Love it. Sums up the anti telemetry 
■■■■ precisely.
The opt out approach, with the opt out right up front like Canonical does is I think the best approach, as pointed out, those who don’t care, or actively choose to send the information, can simply next past the option, those who have an aversion to sending anything to their distro of choice, can take a moment to Opt Out. Problem solved.
If you hover your mouse pointer over the colored sections of the pie chart in the poll at the start of the thread, then you will see exactly what is meant. 
Technically, if the option is offered via a popup — or perhaps as part of the installer, but I don’t think calamares would be geared for this — then it would be just as easy to have the default be “No”, and thus, for the option to qualify as opt-in. 
per the Survey at the beginning of the thread, the Opt-Ins have it. There seems little else to discuss.
This seems the better opt-in approach.
Although, the problem then becomes how best to place the option in such a way that Users are fully aware of potential benefits.
We already know, for example, that if the choice is implemented in Manjaro Hello as I have seen suggested elsewhere; and as most people will automatically de-select ‘Launch at start’ and give it no further thought; the Opt-in opportunity might be missed completely.
Of course, if the option were Opt-out, and the choice implemented in Manjaro Hello; the option to opt-out could be missed completely for the same reason. This would be sure to spark outcries of ‘Nobody told me’.
OPT-OUT:
By default > accepted < to transmit data, by a systemd or cronjob as example,
you as user have a button to disable it.
As user you silently accept every change for the future, as long as you do not turn it OFF.
OPT-IN:
by default > not accepted < to transmit data, till user accept and turn the needed button to ON,
you as user have a button to disable it.
As user you silently accept every change for the future, as long as you do not turn it OFF.
Double-Opt-In (like GDPR):
You as user got asked: can i have… yes u can… (first click by user, like OPT-IN) → ok collected then take a look → this i have collected, may i transmit it? yes (second click by user)
The second confirmation pop up’s on every transmission, as you have/able to check data of next transmit.
At the end of the day, Linux (Manjaro) is a widespread OS and have to deal with all kind of differend hardware. I think this information are not needed for improvement.
This is what i was thinking (assuming) to.
What im missing here, is the clear info and a official statement why this is introduced besides user counting and to tell us exactly how/when the data will be send.
Its specially even more important when @romangg and @philm trying to open the gate for a Opt-out Telemetry Topic.
TBH i think that Manjaro open the Pandoras Box with that and lost a big amount reputation & trust (even when this will be rejected). Was this worth it? Time will tell…
I personally will keep staying with Manjaro and i will even to deal with disabling this Telemetry. Which probably can be used to identify (in combination with browser fingerprint) my device when i use a VPN, when its unasked to send home.
Alone this question about Opt-in vs Opt-out comes up, gives me a sick bitter taste in my mouth, that i don’t really like to donate my money to Manjaro anymore. And this makes me really sad, because i was really behind Manjaro’s decisions in the past but this 2 Topics are going to far.
I hopefully can change my mind in future about Manjaro but for me this is a big deal.
They also said, that (amongst other things) they will be able to prove the size of manjaro’s userbase and attract financing this way - or at least that’s what I understood.
Sorry to intervene. It’s not a matter of who’s right or wrong - it must be really clear what opt-in and opt-out are.
Opt-out means that data is being sent, until the user deactivates it.
Opt-in means that no data is ever being sent, until the user (ever) activates it (if he chooses to activate it).
What option box will be pre-ticked might make the request more or less polite, more or less “pushy”, more or less whatever, but nevertheless opt-in and opt-out are not being defined by the preselected values in a settings dialogue, but by the preselected (predefined) behaviour of the telemetry program (if it sends dat or not, until the user configures it).
Opt out data collection is the wrong way to do it in my opinion. Even if you assume that people don’t understand what you are collecting from them (which from my POV is a wild suggestion considering the average type of user who uses linux), just taking the data anyway is not exactly a good look.
“Hey I want this from you and because I don’t think you understand what I am taking, I think I don’t even have to ask for it and I can take it anyway, doesn’t matter if you agree.” - This is obviously a bit of an polemic exaggeration, but at the end of the day it is what opt out data collection suggests.
I want to point out that (as long as it is opt in and optional) I am actually willing to provide my hardware data and stuff. The moment it becomes opt out, I will immediately opt out just on principle.
