2023-10-25 - RFC Process Improvements

RFC Process Improvements

• Craig McNally provided an update on the changes to the process description that he is working on, including:
  • clarifying what happens when an RFC is rejected during the preliminary review
    • Marc Johnson asked whether rejected RFCs should be recorded in the mainline? Tod Olson suggested those could be kept in a rejected folder (separate from accepted in the root). We agreed to merge rejected RFCs
  • introducing a preparation stage
• Marc Johnson provided some comments on the changes prior to the meeting (mostly about tense of the advancement critieria). Craig McNally agreed to address those outside the meeting

Draft Refinement Stage

• Jenn Colt asked about what expectations the submitter might have for the TC forming the draft refinement working group e.g. how long it should take to form the group or whether a TC member needs to be the convener of the group?
  • Tod Olson does not think there is a need for the TC to provide leadership during the draft refinement stage
  •  Craig McNally suggested the TC needs to receive regular updates on the progress of the group
  • Tod Olson suggested what might be needed is a liaison, similar to how the PC operates and that doesn't have to be a TC member
  • Marc Johnson raised concerns about adopting a passive model similar to the PC for this process due to the TC being responsible for this process progressing quickly
  • Tod Olson suggested that the TC might not be responsible for the draft refinement stage
  • Marc Johnson asked who is responsible for the draft refinement stage progressing sufficiently quickly, given that delays could either lead to development work being blocked or outpacing the RFC process (which then invalidates the benefits of the RFC process due to an unwillingness to provide feedback after significant investment has been made)
  • Craig McNally provided a clarification that it is valuable for PoC work to occur concurrently with the RFC process, and that being open with that is important, however that work would not be merged to the mainline prior to completion of the RFC process
  • Craig McNally suggested it could be the submitter who drives this stage forward and advocates for moving the RFC to the public review stage
  • Jenn Colt agreed that it was important that someone be empowered to suggest it's time to move on or raise concerns with the TC chairs
  • Tod Olson asked why the submitter (or any other interested party) couldn't be the organiser of the draft refinement working group?
  • Jenn Colt suggested it would be too much work for the submitter
  • Jenn Colt suggested we could try the process out and reflect upon it at the next process review
  • Craig McNally will add some clarifying language around the group being formed in a timely manner and that there needs to be a liaison from the group to the TC

Final Review Stage

• Craig McNally asked if further changes requested by the TC at this stage could invalidate the feedback from the public review?
• Marc Johnson suggested that would depend upon how involved the TC members are in the public review
• Craig McNally and Tod Olson agreed that the public review should be where most of the conversations / feedback happens
• Craig McNally asked about how we defer feedback not intended for the preliminary review / draft refinement to the public review?
• Marc Johnson stated that he would likely try not to participate in those earlier stages because it is too likely that he would provide the inappropriate feedback at those earlier stages and is unsure when to provide feedback in a timely without triggering reluctance due to sunk costs
• Jenn Colt suggested we should design a process that makes it easier to do it right and harder to do it wrong
• Jenn Colt suggested that folks participating in draft refinement should feel comfortable providing any feedback
• Olamide Kolawole suggested that the draft refinement group is the best place for providing dissenting feedback
• Marc Johnson we likely need to assume that folks are going to provide that kind of feedback and that it will be received well by the submitter and the TC
• Craig McNally asked if we could combine the final review into the end of the public review?
• Marc Johnson suggested that we may have kept it separate to have a distinction between completing review and the TC being ready to make a final decision. And that the vote is only valuable if the TC is willing to reject RFCs
• Maccabee Levine stated that he thought the TC could reject RFCs. And suggested that combining the two stages could help the process move faster
• Jenn Colt suggested that the most important aspect is that folks are aware that the opportunity for public feedback is now over
• Marc Johnson suggested that we need to work on the basis that all TC members will not have be capacity to monitor all RFCs throughout the whole process. And that for them to make an informed choice, there needs to be time for members to read the RFC prior to voting on it
• Craig McNally stated that the key challenges are determining the timing of closing feedback and having the final vote

Marc Johnson  to  Everyone 11:20 AM
We are struggling with forming groups, so we may well hit a bottleneck on those at some point

Ingolf Kuss  to  Everyone 11:22 AM
i have to be absent

Marc Johnson  to  Everyone 11:29 AM
My past experience is that things tend to go from POC to production code e.g. the inventory search work with ES
We know from recent experience that the community isn’t equipped to move quickly

Marc Johnson  to  Everyone 11:35 AM
We’ve had feedback from most of the previous RFCs that they take a long time and that is frustrating for the submitter

Tod Olson  to  Everyone 11:35 AM
That sounds reasonable.

Jenn Colt  to  Everyone 11:36 AM
Brb cat knocked over coffee

Marc Johnson  to  Everyone 11:36 AM
In the past, we’ve had almost no feedback in the public review

Marc Johnson  to  Everyone 11:41 AM
My concern is about timeliness of feedback and sunk cost for both the RFC process and the likely concurrent dev work

Marc Johnson  to  Everyone 11:49 AM
That’s ok, carry on without me

Maccabee Levine  to  Everyone 11:51 AM
To repeat what I said at TC on Monday, I definitely value Marc's feedback and poking holds both on process and on architectural details.  And there were lots of 👍🏻when I said that.

Tod Olson  to  Everyone 11:52 AM
It would also seem that preliminary review would also be a time to say "I think this is the wrong direction entirely."

Marc Johnson  to  Everyone 11:53 AM
I can try that, I find even summarising that I have concerns tends to be poorly received It’s happened repeatedly for years, which leaves me to think much of the burden is on how I deliver the feedback

Tod Olson  to  Everyone 11:55 AM
Actually, I need to drop now. Thank you all for the considered discussion.

Marc Johnson  to  Everyone 11:56 AM
I think that comes back to the political support aspect
The translations proposal was very different in that regard to other proposals

Maccabee Levine  to  Everyone 11:57 AM
Apologies I have to drop off at 12.