Let's stop making privacy a techie-only thing, though. This should be a question a user chooses the first time they boot the browser, and Firefox should do its hardest to honor it.

Telemetry isn't inherently bad or privacy violating.

I suspect that for every scenario you can think of someone will be able to reply with solid logic about how it could be used in a way that’s bad and/or privacy-violating.

But is though IF it's not up front and center shown to the user. All telemetry can be backtracked to an IP in most cases. That's a bad thing. It's a form of history. It should be a switch that turns ON so that you OPT IN. Not the reverse, nor should it involve having to change a hosts entry.

> Telemetry isn't inherently bad or privacy violating.

How can you tell?

In Firefox:

point your url bar to about:telemetry

It shows you all the data that has been gathered. (Though IIRC it might still show stuff even when you've disabled telemetry -- in that case the data is being aggregated locally but not sent.)

Go to https://telemetry.mozilla.org

To look at the data on the server side. There are more sophisticated ways of querying it, but obviously not everybody can just be handed access to run arbitrary analysis code.

Probe dictionaries:

https://searchfox.org/mozilla-central/source/toolkit/compone...

https://searchfox.org/mozilla-central/source/toolkit/compone...

https://searchfox.org/mozilla-central/source/toolkit/compone...

Just checked about:telemetry, it seems most data are about CPU and memory stats, nothing related to browsing history.

Mozilla doesn't collect browsing history, period.

You forgot safe browsing.

What about it?

How you probably think it works in Firefox and how it actually works in Firefox are two different things, by the way:

https://feeding.cloud.geek.nz/posts/how-safe-browsing-works-...

As soon as i start firefox on windows it connnects to 1e100.net and makes a lot of connection to this domain even when no page is displayed (just the start screen). This is another reason why i'm in the process of switching from Firefox to another browser. A program shall not initiate network connections if not asked to.

It's open-source software.

That means nothing. Go read the audacity threads pertaining to trying to sneak in Google and some Russian analytics clients.

From the latest Audacity release notes:

>A journaling feature has been added for QA purposes.

What exactly does that mean?

https://github.com/audacity/audacity/releases

That's not what I meant. Search the PRs for the most commented upon PR. It'll invariably be the one about adding Google analytics and telemetry to Audacity.

What is privacy really? The browser? The ISP gets your data, the site gets metrics, and VPNs will just redirect traffic.

Having control over which information about yourself, your life, your habits, etc. you choose to give out, when, for how long, and to whom.

It's not this convoluted. Give people their privacy.

I’m not anti privacy, it just means something different to everyone. If you don’t control the metadata for servers you use, does that mean you cannot ever expect privacy?

I argued with someone about how we shouldn’t expect privacy with a cell phone, I don’t see how a 24/7 location tracking device that reads data telemetry (even if you use a custom privacy rom you connect to the towers and send telemetry) can ever be seen as private. PCs are usually compromised with Intel ME or AMD ST (unless you have a specially modded one or use old government contracted ones with them disabled).

Not having to use servers at all is kind of a preference. Why do browsers need to phone home at all?

When you use a site, you’re using servers and they have logs. This is unrelated to the browser itself, although ideally it shouldn’t have telemetry. But even so the servers of sites you use will and you can’t control the data collection.

When I boot Firefox, that operation should not automatically connect to a server unless I have a default web page set.

Maybe I am opening a local file? Using a local network resource? On a limited/expensive connection and don't want it auto-updating some BS? Running the browser in kiosk mode as a glorified HTML/CSS/JS GUI for a local app (apparently a thing now)?

It's really apparent when I socks proxy into a remote site (via limited 2G) to view a few kB worth of HTML admin page and firefox starts autoupdating a bunch of remote resources.

> The ISP gets your data

Browsers know much more. Most channels are encrypted nowadays so ISP doesn't know what you talk about on internets.

I've been a Firefox advocate for years, but after the recent update to 94.0.1 I removed it from my Mac, Linux, and Windows machines because of how it obscures multi-account containers - they are now a hidden icon that you can't click in the toolbar, and it's been moved into the address bar where its functionality has changed. This disrupted my workflow significantly.

I just switched the Librewolf, and installed the plugin there. It's working as expected, and I won't be forced (or incessantly nagged) to update the browser as I was with Firefox.

Thank you for your work on a great browser, but if you have any pull at Mozilla please tell them they're losing core users for good with these forced, unannounced, irrevocable changes to how users work in a browser.

Of the 180 lists we track at RethinkDNS, all the top ones contain *.telemetry.mozilla.org https://rethinkdns.com/search?q=telemetry.mozilla.org

That said, on Android, I don't see a single telemetry.mozilla.org entry in my DNS query logs.

Hi what exactly is your offer with RethinkDNS vs lets say, nextdns?

Hi: NextDNS is the gold-standard and many capable alternatives have cropped up in its wake in the recent times like windscribe's controld.com but RethinkDNS is not it.

Our primary objective is anti-censorship, and so, we make it simple for folks to deploy their own DoH (which means all our code is open source). Today, the serverless DoH stub resolver deploys to cloudflare workers but we plan to support deno.com/deploy and fly.io soon (already have it working in our development branches but the incompatibility between node and deno is causing a bit of pain to merge with mainline). RethinkDNS is the reference serverless DoH deployment, if you will.

