2 ms·
To be honest, I'm pretty against telemetry in Go, but it's not about the compiled code itself. They want to telemetrize the go command and compiler itself to co
by bayindirh 4y ago
To be honest, I'm pretty against telemetry in Go, but it's not about the compiled code itself. They want to telemetrize the go command and compiler itself to collect data about which language features used (I assume), and how the go command is used.
Again, I'm pretty against it. My red-line is opt-in telemetry, and even that's a place I don't want to arrive.
- vivegi 4y agoI think we are in agreement. The problem I have with the principle of forced telemetry is I have no way (or more precisely, time) to understand that their telemetry red-line stops with command line logging/analysis. So, I have to assume the worst with the toolchain. Microsoft .NET also introduced opt-out telemetry (ie., turned ON by default and you could opt-out by setting an environment variable DOTNET_CLI_TELEMETRY_OPTOUT=1|true|yes). These are just dark patterns by big-tech and erode developer trust. Ken Thompson's Turing Award lecture on "Trusting Trust" comes to mind. https://www.cs.cmu.edu/~rdriley/487/papers/Thompson_1984_ReflectionsonTrustingTrust.pdf https://www.cs.cmu.edu/~rdriley/487/papers/Thompson_1984_Ref...
- bayindirh 4y agoBoth of us are completely on the same page. Funnily, I have made the same reference to the same paper [0]. The comment has a "let's talk about how we stop there and not move forward than that" tone, because frankly, I don't think Go maintainers and Google will say "Oh, we overreached, sorry. Let's abandon it and never talk about it again". However, even discussing telemetry and being so adamant about it made me reconsider the extent I'm going to use Go in the future. [0]: https://github.com/golang/go/discussions/58409#discussioncomment-4928702 https://github.com/golang/go/discussions/58409#discussioncom...