"I think I've mentioned this a while back, but one only needs to look to the Guava library team to see what Google's general position is on community contribution:"
Can you explain why it makes sense to generalize a company of 50k developers, which has released 4000+ open source projects, on the basis of precisely 1 of them?
Why doesn't it make sense that it will end up with a variety of projects with a variety of models?
Especially given that is true of the world writ large?
I simply do not understand this fixation with trying to rigidly pigeonhole Google on open source.
I think the point rmrfrmrf is making is that Guava is a typical example of Google's attitudes towards the two-way relationship of open source. Google developers (and many other big software companies) are more than happy to open their code and promote it as something that the community should adopt, but seem rigidly dismissive of any code contributions to the project from outside. That may not be the case for every Google project but it does feel like it's a common one to me.
What Google projects would make good counterpoints by demonstrate a welcoming attitude to external contributions?
Kubernetes is a reaction to that state of closeness, and was explicitely designed to bring Google’s philosophy (Borg) to the world. Its strong focus on getting outside people involved is the core reason of its existence, so it’s a bit of a special case IMO.
An issue is copyright ownership. Even if a contributor assigns copyright, it is ineffective if they didn't hold copyright. This opens the project to legal peril.
That sounds reasonable except Google could write some boilerplate for maintainers to include in a project's README.md, contributing.md or license file that explicitly states they won't accept community contributions because copyright would be an issue, and they don't appear to have done anything like that.
Can you explain why it makes sense to generalize a company of 50k developers, which has released 4000+ open source projects, on the basis of precisely 1 of them?
Why doesn't it make sense that it will end up with a variety of projects with a variety of models? Especially given that is true of the world writ large?
I simply do not understand this fixation with trying to rigidly pigeonhole Google on open source.