Yet another browser thread: Where do we go?

So let's set some foundations:
Chrome and Chromium are google botnet. That's obvious.
Brave has recently been discovered to be making major compromises despite their claimed focus on privacy. They're botnet.
bleepingcomputer.com/news/security/facebook-twitter-trackers-whitelisted-by-brave-browser/
Firefox has built-in telemetry. This can to some extent be disabled by the user, through the micromanagement of dozens upon dozens of about:config and user.js values.
Firefox is also notorious for slipping in stuff behind users' backs. Never forget this.
itsfoss.com/firefox-looking-glass-controversy/
ghacks.net/2017/10/06/mozilla-to-launch-firefox-cliqz-experiment-with-data-collecting/
It would seem that Moz://a doesn't have any concept of 'opt-in'. If they're willing to do this shit without telling users, what makes you think they won't silently revert your config changes and pass it off as one of their 'experiments'? Long story short, Firefox is botnet.
Safari, IE/Edge, Opera, and Vivaldi are proprietary.
[insert loonix webkitgtk/qtwebengine browser] is too barebones. Doesn't have any capacity for privacy/tracking protection besides occasionally a very basic adblocker.

With all of that being said, where do we go from here? What other browser options exist? What do you think of the above kikery?

Attached: browser-logos-30.jpg (465x475, 38K)

Other urls found in this thread:

gnu.org/software/gnuzilla/
github.com/muslayev/icecat-win64/
github.com/muslayev/iceweasel-win64
spyware.neocities.org/articles/browsers.html
dev.chromium.org/Home/chromium-security/client-identification-mechanisms
spyware.neocities.org/articles/http.html
spyware.neocities.org/guides/firefox.html
spyware.neocities.org/articles/iridium.html
support.mozilla.org/en-US/kb/lookingglass
twitter.com/SFWRedditVideos

Let's make our own web browser based on Gecko

What about ungoogled

Compile FF yourself, disable crap at ./configure time.

lol learn 2 code

>I'll make the logo

What happened to that moonman browser anyway?

How do you disable telemetry at compile time? What other modifications do you do during compiling?

Does it even matter anymore? The web itself is fucked, all Gecko based browsers will eventually die together with Firefox and Google will have a monopoly with their Blink engine (except on Apple products, which force the usage of Webkit).

Attached: 1538556085258.png (1280x720, 602K)

>No botnet
GNU IceCat (slow, sucks shit)
Ungoogled Chromium (chrome without botnet, no auto updates or google integration AT ALL)
TOR Browser (80% of the internet breaks, nobody can track you - unless you're retarded)
>I dont care about botnet
Might as well use the best version of the respective engines
Vivaldi (chromium based)
Firefox Nightly

A Moon Man browser? isn't that racist!!!!

Attached: moon luan 23904.png (490x417, 381K)

Check the patchset which essentially makes IceCat. Some can be done with configure flags, others need code removed.

Use ungoogled chromium, icecat, or palemoon. These are your best options for non text based browsers. I would say TOR too, but we all know TOR is a honeypot.

Never give up hope! Be optimistic!

Attached: 1549408911847.gif (549x598, 212K)

use Firefox if you’re not a brainlet, otherwise Iridium

>but we all know TOR is a honeypot.
explain. TOR browser is generally considered across the security community as being the best way to access the darkweb/use onion routing. More specifically, most will recommend you use Tails, which uses the standard TOR browser as its browsing option.

Also, I like your suggestion of Icecat. I definitely trust GNU more than Moz://a at this point to not screw up user privacy.

i use falkon but it krashes all the fucking time
waterfox with all the fingerprinting shit turned off is based enough for most usage
i may need to switch back to pale meme

ungoogled-chromium

>krashes
kringe

i have ungoogled chromium its pretty ok
you forgot to mention opera 12.8

>How do you disable telemetry at compile time?
You can prevent pings from being sent to Mozilla by adding
ac_add_options MOZ_TELEMETRY_REPORTING=0
to your mozconfig. I don't know how to disable the creation of pings though. The IceCat devs probably had to edit the source code for that.
>What other modifications do you do during compiling?
I disable a bunch of stuff like the updater, the crashreporter, WebRTC, EME, etc. and enable more extensive optimization (including PGO; never get LTO to work though).

>I don't know how to disable the creation of pings though. The IceCat devs probably had to edit the source code for that.
Scratch that. Apparently any information collection in about:telemetry gets disabled by setting toolkit.telemetry.unified to false, but only if toolkit.telemetry.enabled is also false (I use FF Beta for which it's locked to true; that's why I was confused).