Other companies with more or less tech savvy customers have shown, that you can collect data prompted and still have a good participation statistic. For instance valve with their Steam Hardware Survey (Which isn’t to different from yours) as a counter argument to the case of KDE. I think opt in is the only proper way forward (Or at least a prompt which allows you to review the data like with steam, before you actually submit anything). Just making it opt out because people don’t understand enough to know what they are submitting is a very similar way of doing things to Microsoft and Google. And they rightfully earn a lot of criticism for it. (And for this point it isn’t important that those two collect more or more sensible data. That is an entire different discussion)
I hope you do the right thing (from my perspective) and introduce the feature as opt in.
I agree. As far as I can tell the member concerned reacted to my (purposely) oversimplified diagram/description of opt-in vs opt-out early in the piece.
When presented with an alternative outline of how it could be implemented, they continued to focus on the initial objection and ignored my attempts to explain further. Until that point I had assumed I mustn’t have explained it well enough. It’s difficult to have a conversation with anyone who cherry picks comments to support their own agenda, whether justified or not; which is indeed why I chose to… opt-out.
Thanks for your comments.
I’ve quoted it at least 3 times, but that doesn’t stop people from saying it’s something else.
So it seems they’ll only be convinced by a team member, and if my interpretation is wrong then the vote needs to be clarified because it says one thing and means something else.
Wow, what long discussions over just a few days. 
With an opt-in approach, Manjaro’s public statistics become less accurate and unreliable since fewer users participate, resulting in skewed data that would be useless or boring, forget that. (That means the project would be dead, I feel sorry for your creation and effort @romangg)
On the other hand, an opt-out approach would make the opening statistics more accurate. Maybe the result would be more insightful and informative for everyone.
Ultimately, it’s up to you as Manjaro members. Having fun with a long discussion.
This is just my only post here, I don’t want to add more unnecessary political spams blabla. Bye 
I agree. To my mind a clearly displayed, at Install time option, that is informative, would be best.
But, clearly the Opt-In vote has the numbers.
One thing @philm mentioned in the video, was the Ubuntu Opt-Out Method which sends the Opt-Out choice to Ubuntu, so they can count the number of actual users. They may not be able to gather any other useful information, but they can know how many Installs they have. That in and of itself is very useful information.
The Opt-In method provides Manjaro with nothing, and therefore the entire Linux Sphere nothing.
This would only be effective if said choice was actually sent after a certain amount of runtime and/or number of reboots though, to eliminate “test” installs.
I’d think some way of filtering out VM installs should also be incorporated; something to detect that it’s running in a Virtual Machine? (Probably already been mentioned earlier in this thread, or the other one).
That information is already covered in the data mdd currently collects. Separating the virtual machines from the bare metal installations is as such also already handled at the server level — see Manjaro Metrics.
Good point. I don’t know how or when Ubuntu gather the Opt-Out Data, as I was unaware of it. Though I suspect it is done at install time.
EDit: …or when the user sets up a pre installed system.
That’s another point regarding when the option is displayed to the user, given that many or most of us here do our own installs, there are many people, and a growing number I would guess, who buy a complete system pre installed.
Off-topic for this thread, but perhaps @romangg can explain to me what fiendish sorcery would be involved in having Manjaro installed on a non-existing filesystem. 
In casu, I am referring to the (thus far) single mdd data submitter who has managed to create an ext5 filesystem, of which there is no documentation to be found anywhere on the interwebs — unless my 
-fu would be horribly inadequate.
Edit: Make that two undocumented filesystems. One data submitter is using something called ruffs. I’ve never heard of that either. There is such a thing as refs — Resilient File System, stylized as “ReFS” — but that’s a Microsoft filesystem, not a UNIX filesystem, and I don’t even know whether the Linux kernel has any support for that.
The only other potential filesystems I know of which could be implied would be reiserfs and reiser4, but reiserfs has been dropped from recent kernels by upstream, and reiser4 has never been included in the upstream kernel.
Maybe it was one of the examples of someone specifically fuzzing the data, as was demonstrated in the other thread.
Good point, and I suspect you would be correct. After all, mdd is only a Python script, and there are many Python coders among our members.
My OCD can thus now safely be laid to rest — or at least, on account of this particular matter. 
Hey this me, not he. I thought I made it clear that I am an anthropomorphic kangaroo.
“She”. @tracyanne is a female. Look at the pouch with the joey.