1. 16
  1. 10

    This post has a curious absence of even mentioning distrust of Google with yet more of our personal data, which was a major theme of discussion in all venues I saw this issue mentioned in, and certainly seemed to play a major part in distrust of an opt-out configuration.

    1. 15

      Because they’ve been very clear: it doesn’t collect personal data. It’s designed to determine how and when people use various language features, anonymously.

      1. 3

        Yes, because Google is not a company that would like to collect more and more data overtime, because it is so trustworthy that a ton of people rejected this.

        1. 4

          The post claims to summarise the discussion. It glaringly does not summarise the discussion.

          Your comment doesn’t actually address what I said, it addresses something else entirely. Do you have an answer to what I said?

          1. 7

            I’m not understanding what you’re angry about. I agree with mperham that it’s quite a stretch to consider which language features are being used by a random golang toolchain installation as “personal” information.

            I also don’t see any claim to summarize the discussion in the original blog post.

            Seems like you just want to be angry at Google and yell at anyone who dares defend them, even when (in this case) Google seems to be trying to do the “right” thing.

            1. 2

              I couldn’t find where it claims to summarize the discussion. Can you quote the passage that leads you to believe that?

              1. 3

                The second paragraph. You’re being disingenuous.

                1. 2

                  I notice you didn’t provide a quote and instead accused me of bad faith. Cool.

                  1. 3

                    See my top-level comment for an explicit examination of the second paragraph. I would characterize it as manufacturing consent or astroturfing; a top-down plan is being described as community-approved.

          2. 1

            I assumed other topics fell under the statement in the second paragraph that said:

            In the GitHub discussion, there were some unconstructive trolls with no connection to Go who showed up for a while

          3. 8

            People who read the posts seemed to agree that there’s a real problem here for open-source maintainers and that transparent telemetry or something like it is an appropriately minimal amount of collection.

            No, this is still a Google-specific desire for user data that isn’t reflective of the wider Free Software ecosystem. I went back to the discussion on GitHub and counted; of the top-level comments that weren’t moderated, 23 of 56 could be read as agreeing both that there’s a need for telemetry and that the proposal is minimal (not counting folks who clearly didn’t read the posts). 41% is not a broad agreement, especially given that there were many folks who said that yes, in general, telemetry is good; but Google specifically ought not to be in charge of such a project.

            1. 16

              I didn’t comment on the project because I agreed with Russ and there’s nothing for me to add there. I think a lot of people feel this way. Too bad there’s no telemetry, so there’s no way to measure the real numbers.

              1. 4

                My tally includes the folks who indicated explicit support. In that sense, I did “measure the real numbers” of people who spoke up. Perhaps you clicked on one of the emoji buttons without leaving a text comment? The top of the discussion currently has 162 thumbs-up, 518 thumbs-down, 4 laughs, 4 rockets, 50 frowns, and 40 hearts. That doesn’t look like a silent majority of support!

                In general, I’m skeptical of “silent majority” claims, because they usually aren’t substantiated; they tend to be used in political discussions (like this one, which is about community policy and thus political!) in order to exaggerate the size of groups of voters.

            2. 7

              I am happy to see this huge change.

              While I understand the motivation and I do believe in the pure intention of the current Go team, I do not trust Google. It would be sad to see Go being the same kind of product as Chrome (https://contrachrome.com/ContraChrome_en.pdf). To some, being opt-out does not mean much. For me, this is the thin line where trust can be lost or earned.

              1. 4

                I think as long as Go is under Google control, this will be needed. It would be nice to know if there is a way to spin up a non-profit foundation to drive Go development in the future.

                Reading this article, i wonder if there would be anyway to change the author’s mind? Is this a proposal of something undetermined, or an announcement of settled matter(in the author’s mind and/or in Google). Similar to the “go mod” proposal in the past: controversial but roll forward regardless.

                1. 6

                  I wonder how long it will take until it switches to opt-out.

                  And whether the reasoning will be “they are only a small sample and have limited resolution.”, which is from the article in the “No Telemetry At All?” section.

                  1. 26

                    I wonder how long it will take until it switches to opt-out.

                    It won’t switch.

                    1. 5

                      First of all, thank you for saying so.

                      Sorry for being skeptical. I just have seen so many instances of something like that is happening. Also isn’t that something that essentially can be forced by Google, maybe even after you leave or something?

                      But I am indeed curious. Will not opting in per default not lead to limited resolution you mentioned in the previous article? One has to be involved enough to turn it on, and preferably there wouldn’t be something nagging about it per default.

                      Another thing I wonder is, how do you think you’ll make decisions on the telemetry? It can already be hard to draw the right conclusions from surveys and reading the Go surveys I feel like reasons for giving an answer were misunderstood. After all it can be hard to know why a person chose a certain checkbox, if there isn’t a text field asking why. I don’t know if telemetry wouldn’t be even worse.

                      Let’s say a feature is barely used. How do you know why? Is it because people aren’t aware of it, is it because it’s bad, is it because it’s not important to people and so on. Based on that documentation/raising awareness, change/improvements or removal/deprecation might be sensible options.

                      And a more general question I didn’t see answered in the discussion (maybe I missed something). Do you have any real decision planned based on this? When I first looked for this I didn’t find it and it was surprising, because this sounds like something where you’d come up with an actual question first.