). This includes web pages, splash screens, tutorials, instructions, announcements, press releases, urls, and anything else where "version" and "code name" and "contents of deliverable" are confusing or meaningless. While typically the date of the release should be added to the code name (such as "Neon.2 [December, 2016]") in some contexts it might be more appropriate to use the project version or build id. Also, improvements to "check for updates" should be made that give some indication that a "whole new stream" is available -- as the Installer currently does (and as other "products" do, such as Ubuntu, though there it is a user preference). But this will require someone to implement a solution in p2 UI. Also, instead of calling the delivered pre-packaged collections "packages" they should probably be called "IDEs".

New levels of IP

- Do we, as Planning Council, want to stipulate a participating project must be of "type B"? Or, do we not care? It may depend on "how labeled"? How do users or companies know? What do they expect?

- Wayne has opened bug bug 501014 for comments.

• In a previous meeting we decided our initial position should be "status quo" -- that is, to say it is required that each project in the Sim. release repository (and EPP packages) be of Type B. A few reasons given:

- Adopters or users may be surprised to get a different legal process than what they are used to. This not only concerns adopters which build standalone commercial product, but some build features which prereq certain EPP packages are installed, so they direct their customers to install those packages first, and then install their features. Plus, while is it commonly believed "users don't care", many of those users work for corporations or government agencies which do care. For example, the corporation may have a policy that "it is ok to download and install things from Eclipse where all the code has been "reviewed as Type B". But those corporations might also tell their developers: "it is not ok to download open source in general and not even from Eclipse for Type A". It is unlikely to get corporations to state what their internal policies are since they would consider that confidential, so it just seems safest to go with "status quo" -- until we learn otherwise or have good reason to change.

- Some projects that want to use this new Type A policy, such as Orion, have always been "part of the Simultaneous Release", but not part of the Sim. Release repository. It is only the repository that we are saying it is required for, not the Release, per se.

- "How labeled" may make a difference. For example, as far as we know, it is still required that any EPP package that contains "incubating projects" label the whole EPP package as "incubating". Could there be something similar for "Type A" or "Type B" projects? And, the methods must be meaningfully named to be usable (by users); not just "Type A" and "Type B".

• New for Nov 2 meeting: Nick Boldt has some new information and questions and a proposal that he posted to PC mailing list message. Let's discuss!

• [Notes from Nick Boldt with slight editing from dw]

- release records will be updated to include which level of due diligence (Type A or B).

- Type A / Type B applies to third party libraries ONLY, not to the project's code itself.

- Wayne [EMO] explicitly do NOT want Type A to be considered bad or stigmatized. [dw: He made the point that "Type A is more than other foundations or open source groups do".]

- if a hypothetical company wants to invest resources to move a project from Type A back to B, they could do so [dw: I do not think this is true, it is essentially Eclipse Foundation's resources that are "spent" on Type B checks of third party software]; they could also do the additional provenance checking and never report back to the Foundation or the project.

- To be resolved: where should license check style be placed in release records? As a new section on release record or a new blurb in the release doc itself?

- To be resolved: should projects use plugins' about.html, features' license.html, or a LICENSE file to identify the license check style?

• [Notes from David Williams -- overlap with Nick's, but since we were entering in parallel decide to enter both sets]

Given the new information provided by Nick, and some by Wayne, it was decided "anyone can be in the Simultaneious Release repository regardless of IP method choosen". But, this was conditional on users and adopters being able to easily know which method applies -- in case it does matter to them. Suggestions were made to provide meaningful names (other than "Type A and Type B") and to provide the information in something like the about.html file. We all agreed with Wayne that it should not be part of a package name or bundle id, etc. Just something more "self documenting" than the "release record" (Wayne's currently planning on providing that).

Wayne did ask for "specifics" on "who needs to know?". For examples, "which adopters would do their own extra IP review for Type A bundles or who would "not be allowed to use it". I pointed out it is very unlikely that a bank or major coroporation would share that openly since they might consider such "internal product decisions" confidential and/or a competetive edge. I do think Stephan Merker implied that SAP needs this in a previous meeting (but they were not represented a today's meeting on 11/2). So, Stephan, feel free to bring the topic up in December's meeting or comment in bug 501014 if you can.

Potential new requirement

• Should the ability to update from yearly release to yearly release be a 'requirement'?

ACTION ITEM from 6/8 meeting. Doug volunteered to "take up" this item to better specify "what does it mean" and "what will it take" to update across major releases. After last meeting, Doug, mentioned he thinks this is a UI issue and he would take it up with the UI task force. I do not think it is only a UI issue, but from our point of view, more a matter of what we expect projects to do differently. As I have listed before:

What would this take? Such as,

features can never be removed but can be replaced with some form of transition. See bug 314165 and bug 371302.

Preferences, views, etc. have to "migrate" (if their ID changes)?

What testing would projects have to do?

What is the effect on commercial products? That is, will their customers get sufficient information that they "... can not upgrade without voiding their warranty", so to speak?

Have we ever had a case where year-to-year updates worked? (For everything.)

For Neon, it was the change in package layouts. (Hence we backed-off having a "streamless URL".)

For Luna? it was the change in MacOS layouts

For Mars? it was the Window executable could not be updated. (Now it can be, as long as it is named exactly 'eclipse.exe').

Eventually, I assume we would want a built-in stream-less URL. I am assuming for Oxygen we do want to have a stream-less URL available, but not built in, to enable testing the update from Neon to Oxygen. (See bug 483786)

-After a brief discusion, it was decided this should not be a requirement though we should encourage projects to test that scenerio and keep track of issues.

-Doug surprised us all saying current "root features" do not work as expected. That items can to be removed or added? Doug, please clarify in bug 505378 what issues you are seeing.

Release Policy vs. Release mechanics

• For details, see bug 483322.

In Neon M6 we changed to have (nearly) all features to be "root features.

Now what? That is, can we "stop" adding "reference repositories" via feature p2.inf files?

Can we make an official policy on "off scheduled fixes" (as we are considering for MPC)?

- Did not discuss at 11/2 meeting

New Business

•  ?

Next Meeting

• December 7, 2016 - Regular First Wednesday Meeting

Reference

Draft Eclipse Project Branding Requirements (Wayne)

Neon Wiki page

Planning Council Members

Simultaneous Release Roles and Simultaneous Release Roles/EMO