How can you accept or decline something you are not offered?
Donāt know.
Isnāt that part of what is being discussed here?
This here is a discussion - but not one about a thing that is already present.
Calm down.
What in my part of the discussion (a simple question) is not calm?
Nothing.
I would have thought you got the meaning behind the phrase.
Iām pretty sure you did.
So how to make it opt out without a pop up?
A hook at update time? Notify peeps in the terminal, something similar to the grub update messages? No need for interaction in the internet with the user.
Why would it be done/implemented that way? - ā¦ without a popupā¦ to opt out ā¦?
When opt-in is still on the table ā¦
I think you assume too much 
The thing first has to work, before it can be deployed in whatever way ā¦
If 0.1% of user base can opt in because they know of mdd from socials and that (and pop ups) is/are seen as an issue by romangg, it is a reasonable assumption imo. Certainly at this point.
Of course this can change, as you say, early days yet.
Iād simply do it - like Fedora does
see my a few posts above ā¦
but that doesnāt solve the already existing installs
I do not know why ātheyā now want to now know things ā¦
presumably to make the default experience better than it is
I canāt really provide - all my Manjaroās are in VMās
Manjaro is just my Hobby - I donāt actually use it (anymore).
1 Like
First of all: Thanks for your great work and IMHO a beautiful distro.
I just ran mdd and it worked fine, so now there is one more Xfce system in the data set 
I am also quite happy with the Manjaro Metrics page. Looks cool 
My 2 cents on the topic:
a) Please make it Opt-In.
I agree that many users probably will not care, but ātelemetryā, ādata collectionā and āprivacy concernsā have become trigger points for heated discussions, as can be seen in this thread.
Why risk to scare off users (or potential new users) or even provoke bad press by implementing telemetry as opt-out?
b) Being a nerd makes it easy to run mdd in the terminal, read through the collected data after --dry-run, check and understand and agree to send the data.
Most average users will have trouble to understand what the data means or even running mdd in a terminal window in the first place.
I recommend a nice dialog box with a short and comprehensive description, followed by a display of the --dry-run output and then ask if the data should be sent. Thatās what I would consider transparent, even for non tech-savvy users.
Honesty and transparency wins.
c) What about runnig mdd once after a big update? And / or if the user switches to a new kernel?
Much like pacnew-checker. I guess this could be more easy to implement than a systemd service? (Iām NO developer, btwā¦)
Honestly, I donāt like the idea of a systemd service to update the data in regular intervals.
The collected data will not change randomly over night, nor every time the system is booted. (Ok, maybe for a handful of geeks who like to switch their DE thrice a day, but how many are there?)
I have no objections to telemetry, as long as it is done transparently, with my consent and without personal data, as does mdd now.
8 Likes
I saw that in the source but it is a problem as itās a unique identifier of a device even after those transformations and wonāt change over time: Therefore it can be used to track devices.
This is intentional: on the receiving/server side you probably want to count each device only once (per day).
Itās even stated in the README.md: mdd/README.md at 580af797d5fdd583231b2b050b78579df758c8b3 Ā· manjaro/mdd Ā· GitHub
So your /etc/machine-id contents wonāt be sent and are still a secret but every one of your manjaro devices is trackable over time.
1 Like
Please - it is necessary to be able to uniquely identify devices - otherwise it is ridiculously obsolete to do.
It is not to track device - as in knowing where it is or who is using it or which applications are installed or your username or your hostname or your ip - it is to avoid duplication and ensure that only unique systems are counted.
How many of you use Firefox or Chromium?
Are you really that naive - thinking the device is not uniquely identified over time?
You are all OK with your browser being identified - but you creating furore over how mdd works?
At least the Manjaro Team is open and transparent about it - even providing a public visualization of the data.
Try opening Firefox about:config - search for the phrase id
Install an application capable of opening sqlite database and go over your browser files in ~/.mozilla and ~/.config/chromium - you will be surprised.
just make it OPT-IN + transparent (as it is already) and there is NO problem,
and plz do not compare a browser with an operating system, that is like comparing apple with pears, the browser you can switch to a better/clean system within 5 minutes, but not an os.
There is a saying by Confucius that fits perfectly, based on it:
Manjaro has three options in my eyes:
-
Think about the feedback, tackle the challenge and then do it right, preferably as an example for others (this also gives you a new unique selling/using point)
-
Ignore the feedback, do it as planned, and accept the bitter pill of āuser lossā
-
or scrap the whole thing and leave it as it is.
But no matter what you decide, the news about telemetry service is already spreading like wildfire, and the comments I found are mostly the same = bad idea as opt-out 
2 Likes
You are all looking at this the wrong way.
Everybody is going on about privacy concerns - yet the data has nothing to do with privacy.
Privacy is you personal identifiables - and your basic hardware does NOT identify you as a person - and there is nothing, not even remotely related to private information.
