hogepodge commented on issue #9057:
URL: https://github.com/apache/tvm/issues/9057#issuecomment-931595264


   +0 as someone who is not actively reviewing a large body of code, I don't 
have a strong opinion on if this is the right move. It's clear that the current 
"tag all code owners" is not workable, and I fully support removing that. To me 
the big questions is: how do we fairly allocate work amongst the community and 
code owners. We can look at how other communities have handled this.
   
   O1) Within the Kubernetes community, code ownership is allocated to Special 
Interest Groups, and SIG leaders are round-robin allocated to review code under 
their ownership. Other reviewers can be automatically added also based on 
codeowners files, and reviewers and assignees can be manually added through 
tags to bots.
   O2) The Rust community allows for volunteers to join a "high five" program 
that automatically round robin assigns them for code review. It's opt-in, and 
is a way to gain experience and merit within the community.
   
   Echoing what @areusch was saying, since every pull request needs a review 
from a committer to merge, I feel we need a way to allocate committers, but it 
does no good if committers have become inactive and we're just spamming them. 
Going back to Kubernetes as an example, their three-month release cycle creates 
a natural period of time where contributors can volunteer for work but be able 
to step back gracefully if work or personal demands take them away from the 
project. Looking at O2, I could imagine a quarterly "call for reviewers" to 
give us space to build an equivalent of that "high five" list, and for the PMC 
to review contributions by community members and identify who is active, and 
who is inactive.
   
   Within a community ladder I like to think of what roles and responsibilities 
people have to be regarded as active in the community. For example, an active 
reviewer or committer should be maintaining a certain level of reviews, and 
similarly somone who is not an active reviewer but wants to be promoted should 
be able to demonstrate an ability to do that. This is somewhat at odds with The 
Apache Way of doing things, but in my mind it helps to distinguish between 
active an inactive, as it helps the community in general identify who is there 
to help, and how to share the work amongst them.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to