Graduation is not meant to be some kind of super impossible bar. It
should be pretty easy to go from successful Incubation to Graduation,
provided social conditions are met. Let's not assume that all
"collaboration" must be between multiple sellers of the same software.
On Thu, Oct 1, 2020 at 6:54 PM Stefano Maffulli via lists.cncf.io
On Tue, Sep 29, 2020 at 10:43 AM Liz Rice <liz@...> wrote:
Problem statement summary: projects that are controlled by a single vendor struggle to meet the current graduation requirement to have maintainers from multiple organizations (for valid reasons such as, they tend to hire the folks who are expert in that project). This multi-organization requirement is intended to address two concerns:
1. longevity of the project (in the event that a vendor is acquired or goes out of business)
2. ensuring that the project roadmap is community controlled, and not only run in the commercial interest of the vendor (we want to avoid feature hold-back)
We recognize that the current multi-org requirement may not be the only (or even necessarily the best) way to address those concerns
I think it's a serious mistake to de-emphasize diversity of employment among project maintainers in a consortium that is all about collaboration. I'd love to explore other venues before throwing the towel.
Maybe the problem is with the word "graduation" and the way it's portrayed as a destination, rather than one of the criteria to assess longevity and community control of the roadmap.
Pieces of the conversation from Matt and Alexis hint at a source of misinterpretation of what "graduation" means. The fact that CNCF is showing a linear progress from incubation to graduation maybe is contributing to the confusion.
On Wed, Sep 30, 2020 at 4:27 AM alexis richardson <alexis@...> wrote:
CNCF Incubation tests for production use and technical DD. It has a
high bar. Graduation is oriented towards sustainability including
some of the matters you touch on below. Graduation is more about
sustainability and governance, than about production use. Those are
all related in the end of course.
Sustainability, governance and production use are correlated but quite independent variables. IIRC the Eclipse and Apache Foundation played have experience exposing a series of indicators in a maturity model. Some adopters of software may have more tolerance than others for things like "employment diversity of maintainers".
How about rethinking the flow from incubation to graduation not as a ladder but rather as criteria for a decision-support matrix?