I really understand this as a starting position, but it can definitely be taken too far. I feel like the details matter a lot.
A few years ago there was a big dust up in the Julia community when they wanted to add a small amount of telemetry to the package servers - basically the plan was to identify real users from things like CI runs, and to be able to identify the number of unique users , which matters a lot, especially for grant writing (and a lot of academics use Julia, so this would be a boon to the ecosystem).
The core devs were super up front about it, offered easy opt-out, and even were receptive to a plan that would switch from unique identifiers for downloaders to some scheme that would give an accurate count without the ability to trace a particular download to a particular user, but a couple of prominent members of the community were incensed.
I really understand this as a starting position, but it can definitely be taken too far. I feel like the details matter a lot.
A few years ago there was a big dust up in the Julia community when they wanted to add a small amount of telemetry to the package servers - basically the plan was to identify real users from things like CI runs, and to be able to identify the number of unique users , which matters a lot, especially for grant writing (and a lot of academics use Julia, so this would be a boon to the ecosystem).
The core devs were super up front about it, offered easy opt-out, and even were receptive to a plan that would switch from unique identifiers for downloaders to some scheme that would give an accurate count without the ability to trace a particular download to a particular user, but a couple of prominent members of the community were incensed.