We support DNS-based content-blocking too (out of sheer need for it from our users) but believe the right avenue to block content is either in-app where possible (like uBlockOrigin in Firefox) or at the IP-layer (not DNS), and hence we also build a Firewall (due to limited time, the RethinkDNS + Firewall app is Android-only, for the time being [1]).

That said, we do plan to achieve parity with NextDNS, but that's going to happen over a period of time since our team of 3 is stretched thin between building the firewall (too much complexity!) and the dns (too much work!).

[0] https://github.com/serverless-dns/serverless-dns

[1] https://github.com/celzero/rethink-app

How valuable is the telemetry data to Mozilla?

Based on their recent design changes (deprecation of compact mode, for example), they are either not collecting enough telemetry about the affected parts of the UI/UX, or they are ignoring what they have collected for whatever reason. Of course, there is a chance that telemetry confirms their vision, but based on the explicit feedback I've been seeing online, I doubt the rationality of their decision-making at least part of the time.

It is also possible that the telemetry shows that the vocal majority you perceive is, in fact, a vocal minority. I don’t have any more knowledge than you about whether that’s the case or not, but the possibility of being in the minority (and perhaps severely so) is absent from your list, and that deserves correction.

While I agree with you and the way you’ve stated it, it should be widely known that tiny groups of highly technical people can unlock huge understandings about how to improve.

Look at the speedrunning community for example: sometimes it’s not just a tiny group, but a single person that sees something that the devs did not, and that can lead to fixing “wtf” bugs for everyone else.

There is a chasm of difference between "We've identified something you have not given consideration to" as you describe (which I wholeheartedly support), and "We will insist we're right about the same points over and over again in every post about Firefox for all time" as HN behaves (which I do not, no matter what the topic). The former is a valuable and essential component of public issue reporting; the latter is not.

I see Firefox doing user hostile activity all the time, so I’m more inclined to think that are using telemetry in the slowly boiling frog analogy. The extension download data even without telemetry shows people want multi row tabs but they’re not putting it in by default. I am not sure how much the data is being used to improve user experience.

Which extension do you mean?

Multi row tab.

How do you compare strong voices of a few on a site like HackerNews or Reddit against many many many millions of data points of users around the world.

Should written feedback overrule a bigger data set?

A technically competent user's feedback should be weighted at least one, if not tens or hundreds of orders of magnitude greater than anonymously gathered telemetry. Remove actual, intentional human communication from the loop and you're lost at sea - anyone can make the anonymous data mean anything, and then it's whomever can make the cleverest chart or analysis of the data that ends up directing decisions.

Ignoring the people who actually take the time to communicate problems in favor of interpreted telemetry is exactly why Firefox is losing. Taking direction from technical users, or so-called power users, can give the application improvements in nuanced and technical uses. Taking direction from anonymous "averages" makes development a race to the bottom.

Firefox developments over the last couple years feels like what would happen if you put grandma in charge of trying to make things better. To put it bluntly, fuck grandma, she doesn't know what the hell she's doing anyway. Firefox used to be a Lamborghini, it doesn't need training wheels and balloon bumpers. Lean into technical excellence and drop the obsessive ui/ux nonsense.

I agree with this so much.

Once it was clear that Chrome was destined to cater to the masses, Firefox should have done a hard pivot with an emphasis on privacy and putting the user in control of their browsing experience. The best time to do this was a few years back when all of this was becoming obvious, but now with massive popular distrust of large tech companies like Facebook and Google making daily news, the second best time is now.

There's a reason I run Linux and BSD on every computing device I own, instead of Windows or Mac. It's not because it's easier to use (they are not), it is not because it has more bells and whistles (they do not). It's because at the end of the day, _I_ am the one in control of my computers, not some product manager who needs bullet points on his or her annual review.

There is no universe in which Firefox is going to successfully compete against Google at their own game, especially when Google is _still_ the majority source of their funding. I have no evidence for or against this, but my greatest fear is that the people at Mozilla who were passionate about the same things that I am passionate about have left out of frustration and the only ones left are there for the lifestyle and hipsterness of "working in tech" at a non-profit in a trendy city.

I cannot agree with it or your comment. I have seen these type of comments pop up in every Firefox thread on HN. It is so common for people to try to play armchair CEO. But when I actually dig into it, I have really never seen anyone with a competent vision of what a "competing browser" is supposed to look like. It all seems to boil down to "put the user in control of X feature and add a bunch of settings for it" or "don't remove Y feature that I used" or "bring back XUL" or something like that which I hope you can understand are not reasonable high-level directions. The various forks of firefox are well intentioned but these are all minor modifications, they don't try to do something different.

To illustrate what I mean here, if you want a fork with an emphasis on privacy you can just use LibreWolf. It is the entire thread we are responding to, the thing you want exists right now. But I don't see people exactly flocking to use that, your comment seems to not even acknowledge that it exists!