The hardware information on the other hand is useful in many ways.
- How many of you have ever speculation on how big the Manjaro userbase is?
- Is it really needed for Manjaro to support 4.x or even 5.x series of the Linux Kernel?
- How much effort do the team need to put in to continue to promote X11 instead of Wayland?
- How many low-powered systems is running Manjaro - as in do the team need to cater for low-end hardware?
- How many Manjaro systems do actually use Nvidia?
- Which Nvidia GPUās are most common - do Manjaro really need to provide those older drivers?
- How many users change the system to use Pulseaudio instead of Pipewire?
- How many users think of tomorrow?
- What if Manjaro Linux announced sunsetting the distribution due to the team deciding enough is enough?
There is a lot of questions like the above - the answers would be helpful and the mdd script is designed to provide data that could possibly give insight into those.
And there is no privacy issues connected with the answer to those questions - so please donāt claim it is an invasion of privacy - because it is not - at least I donāt see it that way.
- Booting Manjaro Linux on a live ISO doesnāt count.
- Installing Manjaro Linux as a test doesnāt count.
- Using Manjaro Linux regularly does count.
9 Likes
i am not talking about privacy concern data, that is all fine so far (at moment),
the point is:
OPT-OUT:
You can do anything ā i need to proove it, every time, if acceptable for my userbase!,
as my userbase perhaps is as example a corp which underlay compliances!
OPT-IN:
You ask user can i haveā¦ yes u canā¦ transmit
Double-Opt-In (like GDPR):
You ask user can i haveā¦ yes u canā¦ ok collected ā take a look ā is OK ? yes
Together we can all, as community, answer mostly of your questions, but call for our acceptance of transmitting data, and do not collect it automatically without our accept! (that is a break even point, to stay or to go to another distro in my eyes)
What is the problem to ask users one timeā¦if the following services will be accepted?ā¦ then the user have the choice, if he not decide (incase of will or knowledge or interest) turn it off by default.
When a user is not willing to share such data, but it is inplemented,
i just block directly the upload url in my dns server, and then u also got no data, then u are at the same point as at the momentā¦ data needed but not providedā¦ as security features block it.
The only one who has more unnessescary work with it, is either the user oder the admin,
but you do not reach your target
No data collected about browsers. Is that for a future version?
Manjaro Team has not yet answered questions about data collected that is not visible
How many users change the system to use Pulseaudio instead of Pipewire?
How many use JACK or just ALSA and will not be included?
2 Likes
I have been watching this topic over the past few days, and have come to the firm conclusion that there are only 2 realistic choices for Manjaro to make: abandon the Data Donor proposal, or make it opt-in. Opt-out is not an option if Manjaro is to remain a highly-rated, popular distro.
In some ways, this is a self-inflicted wound by Manjaro. This topic should have been posted in the private Member Lounge to get initial feedback from experienced members. Making the proposal public from the start was a big roll of the dice.
It doesnāt take much for a top-rated Linux distro to become a once-was-good distro. Manjaro is a very, very good distro. Well maintained, user-friendly (at least compared to other Arch-based distros) and, very importantly, with great support on this forum.
My fear is that a not-insignificant proportion of the members who provide much of the support on this forum could decide to change to another distro if mdd opt-out is adopted. If activity on the forum lessens, and support requests take longer to be answered, or remain unanswered, then other users may start looking around for an alternative which has a more active support forum. And the popularity of Manjaro will subsequently decline.
The fact is that many people choose Linux because it offers better privacy than the proprietary operating systems. And if they perceive that Manjaro wants to reduce their privacy, that could very well be the deal-breaker. It really doesnāt matter that the info collected by mdd does not contain identifying data. It is the perception that the data could be used for tracking users or machines that matters, irrespective of whether that perception is right or wrong.
Edit: I should add that I donāt have an issue with using the Manjaro Data Donor & have already sent my mini-PCās data.
12 Likes
Better you list as numbered list, to spec answer to num, see if u ask you can get the data 
- ???, do not know
- no, i have no usecase for this, old kernel perhaps have vulās
- minimal, as wayland is the future way to go (based on internal workflow)
- ???, me 0 at moment, but rpi 4/5 would be perhaps interesting
- ???, me +9
- ???, me ā rtx 30xx series
- ???, me, mostly pipewire at moment
- ???, me, but not tomorrow, i think of next year as i need stability on systems
- ???, depends on details (could be good, could be bad)
If theyāre hashed, the high probability of non-collision would still be enough to argue that it can identify someone. Even more so with dynamic IPs. Although that could depend on the hash algorithm. The question is, why store the IP in the first place?
Is the btrfs count on root partition wrong?
A hash of an IP does not help at all. All available IPs are easily enumerable and the hashes are not hard to check.
1 Like