Opera is closed source

Yeah but if you compile ungoogled-chromium from source how do you update it? Just check every few weeks if there's an update and recompile it?

>Just check every few weeks if there's an update and recompile it?
That's pretty much how it works, yes.

Sooo then you have to update extensions manually as well? Downloading and installing the CRX file everytime? jeez

No idea. I don't use Ungoogled-Chromium.

Seriously someone with the knowhow here just make one. Fully open source, transparent updates. If theres anything a tech retard can do to help ill do some dirty work.

There are many niche browsers like that out there. The problem is that they can't compete with the big browsers in terms of functionality (modern browser engines are massive projects), have no proper addon support (because it takes a big community to get that going) and stand out like a sore thumb when it comes to trackers.
If you don't mind these drawbacks then look into Dillo, NetSurf, Netrunner, etc.

Use GNU IceCat, a fully libre fork of Firefox with no telemetry.
gnu.org/software/gnuzilla/

It's even more botnet than Chrome itself since it can be fingerprinted easier. That's also why stuff like Brave, from day one on was bigger botnet than Chrome.

Icecat 60 is very good. Much better than icecat 52.

Congratulations by using Icecat or Librefox you fall for the typical brainlet filters. Their developers have no clue how fingerprinting works.

By using anything that isn't standard Firefox with resistFingerprinting enabled you inevitably give away your entire browsing history to Google and countless other advertisers and trackers.

That's also why Ungoogled Chromium is non ironically worse than Google Chrome for privacy. Imagine how many people run either browser worldwide.

I'm satisfied with Vivaldi, for now. Nice-looking browser, it's not closed-source, the authors have expressed disapproval against Google, and others.

hf being fingerprinted everywhere and instantly uniquely targeted by even the shittiest scripts worldwide. Either use stock Firefox and toggle the config yourself or just don't bother and use Chrome.

All tracking can be stopped by disabling images, css, and javascript.

>What is fingerprinting

Wrong. Imagine how many people worldwide do this. Should be easy to guess how easy it makes you be fingerprinted especially with some other vectors left open that do not need images,css or javascript.

The best way to avoid tracking is to non ironically load all ads and trackers as long as you have a non unique fingerprint

Wrong reply?

Nah, he's implying you shouldn't use anything but FF with privacy.resistFingerprinting enabled or Tor, because everything else makes you easy to track because of a unique browser fingerprint.

>no mention of cURL

Ah. Actually, I did thought about this matter before. I'm using an extension called Canvas Fingerprint Defender. It adopts a different fake value each time it does its job, instead simply blocking. I'm satisfied with it.

I know Vivaldi and other Chromium-based browsers are more on spying than Firefox, but I really like it and its unique visual and structure. Anyway, I will switch to this, if Vivaldi gets affected by the effect that shutdown uBlock Origin.
github.com/muslayev/icecat-win64/

Is there a Mac binary anywhere?

Install Vivaldi.

Attached: tetzchner.jpg (2524x3599, 530K)

Attached: uc-browser.jpg (640x427, 33K)

Furrymoon is just so much more economic with resource use that I can't drop it.

Both Icecat and your mentioned extension makes you more unique. As already said either use Tor or Firefox with resistFingerprinting enabled or use Google Chrome itself. There is no in between.

>Both Icecat
Actually, I meant to link this. Sorry.
github.com/muslayev/iceweasel-win64

>your mentioned extension
Explain?

>fingerprinting is too easy
>many websites require javascript
>by enabling javascript you lose privacy even further
>it's possible to track via. css as well
>ad blocking isn't enough anymore
>isp's get your dns requests, they know which domains you visit
>google working on recaptcha v3 where website owners send your data to them during server request. you Jow Forums google. you will never even interface with google anymore, they will get their data directly from the server.
>google probably working internally on AI to track users more reliably with even less data

it's over boys.
it was fun while it lasted.
the only thing that can save the web is a complete overhaul from ground up, scrapping the entire ecosystem as it is (html, css, javascript).

the content rich web was a mistake.

We IPFS now?

Attached: 1473474704102.png (756x715, 889K)

Let's just use Gopher.

