Updated Aug. 28, 2024. Take back your privacy Firefox is rolling out Total Cookie Protection by default to more Firefox users worldwide, making Firefox the
Advertisers track you with device fingerprinting and behaviour profiling now. Firefox doesn’t do much to obscure the more advanced methods of tracking.
Lots do. But do you know anyone that turns JS off anymore? Platforms don’t care if they miss the odd user for this - because almost no one will be missed.
“Anymore”? I’ve never met a single soul who knows this is even possible. I myself don’t even know how to do it if I wanted to.
I do use NoScript, which does this on a site-by-site basis, but even that is considered extremely niche. I’ve never met another NoScripter in the wild.
Roughly similar to using Adblock Plus with many filter lists + NoScript with 1st-party scripts/frames automatically trusted. Unlike NoScript however, you can easily point-and-click to block/allow scripts on a per-site basis.
The people who I’ve tried to get on NoScript seem to have the brain capacity of goldfish. If the site doesn’t instantly work, it’s as if the sky has fallen and there is no way to convince them to pay attention to which scripts are actually needed.
It’s a rare breed that is willing to put up with toggling different scripts on and off. I’ll also acknowledge that too many people (including me) are in a giant rush. For work-type stuff, I have the laptop without noscript, because sometimes I do need something to work absolutely right now.
People whose lives revolve around fashion probably think you dress like shit.
People who love food probably think you eat like shit.
People who love cars probably think you are a shit driver.
You probably love computers and care about privacy, and you are shitting on regular users(assumption, admittedly) for not being invested.
They had something that was working, you present noscript, thing no longer works. If you are not invested, how are you going to see the appeal of extra work?
Well, you know what they say. You can lead a horse to water, but you can’t make it interested in learning about the water cycle to have a deeper understanding of why the river flows in the first place.
I go hard with DNS-based ad blocking and I’m constantly confirming it works by checking the network tab in developer tools. I’m basically only seeing first party scripts and CDN assets — 99% of websites load all the tracking garbage from third-party domains that can be easily blocked.
Is it something you can talk about? I’m currently in the process of trying to switch from pihole to pfblockerng but am interested if there are better alternatives
At a high level it involves a terrible custom parser written in Ruby for several formats of DNS blocklists. It finds the proper domain then outputs a large configuration file for Unbound.
I’ve attempted to Dockerize it but honestly, I think it would be better to use a superior parser written in another language that can be statically compiled.
I was using Fly.io to host it in various regions using an Anycast IP, but since I’ve moved onto using VPN for everything I’ve moved it to a few hosts acting as Tailscale exit nodes. Those exit nodes provide the blocking DNS service along with rewriting incoming Tailscale client traffic to route out of another network interface assigned to a VPN provider.
Had I unlimited free time I’d rewrite the parser in Crystal, but part of me thinks there’s got to be something already written by someone in Go.
uBlock origin + NoScript for me. I deal with the bigger umbrella of scripts with uBlock and then fine tune permissions to the ones that uBlock allowed with NoScript.
They might be fingerprinting me using these two extensions though.
Not all but most, yes. But TBF, sites that still function with JS disabled tend to have the least intrusive telemetry, and might pre-date big data altogether.
Regardless, unless the extent of a page’s analytics is a “you are the #th visitor” counter, all countermeasures must remain active.
It’s really strange how they specifically mention HTML5 canvas when you can run any fingerprinter test on the internet and see that Firefox does nothing to obfuscate that. You can run a test in Incognito mode, start a new session on a VPN, run another test, and on Firefox your fingerprint will be identical.
Well yeah, they’re just blocking known fingerprinting services. If you use a tool that they don’t recognize, it’ll still work, but their approach will still block the big companies that can do the most harm with that data.
The only alternative is probably to disable WebGL entirely, which isn’t a reasonable thing to do by default.
I wish Firefox had a per-site or per-domain preference for WebGL (as well as for wasm, etc), the same way we have per-site cookies or notifs preferences. It’d help clear most issues regarding this.
Honestly would be hard to do. There a perfectly legitimate and everyday uses for pretty much everything used in fingerprinting. Taking them away or obscuring them in one way or another would break so much.
Librewolf has Resist Fingerprinting which comes pretty far.
Every Librewolf browser uses the same windows user agent, etc. But there are downsides, like time zones don’t work, and sites don’t use dark mode by default.
And even then, EFF’s Cover Your Tracks site can still uniquely identify me, mainly through window size. That’s one of the reasons why Tor Browser uses letterboxing to make the window size consistent.
I don’t know what letterboxing is. But if window size is used to identify me, can’t it be circumvented simply by using the window in restored size, and not maximised?
Your restored window size is even more unique than your maximised window size!
The correct solution is to just not make the window size available to JS or to remotes at all. There’s no reason to ever need specifics on window size other than CSS media-queries, and those can be done via profiles.
Changing the source code of the browser, unfortunately. I don’t know what Tor Browser does or how, but basically you’d have to do about the same as they do.
Advertisers track you with device fingerprinting and behaviour profiling now. Firefox doesn’t do much to obscure the more advanced methods of tracking.
