The change proposal F40 Change Request: Privacy-preserving Telemetry for Fedora Workstation (System-Wide) is — as is appropriate for such a big, important topic! — getting a lot of discussion. In order to keep the conversation from becoming one long list, I’m making a number of break-out topics for various important sub-topics that are emerging in the discussion.
This topic is about the specific issue of opt-in / opt-out. Post in the main thread that are primarily about this will be moved here. If you have more to add on this particular topic, this is the best place for that.
See also! I’ve made an informal, non-scientific (and certainly non-binding) poll specifically on this topic. See Straw poll on your preferences about opt-in / opt-out for possible data collection
There must be an internal contest at Red Hat to see which group can generate the most negative press and/or piss off the most people. The fact that there appears to be a strong desire within Red Hat to collect user data from user’s private machines using an opt-out mechanism is troubling, to put it diplomatically. If you want to know how users interact and value different aspects of an offering, you need to interact with your users. Telemetry isn’t a substitute for that interaction and won’t overcome ineffective and/or bad decisions by management too disconnected from users.
I understand the need for telemetry. But it needs to be opt-in. Even if it’s one of those checkboxes in the initial setup steps (I think they call it “GNOME welcome”?), similar to “Enable third-party repositories”. It also then needs a toggle on the settings page.
But yeah I’m not against it, though I have to admit, highly reticent.
I’ve written and rewritten this reply several times, and I just can’t find a more constructive way to put this, so: if this isn’t explicitly opt-in, I hope FESCO rejects this.
I have a lot of specific concerns, but honestly, that’s the one I care about. Since that’s been preemptively (and repeatedly) rejected in the proposal, it’s clear we’re not going to be able to convince the proposal authors otherwise. And because I suspect FESCO will approve this, it’s not a productive use of my (or anyone else’s) time to write a novel detailing my individual worries when the one I’m most concerned about is unlikely to change.
But I’m at least going to object.
This is one of the worst ideas I’ve read in the past 15 years I’ve been using Fedora!
Well where do i start.
The first Problem privacy-preserving Telemetry does simply not exist. RedHat is controlling all the Servers which the data is send to. Nobody will run their own telemetry server, for which reason?
No more further explanation required.
The only privacy-preserving way is no telemetry.
What’s wrong with your wording?
I’m not interested in Fedora anymore and now???
If you really want to go this path to telemetry hell than OPT-IN is the ONLY solution which is kinda acceptable.
(There will be a privacy focused fork of Fedora with telemetry removed, if this is your goal)
Nobody in the whole Fedora Community will accept this opt-out nonsense. The only way, this going through is, if RedHat is taking the control away from the Community and measured by the last decisions they made, this is not that unlikely.
Opt-In or nothing, that are the only options.
Ed Marshall is right: this is the one point of the change proposal that I’m going to be extremely inflexible about. I don’t think it’s worth implementing telemetry at all if it is opt-in rather than opt-out. Opt-in telemetry is basically garbage and we don’t want to be making development decisions with garbage data. We’ll wind up collecting data only from our most hardcore Fedora fans, who are not representative of typical users.
I feel the need to add to this. Although the exact data being collected isn’t entirely clear, based on the wording in the proposal it sounds like we would be collecting what I would call highly intrusive data.(At the very least, it talks about tracking which applications are installed and run)
I wouldn’t want that data to even be collected locally never mind transmitted to a 3rd party.
I honestly don’t understand how collecting information that detailed could even be considered as anything other than opt-in only.
Yeah if you’re going to be inflexible about that part, that might be a problem. There might be a chance if you completely open source the data resulting from the telemetry. “All cards on the table” so to speak.
But I’ll be completely honest, as I pointed out in Matrix earlier, telemetry is one of the reasons I switched away from Windows.
opt-out telemetry will show you one thing, users opt-ing out of using Fedora.
I’m not involved with Red Hat at all and I can assure you that I enable telemetry in FOSS whenever I can (Firefox, KDE Plasma, etc.). I would enable all telemetry on Silverblue and GNOME if I had the option to.
Sure ‘no one’ wants telemetry, but unfortunately, we don’t live in an ideal world. We need to gather insights over some data. Making it opt-in means that no one will enable it, so it’s best to have it enabled by default, and improve the experience as much as possible.
Asking for users’ feedback is very limiting as well, because a large portion of users don’t feel like answering questions.
And it’s nice that people who would like to contribute in that way have the option, but it shouldn’t be imposed on users. They should have to explicitly opt-in to telemetry and have the option to completely remove all of the telemetry related packages locally if they so desire.
Again, explicitly opting in doesn’t work, especially not at a large scale. The majority of computer users aren’t technically savvy. The people who opt in aren’t representative of the majority of computer users. What we need is something that can be representative.
Hari you just contradicted yourself, you said you enable telemetry where you can but said that nobody wants to enable it. I also enable telemetry in software I trust (KDE for instance) but only because I want to and only because it’s opt in.
Well yes, I enable them because they’re opt-in. If they were opt-out, then I wouldn’t be enabling them, wouldn’t I?
Also, I put “no one” in quotes. It wasn’t literal.
The point I’m making is that only the people who want to provide feedback will. Again, this isn’t representative of the majority of users. It’s likely going to be misguided if we make it opt-in.
Either way, if this change proposal goes ahead as an opt-out I will no longer be contributing to Fedora (KDE SIG) and will find another project to contribute to.
I had a long reply composed, but let me step back first and ask a question to make sure we’re on the same page. For several years now, Canonical has been including “opt-out” data collection as part of Ubuntu. By opt-out, I mean that it is presented like this:
This is what I take opt out to mean. As I understand it — from Canonical and others — having the option on by default really does make a big difference in the quality and quantity of data the project gets.
Is this also what those of you upset by “opt out” are thinking? Or, are you envisioning something buried deep in secret settings which no one will ever find or notice?
I guess I’d be fine with something like this if
No, don't send system info was the the default choice.
That’s easily skipped in the next next next world that users live in.
That’s opt-in and what most people here want.