Opt-in / Opt-Out? A breakout topic for the F40 Change Request on Privacy-preserving telemetry for Fedora Workstation

Why not have the user have to choose an option before hitting next? Opting in could be set as first choice, making it the easier one to go with, but people will still be more aware of the entire setting.


What I don’t get is if opt-In turns the data into “garbage” because only the dedicated will turn it on, how does it further the goal of “making Fedora the premier cloud developer platform” to make it opt-out, if it’s trivial for developers to disable it but not for the ignorant and uncaring? Wouldn’t that be “garbage” data as well?

As a dev, it is a lot easier for me to disable all telemetry without ever researching what it does, than to painstakingly look at everything it sends and see if it’s compliant with my company or client’s policies or NDAs.


I could theoretically be useful, but most likely it isn’t, because those 5% are people who for instance care a lot more about Fedora than the average user which most likely puts them in the expert linux user category. So when you look at their data and if you assume it is representative you might conclude that Emacs is the most used IDE out there (random example).

Opt-in is the way to go if Fedora doesn’t include an unskippable button that forces the user to read about what they will click. Guarantee consent!

Expecting that people will read what the telemetry does is optimistic. They will click “next next next” without reading, I know more people who do that than actually reading.
Opt-out can take advantage of people’s ignorance and gather their data without permission, as reading and understanding can be avoided.
This is exactly what Windows does, they collect data from users and get away with it by making the dialog skippable and boring to read.


Please do try to understand the point.

I think I do. Like I get how the data is better. I get how people from RH want the data to improve the product and have a justifiable business driver to show their value/hours. I’m all for those things.

While it is not morally the best option […]

This is where we disagree. I don’t think there’s enough justification to damaging our other values for the potential gain.

The only other option I see here that I don’t think has been discussed is that since this is already limited to Gnome, then it should be upstream. This doesn’t have to be a fedora related effort and could instead be proposed in the Gnome project.


You have to decide I thought you don’t want our garbage data.

1 Like

Why not have the user have to choose an option before hitting next? Opting in could be set as first choice, making it the easier one to go with, but people will still be more aware of the entire setting.

This has been proposed, and it seems like people are good with it as long as you can’t just hit ‘enter’ or next that it still needs to be physically selected. I believe that it was still “opt-in” and they’re worried about the user base size.


So are you saying you be fine with the metrics gathering if it came as an integral part of upstream GNOME?

1 Like

A loud minority does not the whole crowd make.

I, along with I believe quite a few people, wouldn’t mind telemetry if done right, if it is actually privacy-preserving and if it does actually help the developers of the software we use to have a better understanding of not only our (as in the people that know how to file an issue or report a bug to those developers) but other people’s issues with the software we use.


I’m saying Gnome could gather that data instead of fedora.

I don’t they’d do that, but it’s certainly an option and then it would benefit a lot more than just fedora.

What guarantees that 5% cares a lot about Fedora? It could be 40%.
Most users could just read and say “oh so that helps them, why not, click.”.

But the problem with opt-out is that if it isn’t guaranteed that the user will read what’s going to happen to their data, you are essentially stealing their data.
Opt-in avoids this problem by requiring input to enable the data.

Whatever the method, making the numbers of users who said yes/no to the question publicly available can answer if the percentage of users who say yes/no is 5%.

This is pretty much the meat of the “Ask yes or no with no default” type suggestions.

  • Is it akin to “opt-in”
  • Will it add undue friction?
  • Will it be meaningful?

Which, would be worth at least some thought and consideration.

Well the data would likely be useful to others regardless, but one of the motivations for doing this is to make Fedora better and attract more users to Fedora, and thus getting data from a ton of different distros might actually make the data less useful for that purpose.

Then I would say the proposed scope of the RFC is wrong.

1 Like

Let’s try it this way, I give you a chance to
understand what users prefer but do it without a preselected dark pattern, ask!!! Don’t assume the normal user won’t care about their data privacy and it’s not ok to just say “you are good when I collect this data about you and send it to our server, aren’t you?” It’s intrusive and not the right way to handle this. Only because many other are doing it doesn’t mean you also have to do it. opt-out is not an option.

And i keep repeating it you said:

This is so incredibly disrespectful for everyone actively choosing to send telemetry.

And about Gnome, you can get the data in the backend without telemetry, just look at the downloads😋.


This option has been presented multiple times as a compromise position, and it is one I am fine with. No default choice, forcing an actual choice, at that point I have no objections about the ethics of using telemetry in Fedora. It can’t be skipped and it can’t work on the hopes that the user ignores it so data can be had. It is probably the only reasonable choice.

Thanks for the excellent post by the way.


For the record, I’m behind making it an explicit choice.


I’ll have to agree on this. It might well be a good compromise position. I initially viewed it as opt-in by a different name, but more thinking and analysis in my head thinks this would be a good way to nudge users either way. We just have to do a good job with the value proposition immediately presented on the page.


This is the only right choice. Let the users decide. Do not set the choice for the user. Any choice the user makes it is completely their fault, and not the designer’s fault which could lead to accidental consent.

But this must be free of any dark pattern. Agree on the left, refuse on the right. Same color, same importance. No deceptive wording. In other words: trust the user and don’t trick him/her.


I’ll take opt-out