Don’t all the advanced ways rely on JavaScript?
Lots do. But do you know anyone that turns JS off anymore? Platforms don’t care if they miss the odd user for this - because almost no one will be missed.
“Anymore”? I’ve never met a single soul who knows this is even possible. I myself don’t even know how to do it if I wanted to.
I do use NoScript, which does this on a site-by-site basis, but even that is considered extremely niche. I’ve never met another NoScripter in the wild.
Am I in the wild? I use it.
Why not just use ublock medium mode?
https://github.com/gorhill/uBlock/wiki/Blocking-mode:-medium-mode
The people who I’ve tried to get on NoScript seem to have the brain capacity of goldfish. If the site doesn’t instantly work, it’s as if the sky has fallen and there is no way to convince them to pay attention to which scripts are actually needed.
It’s a rare breed that is willing to put up with toggling different scripts on and off. I’ll also acknowledge that too many people (including me) are in a giant rush. For work-type stuff, I have the laptop without noscript, because sometimes I do need something to work absolutely right now.
You don’t think you are being a tad judgemental?
People whose lives revolve around fashion probably think you dress like shit.
People who love food probably think you eat like shit.
People who love cars probably think you are a shit driver.
You probably love computers and care about privacy, and you are shitting on regular users(assumption, admittedly) for not being invested.
They had something that was working, you present noscript, thing no longer works. If you are not invested, how are you going to see the appeal of extra work?
Well, you know what they say. You can lead a horse to water, but you can’t make it interested in learning about the water cycle to have a deeper understanding of why the river flows in the first place.
I go hard with DNS-based ad blocking and I’m constantly confirming it works by checking the network tab in developer tools. I’m basically only seeing first party scripts and CDN assets — 99% of websites load all the tracking garbage from third-party domains that can be easily blocked.
Pihole?
It’s a common solution but I do something more involved and manual, but it’s the same concept.
Is it something you can talk about? I’m currently in the process of trying to switch from pihole to pfblockerng but am interested if there are better alternatives
At a high level it involves a terrible custom parser written in Ruby for several formats of DNS blocklists. It finds the proper domain then outputs a large configuration file for Unbound.
I’ve attempted to Dockerize it but honestly, I think it would be better to use a superior parser written in another language that can be statically compiled.
I was using Fly.io to host it in various regions using an Anycast IP, but since I’ve moved onto using VPN for everything I’ve moved it to a few hosts acting as Tailscale exit nodes. Those exit nodes provide the blocking DNS service along with rewriting incoming Tailscale client traffic to route out of another network interface assigned to a VPN provider.
Had I unlimited free time I’d rewrite the parser in Crystal, but part of me thinks there’s got to be something already written by someone in Go.
uBlock origin + NoScript for me. I deal with the bigger umbrella of scripts with uBlock and then fine tune permissions to the ones that uBlock allowed with NoScript.
They might be fingerprinting me using these two extensions though.
Not all but most, yes. But TBF, sites that still function with JS disabled tend to have the least intrusive telemetry, and might pre-date big data altogether.
Regardless, unless the extent of a page’s analytics is a “you are the #th visitor” counter, all countermeasures must remain active.
It’s not much but they try a whole lot more than most.
It’s really strange how they specifically mention HTML5 canvas when you can run any fingerprinter test on the internet and see that Firefox does nothing to obfuscate that. You can run a test in Incognito mode, start a new session on a VPN, run another test, and on Firefox your fingerprint will be identical.
Well yeah, they’re just blocking known fingerprinting services. If you use a tool that they don’t recognize, it’ll still work, but their approach will still block the big companies that can do the most harm with that data.
The only alternative is probably to disable WebGL entirely, which isn’t a reasonable thing to do by default.
I wish Firefox had a per-site or per-domain preference for WebGL (as well as for wasm, etc), the same way we have per-site cookies or notifs preferences. It’d help clear most issues regarding this.
Honestly would be hard to do. There a perfectly legitimate and everyday uses for pretty much everything used in fingerprinting. Taking them away or obscuring them in one way or another would break so much.
Librewolf has Resist Fingerprinting which comes pretty far.
Every Librewolf browser uses the same windows user agent, etc. But there are downsides, like time zones don’t work, and sites don’t use dark mode by default.
And even then, EFF’s Cover Your Tracks site can still uniquely identify me, mainly through window size. That’s one of the reasons why Tor Browser uses letterboxing to make the window size consistent.
Librewolf supports letterboxing as well, though the setting might be disabled by default
I don’t know what letterboxing is. But if window size is used to identify me, can’t it be circumvented simply by using the window in restored size, and not maximised?
Your restored window size is even more unique than your maximised window size!
The correct solution is to just not make the window size available to JS or to remotes at all. There’s no reason to ever need specifics on window size other than CSS media-queries, and those can be done via profiles.
But the restored size keeps changing - can’t be profiled, right?
And how do I not make the size available “to JS or to remote”?
Changing the source code of the browser, unfortunately. I don’t know what Tor Browser does or how, but basically you’d have to do about the same as they do.