Hi folks, I've been thinking about this for a while and wanted to share it with you and have a discussion about it.
Our knowledge levels about Linux are certainly different from a person to an other. So when It comes to debugging and feedback, there is very few people that can properly communicate it to the devs. The majority either won't know how to, don't have the time, or are like me : a lazy bastid 😂 .
So, a telemetry program that collects this important bits and bytes and get it in a convenient way to the dev while it leaves out your most personal info would be very helpful to accelerate the OS's development. This is not only good for Solus, but a good technique for the whole Linux ecosystem IMO. Too much data is already collected without our consent, but if I can give my data willingly in favor of a better OS for us all, I'm doing it.

What about you? Yay or Nay ?

    Since the Solus devs haven't shown any interest in implementing telemetry so far; and every question about it has been answered along the lines of "We don't do telemetry", I would think it's a "Nay" ^^

      eljondy a telemetry program that collects this important bits and bytes and get it in a convenient way to the dev while it leaves out your most personal info

      Is this even possible?

          brent it should take the logs and the config while leaving out the user name and such personal configuration. something similar to Ubuntu's approach

            Staudey Yes they don't do telemetry. And just say the word is kind of a sin in the open source world. But if we give the subject a deeper thought and do telemetry the right way, we may end up with far better results

              Perhaps there are still people who believe in relationships between humans being part of a community also means asking for help and reporting issues.

              You know a distribution is a collection of packages coming from many different sources, even if there was some sort of telemetry on Solus developed software - which is not something the team has any interest in - it wouldn't cover the stuffs that aren't developed by Solus, so it would be just useless.
              Also people behind Solus are volunteers, not robots so implementing a system to automatically send error reports for users who "don't have the time" hum hum...

                kyrios So the next task would be figuring how to mechanize/automate our volunteers, right 🤣

                    kyrios for users who "don't have the time" hum

                    this is a fraction of users there that don't value the importance of debugging and feedback , but most of people we're getting on Linux lately are just not that tech savvy but may decide to help with activating an option during installation. I get it, it's a project not a "product" . and the human interaction is highly valuable. but this technique still can be very handy when filing bug and DMing the computer's telemetry ID and time of the incident for debug purpose would make it a lot easier for devs and users .

                    Anyway, It seems I have a different and maybe VERY optimistic opinion about the right way to use Telemetry that isn't what anyone agrees with.
                    Well, kyrios @brent @Staudey thank you guys for your thoughts . Much appreciated

                          eljondy Not sure I contributed anything to the discussion. After an IP address and remote entry, anything bundled in your program can 'scrape' for anything. The whole "good/evil" thing starts there and I won't be the one leading the discussion🙂--far too subjective this issue.
                          I think you raise a fair question in the name of debugging.
                          If Solus has no interest in telemetry, I am very satisfied with that.

                            eljondy but most of people we're getting on Linux lately are just not that tech savvy but may decide to help with activating an option during installation.

                            People moved to Linux with the pretense that it requires you to be a bit more tech savvy than usual and need to be open to learn. Having said that and this being the case for as long as it has; you want to make a Windows 2.0? Where everything is simply sent whenever we deem necessary in the hope that we can "help" users? How? What do you want to sent? What systemctl status <package> returns? And for what? You want to push them an update right away? Want them to submit an error of it? Why not give them the option to, and make that as easy as possible. A simple pop-up on any error and a link to a respective issues-page would do just fine.

                            But of course, it's easier to have users sent everything to you for you to sort through. 👍️

                            I get it, it's a project not a "product" .

                            So you do understand the privacy aspect of it, but you say f-it anyway?

                            but this technique still can be very handy when filing bug and DMing the computer's telemetry ID and time of the incident for debug purpose would make it a lot easier for devs and users .

                            Of course it can be very handy, just ask Microsoft! 😂 On a serious note though: Why do you need time of the incident? Don't you get that when the telemetry comes in anyway? And why does it need to have an ID? To catalogue the specific computer? Don't you already get that with the IP address? You know, 'cause they're already pretty unique. Oh but for people on dynamic IPs, right? Or for VPN users? So you render the VPN useless (in a way)?

                            With all that I'd rather just willingly post my information on a Github Issue page than have you sent it via your program with who knows what else.

                            If you're only surviving because you get telemetry, you need to up your game and change the way you work.

                            Anyway, It seems I have a different and maybe VERY optimistic opinion about the right way to use Telemetry that isn't what anyone agrees with.

                            The very idea of telemetry is something that's not supported by many for exactly the reasons stated above. If I want to ping a server, I'll do that myself, thank you very much. If I really can't figure it out, it's up to the developers to ease the user into getting the right information and then proposing them with the option of submitting that. And not just sent anything over the internet as soon as an event happens. Just make it as simple as possible, and stop taking the easy way out which almost gets abused by default.

                              dbarron Don't get me wrong, I don't tell that telemetry is always bad. On large projects, it can be a precious help, but on human-sized project like Solus it is not. Not even mentioning it raises legal & privacy concerns. Last but not least when you distribute a software you make, even if it relies on 3rd party libraries it could make sense but Solus doesn' develop the linux kernel, the window manager, nor the huge majority of the other pieces of software it ships -> Telemetry on what exactly ?

                              Here what I personally see is someone lazy that wants a system that would report errors automatically and wrongly believing it would speed-up their resolution while in small teams projects it would probably be the opposite (unless they are automatically sent to /dev/null) because it could generate a too much data for a small team resulting in a huge backlog and the devs would spend more time analyzing/understanding the data and trying to reproduce the problems rather than actually fixing them and all the time spent on that is time that isn't available for working on something else. It's so much easier to ask the user "could you try this?" or "what if you do that?".

                                  So let's be perfectly clear here, Solus does not collect telemetry data and has no intention of doing so any time soon. GDPR and other legal concerns aside, it is very difficult from a technical perspective to do so without accidentally fingerprinting a specific device. I have considered the possibility of a Solus Hardware Survey when I have time to circle back around to my work on Linux Hardware Qualifier, but that would be purely opt-in and completely open-source to allay any privacy concerns. Even for that I would have to spend a considerable amount of time investigating ways to prevent duplicate entries without accidentally fingerprinting. But on the subject of automated bug reports, absolutely not. I would much rather have on-demand reports that you could attach to a bug-report. And still, significant effort would be needed to protect your privacy.