openstack team mailing list archive

[Vishvananda Ishaya <vishvananda@xxxxxxxxx>]

On May 8, 2012, at 3:19 PM, Brian Lamar wrote:

> I'll be the second to admit I haven't been doing a ton of reviews lately, but that doesn't mean I'm not dedicated to making the project the best it can be. That is not to say your email implies any judgement but I'd love a couple clarifications.
> 
> What are the minimum requirements, in your mind (since you're proposing members for removal), for keeping core membership? (Your answer could be just pointing me to the wiki page where guidelines for minimum flair have been described.)

There isn't an exact set of requirements, but I feel that people with rights to approve things into trunk should be people that consistently do reviews.  I personally average > 2 hours a day doing reviews. I expect to take a larger load than most because of my role as PTL, but I think it is reasonable to expect people to average 1 review per work day each month or ~20 reviews.  The numbers in the previous mail were since Feb. 22nd, which means ~70 days.

> Is the re-joining of nova-core the same process as joining in the first place?

That was my assumption, but we could come up with something different.

> 
> What are the benefits of having a smaller team in your mind? Many people on this list have dedicated countless hours to the project and have contributed thousands of lines of code. One huge benefit I see of being nova-core is the ability to -2 code which  you believe is going to be detrimental to the project as a whole, perhaps because you have worked with or have contributed that code in the past.

The -2 issue is a good point. I personally treat a -1 (or +1) from the author of a given piece of code quite strongly when I do reviews, but you're right that the -1 could be more trivially overridden. The removal is primarily to keep core a manageable size.  We currently have 25 core members and still have many patches that are not being quickly reviewed.  Giving too many people the ability to approve patches leads to inconsistency in code and the review process.  It seems like overkill to have > 20 people. I expect this number to decrease further if out plans to create substystem branches materialize.

Vish