alt Hacker NewsI can see a lot of time was put into the report, and it helps to have the detail, but in my mind it glosses over one of the most important parts: The dispute in the stewardship of the bundler and rubygems open-source projects.
As I understand it, Ruby Central controlled the rubygems and bundler github organizations, but did not "own" the projects in the traditional sense - the individual contributers have copyright on the code, and potentially even trademark rights. By then removing access of core maintainers to those projects, they removed access to something they don't "own" themselves.
This is all complicated by the fact that controlling a github organization or repo is different from owning the trademark or copyright. But some of the original maintainers clearly felt they had more of a right to those projects than Ruby Central did.
I believe not clarifying this before making these access changes was the biggest mistake that Ruby Central made, and it's not even mentioned in this report.
Replies
There’s a ton of detail in the report so perhaps I missed it, but yes, the underlying structural/governance flaw of conflating a service, with the IP that runs that service, is a root cause here and seems insufficiently called out. The tragedy of misconception -> misconstruction -> misconfiguration is common when the bridge between governance and engineering is crossed.
The takeaway for the rest of is that separation of such concerns isn’t an abstract notion but needs to be reflected in the mechanical implementation of organisations, lest you get a train wreck later when perspectives don’t align and the whole picture crumbles.
> individual contributers have copyright on the code, and potentially even trademark
They're not the original authors of Rubygems so it's doubtful they have anything more than copyright on the code they contributed.
I don't have much skin in the game but as a passerby, I agree that the report obviously was made with a lot of time/effort but wouldn't dramatically change someone's view of Ruby Central or assure anyone this won't happen again. This is like writing an outage postmortem without really getting to the root cause and identifying what can be done to prevent in the future.