Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [microprofile-wg] [External] : Proposal to standardise the process for creating new proposals

Thanks for composing a proposal, Emily. I have a few comments.


Sandbox approach works well in case some work has been done already and some version of APIs already defined. It can be a starting point for a new spec, but it may be not. APIs are defined as part of the work process based on the spec goals and not goals. Jumping over the goals step usually results in faster spec development, but it may have a drawback of quality. I think that the sandbox step must not be required. It’s fine if proposers can show something, but it’s also fine if they just present strong goals and justification why the spec is needed without providing any API.


I am also strongly recommending adding spec goals and justification as requirements for the plan review. It will be easier to understand why this spec is needed.





From: microprofile-wg <microprofile-wg-bounces@xxxxxxxxxxx> On Behalf Of Emily Jiang via microprofile-wg
Sent: Wednesday, June 30, 2021 12:31 PM
To: Microprofile WG discussions <microprofile-wg@xxxxxxxxxxx>
Cc: Emily Jiang <emijiang6@xxxxxxxxxxxxxx>
Subject: [External] : [microprofile-wg] Proposal to standardise the process for creating new proposals


Further to yesterday's MP technical call, there was a consensus to standardize the way of creating new specifications. In this year's MicroProfile Technical Goals, the community expressed the desire to work on new proposals of adopting new technologies such as OpenTelemetry,  gRPC, etc. How to proceed with this?


The current MP process is documented in this wiki. There was some feedback regarding:

If you (proposer) are already active in the MicroProfile community, lazy consensus applies.

If you (proposer) are not yet active in the MicroProfile community, consensus is required. Those not yet active are highly encouraged to leverage the sandbox as a great way to show the essence of ideas.

It seems like adding an extra barrier to new proposers. Besides, what happens if no consensus is reached. This seems one -1 will stop a proposal going forward, which contradicts with our voting policy of supermajority.

With these concerns, we discussed in the meeting with some initial suggestions:

Step 1: Start immediately via `microprofile-sandbox` repository

  • Fork the repository
  • Create a distinct sub-directory
  • Code
  • Submit as many PRs as you need to explore the ideas behind the proposal

No approvals or prior notification required.

The `microprofile-sandbox` has an intentional zero bar to entry to capture ideas when time permits, from anyone, even if not yet active in the MicroProfile community.

When ready, proceed to the next step.

Any form of contribution on any topic is allowed in the `microprofile-sandbox`. Moving from the `microprofile-sandbox` is at will or upon request of the MicroProfile community.

Step 2: Create Project Creation Plan

Follow what Jakarta project creation plan template (see Jakarta config plan creation)

  • Background
  • Scope
  • Why here
  • License
  • People

Create an issue under microprofile-wg with the label of Plan Creation Review and then send it to this WG to start a ballot. If it passes the ballot with a super majority, a new project will be created under microprofile. 


Please share your thoughts. With some discussion, we can then move towards a ballot to make it final. 




Back to the top