doc: add definition and process of WG

[FeynmanZhou]

Fix #33 due to the WG proposal received a super-majority approval from the existing ORAS organization owners.

[sajayantony]

Testing if tagging @oras-project/oras-org-maintainers works.
/cc @lachie83 @toddysm

Thanks @FeynmanZhou for working on this. I'm hoping we can close on this soon.
I couldn't interpret if there is a clause that enables the WG to recommend that as per its investigation, the recommendattion would be pursue working with the larger group outside of ORAS.

For e.g. if the search working group decides to pursue working on CNCF tag post initial investigation/prototying etc.

Reading these clauses means that the WG limited to

• an update to an existing specification (for example, adding support for a new set of features);
• a new specification (for example, creating an specification for a new topical area); or
• a new ORAS sub-project other than a specification (for example, creating a reference implementation of an specification).

[SteveLasker]

I'm not yet convinced ORAS needs the additional governance and complexity of a working group definition, as opposed to setting the scope of a sub-project. Or, simply clarifying the existing sub-project definition.

However, I do believe we need code (line 83) to sketch, prototype and design an idea.

[SteveLasker]

I appreciate some super large projects like the kubernetes project may have working groups, but those are significantly larger. Other projects like Helm have Helm Improvement Proposals (HIPS) for ideas that enable iteration.

Some more info related to CNCF Working groups, and how they are a component of a sub-project, but not yet another thing.

• CNCF Governance By Subprojects Template

  • For example, you can pretty easily decide that each “working group” in your project is a subproject

  • Subproject: The individual group/repository where work on the project gets done; alternately called projects, SIGs, repos, drivers, plugins, operators, working groups, or other units of work that each have their own maintainers.

However, CNCF does have working groups, that aren't tied to a particular project:

• Contribution Guides to the CNCF Ecosystem - Working Groups

The current search effort is focusing on the CNCF TAG to iterate. As a result, I'm not seeing good justification to complicate the ORAS governance with a working-group definition, above what a sub-project is already doing.

[FeynmanZhou]

@SteveLasker @sajayantony @toddysm Thanks for providing your suggestions. I see the biggest confusion seems to be lacking clarity about the difference between Sub-project and the Working Group. I would like to elaborate on the major difference and justifications for both after I investigated a few more CNCF projects:

Differences between Sub-project and the Working Group

Working groups are organizations responsible for the design and implementation of large architectural aspects of the overall ORAS project. Working groups operate with a fair amount of autonomy within the broader scope of the project. They tend to be long-lived or temporary.

The sub-project is more focused on implementation and solving specific requirements. It could be the outcome of the Working Group. A Working Group's scope may be across multiple sub-projects.

Working Groups are running in multiple CNCF projects.

Apart from the latest Kubernetes community in CNCF, WGs are also popular and are actively running in other CNCF projects/communities either.

• Knative
• Istio
• KubeEdge

[SteveLasker]

Thanks @FeynmanZhou,
There's a lot of great detail for working groups. Much, if not most, or possibly all would seem to apply to sub-projects.
Before we complete the delta definition of a working group, can more clarity be created around a sub-project?
That might help understand what/why a new governance would be required.

[sajayantony]

Approving this as a baseline and recommend we add/modify as needed.

[TerryHowe]

LGTM

[shizhMSFT]

LGTM

[FeynmanZhou]

@sabre1041 Looking forward to your opinion on this WG process proposal.

[sabre1041]

This looks really good. However, one area that I would request additional clarification/discussion is surrounding the length of time for a working group.

In the provided text, it states that a WG can either be long lived or temporary. Working groups should be short lived in order to provide a concrete scope of work that should be accomplished. I feel like the text should read short lived rather than long lived.

The rest of the content looks great.

[FeynmanZhou]

@sabre1041 It makes sense to me. I revised the WG description at line 60 and clarified WG tends to be short-lived. Could you pls help to review it again?

Sorry for the late response. I just back from a holiday yesterday. We can continue to work on ORAS things.

[sabre1041]

LGTM

[sajayantony]

Added to agenda for tomorrow. /cc @SteveLasker

[SteveLasker]

I’m on vacation, and defer to the majority of the maintainers for next steps

defer