>FF with privacy.resistFingerprinting enabled
>Firefox with resistFingerprinting enabled
>yes, use an actual botnet browser with telemetry up the fucking ass that phones home to CIAniggers constantly and is shady as hell, all in the name of avoiding """fingerprinting"""
>just submit, goyim
Seriously, fuck off. If this is about user agent fuckery, it's called an extension. You can use extensions to change the user agent to whatever normie shit you want, while still having actual, legit privacy. Don't tell users to install a fucking unethical spyware botnet browser just to maybe avoid one part of what can track you. Did you not read the fucking OP?

spyware.neocities.org/articles/browsers.html
all you need to know

Attached: 1474622657441 001.jpg (1000x1000, 148K)

dev.chromium.org/Home/chromium-security/client-identification-mechanisms

IPFS would be even worse. Imagine companies like Cloudflare/Google/Spywareinc./etc. all pinning IPFS sites. They'd know exactly which sites you are using even without serving ads or trackers, and not only them, but everyone who is interested.
We need something that's:
a. not web based
b. very limited in what it can send about your machine, i.e nothing (why the fuck is it even possible for someone to get your list of fonts and send it to them via. javascript?)
c. still allow rich content without all that web garbage
d. able to replace most of the web and offer advantages over it (give companies / individuals good reasons to adopt it)
e. led by a committee that actually cares about privacy and their users and doesn't sell out like the web consortium
f. entirely declarative
g. extremely fast to render (html/css is fucking bloated and dated)

Let's say it as it is:
the web is a platform that allows you to easily download a client program to interface with a server via. js and display data via. html/css, nothing more.

I'm 100% certain there are better and more efficient ways to do this than use this abomination that the web has become.

what are you even proposing? the only thing i could think of that comes close to this would be using a dedicated program to render content that is obtained via ftp or irc or other non http/s means which i wouldnt even be opposed to but its not business friendly. there is never going to be something as feature and content rich as the modern web without the downsidses we have right now. wikipedia cant run on some NEETs hobby server
link for http problems:
spyware.neocities.org/articles/http.html

>Safari, IE/Edge, Opera, and Vivaldi are proprietary.
All the source code for Vivaldi is available though. Go look at it if you're so worried

>Not using Yandex in 2019

Imagine being worried about the Kremlin stockpiling your info rather than the U.S.

Yandex is a search engine, not a browser.

Is there a comprehensive guide on how to disable all telemetry and configure Firefox for maximum privacy?

Attached: 1546513399900.jpg (2048x1642, 572K)

Do you ever google things before making dumb statements like that?

Attached: 1538449491449.gif (222x330, 545K)

I'd go Vivaldi, but it has some shady shit going on.
Some files are not downloaded.
GitHub captcha is unsolvable and so on.

Therefore chromium is the only sane solution.

If you have pihole, however, I'd suggest vimb or the alternatives.

What you write makes sense but I'm also heavily inclined to believe that you're a shill

spyware.neocities.org/guides/firefox.html

Thank you very much for this. Taking the time to properly do this on a phone was tiring but necessary

Attached: Moe_portrait.png (334x454, 148K)

Nice, that's exactly what I was looking for

Fucking paranoid cucks. Use chrome

Unless your willing to ditch basically all modern web functions and use a bare bones browser everything is fucked

Revive netrunner.

I just decided that vimb is good enough

There are far more things to consider when it comes to fingerprinting than your user agent. Are you blocking cookies? Are you blocking Javascript? What system fonts can a website use? What system language are you using? Is your browser sending DNT requests? And so on.
Making your browser not stand out is difficult and IMO not worth the effort. Might as well stand out and use whatever browser you want. The web is already fucked.

Go through ghack's user.js. They are very extensive and have a good documentation. It's important to understand what you are tweaking instead of mindless copying.
You might also want to compile FF yourself.

I'm honestly considering it. Getting some bare bones browser and keeping one of the big ones as backup. It would probably improve my browsing habits as well.

Falkon ftw

What about Iridium?
spyware.neocities.org/articles/iridium.html

Attached: file.png (375x375, 9K)

Is it true that it hogs RAM like crazy?

>I definitely trust GNU
You should't trust GNU for anything, but OK.

I'm a brainlet. Can you explain what you mean by fingerprinting and why do you claim browsers like Icecat/Librefox/Ungoogled Chromium are the worst at it?

>You should't trust GNU for anything
why?

He's saying that sites can identify you by how unique your setup is. If you have certain config changes, extensions, user agents, or browser itself that are not the norm, the theory is that you are more 'unique', and sites can actually determine who you are based on that.

It's a kike scam though. Yes, fingerprinting can be a method of discovering who you are, but the proposed methods of decreasing fingerprintability are to dramatically decrease your privacy in other areas. "Just use TelemetryBrowser, user," "Just disable your privacy extensions, user," "Umm, Chrome is ackshually MORE private, user," "Just submit, goyim."
In other words, I don't buy it.

isn't ungoogled chromium reports itself as chrome?

