Fill in missing gap in Beman Library Maturity Model - require some approval when moving to production ready states

Beman Website and Docs
1

Fill in missing gap in Beman Library Maturity Model - require some approval when moving to production ready states
Current issue: Fill in missing gap in Beman Library Maturity Model - require community and leads approval to go production ready states · Issue #137 · bemanproject/beman · GitHub

Context: We have [beman/docs/BEMAN_LIBRARY_MATURITY_MODEL.md at main · bemanproject/beman · GitHub](https://The Beman Library Maturity Model), but we don’t specify who decided a transition to right.

@Jeff-Garland pointed out that maybe we should not let the authors decide that.
Previous thread: add badges for c++26 by JeffGarland · Pull Request #5 · bemanproject/indices_view · GitHub and Add guidelines for code review by neatudarius · Pull Request #134 · bemanproject/beman · GitHub.

I do think we should require a formal review from community.
Jeff proposed a 4 step process:

  1. submit a PR to update the status
  2. post on discourse the start of a 1 week library review
  3. community to download, test, and review entire library
  4. Such PR requires approval of 2 leads and other community members

Please comment here. I also pushed this subject on our agenda.

1 Like
2

If this is purely an administrative rubber stamp where a merely cursory review of the library is made, I’m opposed. More bureaucracy without motivated purpose or specific desired outcomes isn’t useful.

What situation are we trying to manifest with such an approval process? What situations are we trying to avoid? What are the other options for avoiding them?

On the other hand if this is boost review, well let’s call a spade a spade. I’m not for or against really, and such a review certainly has a clear and well-defined purpose, but you shouldn’t try to smuggle such a thing by. If you’re suggesting moving a part of boost’s process into Beman it should have big blinking lights on it, tassels and bells.

3

The clear purpose is to ensure a library isn’t just declared as production ready without a single review. If you follow the links in the post you’ll see that a brand new to the organization author just went ahead and did that. Are you willing to bet your production on that standard?

Ensuring that there has actually been a review of libraries that are going to declare the highest rediness state. Since it’s now already happened with only cursory review from myself I don’t feel comfortable – but I don’t have a process to say to the author – like you need to do this first before putting on that badge.

That’s just one idea used as an analogy where we get community eyes on the library. Boost review is a completely broken concept in 2025 in my view, but if a new author can show up and just say yeah this is production ready with not a single person putting eyes on it then our process is broken too.

My first thought was just getting leads approval – but what are the actual criterion for an author? And the leads already have too many things to do so we’re likely to do a crappy job. Meeting the Beman standard is fine and all, but it doesn’t ensure there’s enough tests to satisfy a production ready state. Code coverage? Well that’s got many well known flaws, but high coverage is certainly a positive sign that the work has been put in.

Anyway, this is why we need to have a discussion about the details of that state and what it means…