3 ms·

Even more so when they claim they don't need to sample more than 16000 reports! Like dude, you can easily find 16000 opting in across millions of users . Hell,

Even more so when they claim they don't need to sample more than 16000 reports! Like dude, you can easily find 16000 opting in across millions of users . Hell, if everyone in Google opts in, that is more than a sufficient sample. It's really a non-issue if you consider their problem statement. The fact that they are so adamant in the face of substantial feedback makes me think if there is some other ulterior motive at play.

joshuamorton 4y ago
The point is to get representative users. So you really want go's development telemetry informed only by the uses of Google employees?
- potamic 4y ago
  That was a bit hyperbolic tbh, but it's not far fetched to expect about 1% opt-ins from your user base (also considering beta channels are generally opt-out). A more reasonable way to go about this would be to launch this as opt-in first, and only bring forward the opt-out proposal if enough people don't opt-in. It doesn't have to go with opt-out on day one. This kind of requirement validation in not uncommon and I would be very surprised if the idea didn't cross them already. It would have been the far easier choice to make especially in the face of strong feedback, but the fact that they just chose to close the discussion instead of trying to move forward on the proposal doesn't bode well.
  - arp242 4y ago
    > the fact that they just chose to close the discussion instead of trying to move forward on the proposal doesn't bode well The way I read that wasn't as a "buzz off, we don't want to discuss this", but rather as a "okay, we've heard all that was said, we will think about it". At some point everything that can be said has been said. Did they need to close the discussion? Probably not, but it avoids the work and overhead of having to moderate things because it attracted quite a few comments that did need moderation (and will likely continue to attract them).
    - joshuamorton 4y ago
      Right, "this thread is not productive, we will deliberate in private and come back with a revised proposal" is the vibe I get, which is probably the most constructive approach possible in a case like this.

joshuamorton 4y ago

The point is to get representative users. So you really want go's development telemetry informed only by the uses of Google employees?

potamic 4y ago
That was a bit hyperbolic tbh, but it's not far fetched to expect about 1% opt-ins from your user base (also considering beta channels are generally opt-out). A more reasonable way to go about this would be to launch this as opt-in first, and only bring forward the opt-out proposal if enough people don't opt-in. It doesn't have to go with opt-out on day one. This kind of requirement validation in not uncommon and I would be very surprised if the idea didn't cross them already. It would have been the far easier choice to make especially in the face of strong feedback, but the fact that they just chose to close the discussion instead of trying to move forward on the proposal doesn't bode well.
- arp242 4y ago
  > the fact that they just chose to close the discussion instead of trying to move forward on the proposal doesn't bode well The way I read that wasn't as a "buzz off, we don't want to discuss this", but rather as a "okay, we've heard all that was said, we will think about it". At some point everything that can be said has been said. Did they need to close the discussion? Probably not, but it avoids the work and overhead of having to moderate things because it attracted quite a few comments that did need moderation (and will likely continue to attract them).
  - joshuamorton 4y ago
    Right, "this thread is not productive, we will deliberate in private and come back with a revised proposal" is the vibe I get, which is probably the most constructive approach possible in a case like this.