Attached: Iceweasel_icon.png (2000x2000, 1.19M)

Isn't that a cancelled debian project? afaik they now just use 'firefox'.
Also, 'Iceweasel' is different from 'Icecat'

>History
>GNU IceCat was formerly known as GNU IceWeasel but changed its name in 2008 to avoid confusion with Debian IceWeasel (who was rebranded back to Firefox in 2017)

So in summary:
GNU wanted to make a fork because they wanted to make it a secure and private browser.
Debian wanted to make a fork because fuck knows why.
There was a bit of confusion because they both wanted to call it Iceweasel.
GNU agreed to call theirs Icecat.
So now there are 3 browsers: Firefox, Debian Iceweasel, and GNU Icecat.
Debian gave up Iceweasel a while ago.
So now there's 2 browsers: Firefox and Icecat.

Debian made a fork because they weren't allowed to use the official Firefox name and logo on their repos.
So they just changed those, there weren't any more modifications.

Exactly my point.

If you want just privacy then sure use your little fork
But if you need both security and privacy then you'll want to be up to date all the time asap
Mozilla are the one rolling out security updates to firefox first
Just remember how they rolled out a patch right the next day after meltdown and spectre came out
And regarding privacy did you forget fitefox is open source?
Sure you'll need to click some options, but only the FIRST TIME you run it after installing
>disable statistics sharing from the options
That should be enough, but if you're paranoid you can also
>about:config
>disable all telemetry
>disable all experiments
There, you just made your own little waterfox, but with instant security updates, now stop being a mindless cuck and stop posting that gay shit

Attached: d8xbrb7-5071ac8d-2add-4e7e-b071-6b9c4a2a76b5.jpg (610x900, 100K)

But you didn't address the real problems of firefox: Mozilla's evil, lie-by-omission practices. Again, if they're willing to do this shit without telling users, what makes you think they won't silently revert your config changes and pass it off as one of their 'experiments'? Long story short, Firefox is botnet.

aren't addons on FF a valid requisite to overriding some of those issues? i would imagine they automatically freeze the changes from happening, unless I'm mistaken.

>aren't addons on FF a valid requisite to overriding some of those issues?
There are some addons that allow you to quickly toggle certain tweaks without having to go to about:config, but they aren't necessary. It's easier to enforce tweaks by using a user.js. That'll revert all listed about:config entries to the specified value at startup.
The bigger problem is when Mozilla changes the role of those entries, deprecates them or ties them to some other setting that might change the expected behavior. That's nothing they do on a whim though (or even can; the process of deprecating toolkit.telemetry.enabled is still going after years).

>That's nothing they do on a whim though
>Moz://a doesn't randomly decide to fuck with users on a whim
oh you sweet summer child

Debian likes to apply their own patches and can’t use firefox branding with a modified source code, so they rebranded it as their own.

They may decide randomly, but you still get some heads-up either through a statement or a discussion on FF's Bugzilla.

Great, so now you have to constantly keep tabs on that, and also hope they don't modify their statements after the fact.
This is their statement now.
support.mozilla.org/en-US/kb/lookingglass
As you can see, they state that this is opt-in. IT WASN'T OPT-IN when it was first rolled out. People just ended up with a random extension installed on them without their consent. They make mention of that in a little sidenote, but that note inherently points out that the shield studies are opt-out. They are opt-out because they are enabled on a fresh Firefox install, and you have to manually disable them. Once again, something that you have to disable, and once again something I wouldn't put it past them to revert with an update at some point.

>They are opt-out because they are enabled on a fresh Firefox install, and you have to manually disable them. Once again, something that you have to disable, and once again something I wouldn't put it past them to revert with an update at some point.
That's what the user.js is for.

>it's possible to track via. css as well
Is nothing sacred? How long until they implement an image format which ties to your processor ID, MAC address, ISP domain and shit?

Attached: 0vq95j3c8rlz.jpg (750x736, 64K)

Falkon if you want to avoid Moz://a codebase completely, Icecat if you like Firefox but don't want botnet

Mhm. Even if you had the most generic setup, you could be identified by other factors.

That's my point. Having the 'unique' setup blocks those other identifying factors like telemetry, ads, third-party cookies, unnecessary tracking javascripts, etc. But doing that also apparently makes you more vulnerable to being 'fingerprinted'.
Overall, I think the best option would be to use a unique privacy-protecting setup, and then fake the other stuff through the use of a user-agent spoofer and/or other methods.

Bump

>krashes
jej