> Once it was clear that Chrome was destined to cater to the masses, Firefox should have done a hard pivot with an emphasis on privacy and putting the user in control of their browsing experience.

Why do you believe that catering to the masses implies not focusing on privacy and putting the user in control of their browsing experience?

Is there a guiding philosophy behind any of Firefox's decisions?

If it isn't customizable by the type of users who care about customization, then what is the reason to use Firefox instead of what ships with your OS or Chrome. Why would "typical users" have chosen Firefox in the first place without some vocal user suggesting it?

I still use Firefox for everything, I'm just sad that the lack of inspiration in the project means that it might not be a viable option in a few years. Maybe they're aiming for making 98% of users happy, and matching 98% of the features of other browsers, but it needs to have some reason to exist. Usability testing without innovation is that different from p-hacking without hypotheses in science.

Anyway, compact density was a non-default option, so it's difficult to understand why the option had to be deprecated. Compare that to MacOS. I didn't upgrade to Big Sur until Apple restored the option `NSWindowShouldDragOnGesture`, which allows you to drag a window from any pixel when you hold down control-option-command. Out of a billion users, I'd be surprised if more than 5,000 users care about this feature. (ie >99.9995% probably don't care) I only use the feature in combination with Karabiner Elements to change the command and Steermouse to recognize mouse button chords, but I invoke the command every couple minutes. Nevertheless it was restored, and it never disappeared in Monterey. Is it the budget alone that allows Apple to be simultaneously opinionated in their UI design and user accommodating, or is it completely different attitudes about users?

As someone who is currently implementing a compact theme for a different app, anything of that nature has a non-zero cost. Which is compounded by the number of designers you have working on the product who now all have to review every change multiple times.

While I broadly agree: data lies too. "X% more people use this now" says absolutely nothing about if they like it, or if it's doing what they want, or if you'll drive them off the system in a couple months because of it. It just says that X% more used it during the time you were watching.

You can use nothing but positive data-driven results to drive yourself out of existence, and it's rather easy. Direct, human feedback is absolutely essential.

If the telemetry-based removal of a feature would turn out to be a dealbreaker for a critical mass of users, it should be reconsidered (think of Mozilla's position in the browser market nowadays: it can't afford to piss off the "power users" and evangelists of Firefox).

And it's not like Firefox has no Nightly or Beta branch to test the waters before making a significant change. For example, during the prerelease phase of the so-called Proton UI, there was no shortage of clear feedback about it. A lot of it was legitimate criticism about accessibility (harder to distinguish inactive horizontal tabs because the separators were removed; part of the new palette did not have enough contrast; etc.) and usability (e.g. in cases of low screen estate, some menus were suddenly so huge that they'd not fit within the height of the screen).

Mozilla is slowly fixing some of these issues, which is a good sign IMO, but also sticking to some other "deliberate design decisions" that still remain controversial. I largely do not believe in design-by-committee, by the way. However, I believe that all valid feedback should be evaluated and taken into consideration if it's critical.

At the rate Firefox's market share is currently declining, it seems unlikely Mozilla actually has a finger on the pulse of the wider "millions [of users] around the world."

Is there anyone who you believe does have a finger on that pulse? If so, why? What can we learn from them?

Unfortunately, a certain employee who was rightfully kicked out of Mozilla in 2014 made a few bets that paid off (users dislike ads, users are stupid and will love it if you offer them money, and they'll go crazy for security snake oil even if it actively harms them).

Firefox doesn't block ads because they're funded by Google and they have ads in their own browser. Nobody likes ads. Especially not in the browser. This contributes to the decline. Firefox doesn't advertise with security snake oil, and they beg for money a lot, which is the exact opposite of the model proposed by the only browser that's rapidly shot up in market share in the last decade rather than roughly stagnating.

Some vocal voices are superusers that can affect others to use or not. Think of them as like current social media influencers. FF userbase has been in decline for sometime now during which those vocal feedbacks had been largely ignore. Perhaps, that give some proof indication "should written feedback overrule a bigger data set"?

Typically on the web more virulent, extreme content should always win against the quiet majority.

Is this sarcasm or sincere? I agree to a certain extent; informed people with good taste tend to be strongly opinionated.

Uninformed people with terrible taste generally seem to be even more strongly opinionated, though. I'm not sure we can draw many useful conclusions!

I'm not an expert on Mozilla's telemetry, but my recollection is the vast majority of telemetry data is performance data (e.g., how long does it take the program startup, how long does it take to query the history database when typing in the URL bar) or features usage (and this is more on the level of "which SSL cipher suites are being used" versus "who clicks this button in the UI").

Here is Mozilla's list of telemetry probes, including descriptions and whether it is recorded in prerelease (Firefox Nightly and Beta) or release versions.

https://probes.telemetry.mozilla.org/

Those are great examples and we do both.

We use the telemetry data as input to many product and business decisions. It is very important.

If that was entirely true, then better communication with users would probably be the ideal substitute for it.

Those are in no way substitutes for each other -- you have to do both. People are not able to self-report accurate measurement data, and telemetry data can't tell you anything about what a person wants or why they do things.

Product decisions I can kind of see, but what kind of business decisions are you talking about?