This poll uses @cassidyjames’s suggestions for opt-in/opt-out terminology. I’m keeping Cassidy’s definitions and removing some of the commentary — see the link for details. Please note that no one is suggesting or considering a “buried opt-out” approach here. That option is just here to be complete. I would be willing to accept[1] Fedora Workstation gathering FESCo-approved aggregate metrics, using the following approach: poll By “willing to accept”, I mean something between “I’m satisfied t...
I think a lot of the arguing people are doing here is about the usefulness of opt-in vs opt-out. And personally I tend to agree with the side of the opt-out group; telemetry that users opt-in to is just less useful overall for figuring the average needs of your users. Opt-in is way too self-selecting and shows you very little about what actually needs to be worked on for everyone. However, if the telemetry is not privacy-respecting then opt-out is not a good thing at all. But I think I trust the endless OS system that fedora is trying to use.
I would say opt-out is fine as long as the option is presented to the user early on in the UX. Like for example during installation. If it’s opt-out but the option to do so is hidden then that’s not good.
I think a lot of the arguing people are doing here is about the usefulness of opt-in vs opt-out. And personally I tend to agree with the side of the opt-out group; telemetry that users opt-in to is just less useful overall for figuring the average needs of your users. Opt-in is way too self-selecting and shows you very little about what actually needs to be worked on for everyone. However, if the telemetry is not privacy-respecting then opt-out is not a good thing at all. But I think I trust the endless OS system that fedora is trying to use.
I would say opt-out is fine as long as the option is presented to the user early on in the UX. Like for example during installation. If it’s opt-out but the option to do so is hidden then that’s not good.
That is a great middle ground