I'd imagine that "opt-in" could even mean you have to install a separate program/package to send data that's been collected. If it were very separate from the compiler itself, would these security concerns still be a problem? I for one would go through the effort of opting in since I want the ecosystem to improve and I have the luxury not to be dealing with high-security code bases.

On Fri, Dec 9, 2016 at 4:48 PM, George Colpitts <george.colpitts@gmail.com> wrote:

I would opt-in. I also agree with Simon that privacy is no longer a big deal although I do believe that most companies do telemetry with an opt in policy. If it's opt-in why would anyone have a problem with telemetry?

On Fri, Dec 9, 2016 at 1:46 PM Tom Murphy <amindfv@gmail.com> wrote:
On Fri, Dec 9, 2016 at 4:50 AM, Simon Peyton Jones via ghc-devs <ghc-devs@haskell.org> wrote:
I have wanted telemetry for years. ("Telemetry" is the term Microsoft, and I think others, use for the phone-home feature.)

It would tell us how many people are using GHC; currently I have literally no idea.

In practice I think the best data we could get is "how many people are using GHC && are willing to opt into phone-home," which seems like a rougher number than e.g. downloads of ghc/HP or number of downloads of base/containers or something similar. I also would not opt in.

Tom
_______________________________________________
ghc-devs mailing list
ghc-devs@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

_______________________________________________
ghc-devs mailing list
ghc-devs@haskell.org
http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs