> A Google spokesperson has been in touch to say the issue is a programming error, and will be fixed: "We are aware of a bug in Chrome that is impacting how cookies are cleared on some first-party Google websites. We are investigating the issue, and plan to roll out a fix in the coming days."
It's really hard to give Google the benefit of the doubt at this point.
The source is public and the version history too, so you don't need to assume, you can look. If you don't see anything particular, you may assume that it's a bug and wasn't easily visible to the developers either.
When I looked at the source, around 2015, there wasn't much difference, other than the name. People called "Google" were allowed to name the result of building it "Chrome". If you weren't Google you had to call it Chromium, and could of course apply patches or change the build flags. The functional differences I saw then were all a result of patches that were applied by my distribution.
Where you'd assume clearing Cookies to be as simple as cookieStore.clear() - I can imagine it's due to some legacy remainder of times when they did do everything to track without consent.
Google acting like this is a programming error. Any programmer would know this would have to be an edge case, perfect-storm type of bug for it to just magically decide google.com and youtube.com are not included in the wipe.
I call bullshit. Google is an evil company, this is nothing new.
As someone who works on Chrome at google and know people working on fixing it.... it really is kinda a perfect storm bug but you obviously don't have to believe me.
I could be convinced, but given Google's behavior they've lost the presumption of good faith, and the bar of evidence required to convince me that this is an accident is much higher than it would be normally.
Well...they already treat Google and Youtube as special, because logging into them also logs you into the browser as well. It wouldn't surprise me if that special-ness had unintended consequences and that's the bug.
It /also/ wouldn't surprise me if it's exactly the way you describe.
> Circumventing otherwise strict security rules for their own domains is shady in its own right.
In most fights between adding a user feature and protecting user privacy, if the feature cannot be added otherwise, feature wins. It's only "shady" if you're already assuming the feature vendor is a hostile actor.
The consistent Chrome experience enabled by logging the browser in is hells useful (especially in this era of mobile device + desktop device; no more having to remember if I bookmarked that article on my phone or my desktop or my work PC, etc.).
The easiest way to protect user privacy is to never implement any features. Can't hack what isn't there.
Once we assume that's not the decision that'll be made (because obviously), the question becomes whether the feature can be implemented at all while respecting user privacy.
And if the answer is yes, we can expect it to happen.
>> The easiest way to protect user privacy is to never implement any features.
> I don't subscribe to this school of thought, personally.
What school of thought do you subscribe to? I had assumed that this
> /Not/ protecting user privacy is always shady
... being an absolutist statement, meant that you were open to the easiest way of protecting user privacy, which is "Don't increase the security surface area at all," i.e. add no new features.
On the other hand, a very large majority of the time that a programmer speculates on the internet about something that's obviously true about code I've written or am otherwise privy to, they're wrong
On the other other hand, most vague and fabricated statistics about anecdotal, speculative or subjective topics tend to have very little weight when making a point...
Too true. I guess my point is that complex products tend to have complex interdependencies that one isn't likely to guess exist from the outside. Things that look clearly intended can easily be programming errors, things that seem like they should be easy can be hard. I think the reverse can be true too but not as often.
So I don't think the thought that any programmer would know this is a ridiculous edge case is one that holds any real weight here.
The commit that did this ~4 years ago was very much in the open, and had rationale in the commit message.
I think it's pretty sad that nobody here has linked to it... It's as if the people who care about this stuff aren't the same ones who write and review code for a living...
Having seen the source code for web browsers, I'm 100% willing to believe it's a programming error.
The likeliest point of introduction is where Google added account login to the browser itself. At that point, either ambiguity of specification or failure to test as though the user thinks Google is a privacy-hostile agency (while using a browser that allows them to login to Google) could have resulted in someone adding a special case to refrain from throwing the credentials away that keep the browser logged in that inadvertantly failed to throw the credentials away that keep a session logged in.
We will need laws to forbid providers of web-based commercial services from making browsers. Google has made it abundantly obvious that control over both sectors is a net negative for the market.
In the same way, browser makers should be decoupled from hardware and OS manufacturers, which have a similar monopoly problem.
Slowing down development of the browser will actually turn out to be a Good Thing. The breakneck speed that has seen Google pushing more and more features has effectively destroyed the ecosystem of alternatives, since nobody could keep up. And after all this, is the web so massively better than it was in 2005? Not really. We just download more and more bloated polyfill libraries.
KHTML was the basis for the original webkit, and that was built basically for free.
The mainstream doesn’t even know archive.org exists. But they are all on youtube, and Google will make sure they keep doing that, “optimizing the experience” of the browser to lead them there.
Mozilla can easily divest from those activities.
Everybody else: yes, cutting them off is the point. To allow for real competition, browsers must be agnostic at both ends. The current field has been entirely captured by giants for anticompetitive purposes.
The law should be crafted in a way that allows for commercial agreements, as long as they are between equal parties, non-exclusive in nature, and time-limited if they involve default options. This way, browser companies will not be directly dependent from any single web or OS company.
Meh. Mozilla couldn’t be dominant if they were the only browser-focused company left on the planet, they are just terrible at steering anything and utterly wasteful. Before the search deals, they couldn’t even dominate the small backyard that is the Linux desktop. By day 2, you would have several different competitors with better tech and faster development cycles busy eating their lunch.
I'll use this as an opportunity to ask: What am I missing, having ditched Chrome ~3 years ago and switched to Firefox? I am very happy with Firefox, and it fulfills 100% of my requirements for a modern browser (must have sane browser dev panel, must perform with >50 tabs open, must be extensible with a huge extension ecosystem).
On my Android phone, I'll just go ahead and claim I'm not missing anything and it's not even close - Firefox is so much better just for being able to run extensions like uBlock Origin.
But on a workstation (macOS) - I'll keep an open mind.
I still use Chrome by default on mobile, but when Chrome has performance issues, I switch to Firefox and it works fine. I've been considering a switch to Firefox on mobile (I already switched on my main workstation).
Try it with a long page, something thousands of lines long. Mobile Chrome on Android chokes badly, stops responding for seconds or minutes, and page search is hideously slow. Firefox handles it fine. It's really obvious on long lists or tables, e.g. this page:
Why is Chrome your default on mobile? The web has become nearly unbrowseable on mobile without an AdBlocker and that's the single greatest advantage Firefox has over chrome on mobile.
In my experience, on many (most?) sites, ads are by far the biggest performance issue, dwarfing any differences in rendering. I can remove those with FF but not Chrome, so they made that decision quite easy for me.
I switched to Firefox for regular browsing, but use Chrome for development because their devtools is so much better. I hope Firefox invests more in devtools development, honestly. More devs developing on FF could lead to higher adoption.
That's very interesting. Personally, I much prefer the Firefox development tools. Mind if I ask what you like about the Chrome ones, compared to Firefox'?
my primary is Firefox, but i definitely prefer Chrome's devtools - especially for performance profiling, which i do a lot of.
chrome also renders UI interactions faster. maybe it's skia vs cairo? e.g. try mouse-dragging the ranging slider and compare the lag between chrome and ff:
I also switched years ago and haven't looked back. FF is fast, solid. However, I too keep Chromium for development. But I use the Ungoogled version (I always plug it, but have no stakes in it): https://github.com/Eloston/ungoogled-chromium
Does anyone know a Webkit browser that runs on Win - preferably with some kind of debugger?
The only thing I notice about Firefox is that some sites don't seem to support it very well. For example, I can't log in to U.S. Bank with Firefox, or successfully place an order on Doordash, but both sites work fine in Chrome.
Until recently, I had thought that browser compatibility issues were not much of a thing anymore.
It's still not worth running Chrome all the time. I only use it to complete those discrete tasks, then it's back to Firefox.
Not sure anyone else's experience but every time I bring up supporting browsers other than Chrome I just get sarcasm at best and belligerence at worst from 90% of engineers.
It's usually down to the remaining 10% to pick up the rest of the pieces.
It's extremely frustrating to hear this attitude as someone who was doing front-end dev in the IE6 era.
That's about the right ratio to expect, since less than 10% of browser sessions appear to be Firefox on average.
The major difference between this era and the IE6 era is that Microsoft wasn't willing to play ball with standards. With few exceptions, Mozilla and Google are implementing to the same standards and specifications; problem is, Chrome's a better implementation, for speed and stability, in the average case and the corner case.
They have both more resources to throw at the problem and more access to data to proactively discover issues.
It is, unfortunately, a situation where the playing field is "fair" (in the sense that there's a uniform standard) and one player is just trouncing the alternative.
>> Mozilla and Google are implementing to the same standards and specifications; problem is, Chrome's a better implementation, for speed and stability, in the average case and the corner case.
This is not the interpretation I make of the current landscape. Google is SETTING standards that directly benefit themselves, just like MS did; Chrome uses it's market share to push in directions that benefit Google before they benefit end users.
I'm probably tilting at windmills, but I develop in Firefox, then spend the extra effort to make it work in Chrome for the few areas where behaviour is different. I tend to build stuff for smaller audiences of enterprise-oriented users though; not sure if this would be appropriate for massive consumer plays.
It may fall of deaf ears, but you should complain to the site owners that they are broken. If enough complaints come through (and others are likely to be complaining as most issues I see are not FF issues, the code is actually broken for everything except chrome), then they'll be forced to fix it, or get better developers.
It's not a bad idea, but to be realistic about the expectations, here's what happens.
1) site owner receives complaint
2) if offered, site owner looks up the user agent
3) site owner finds that the user agent matches one that is sub-10% of their traffic and samples the complaint signal accordingly
High-volume sites don't actually optimize for 100% of potential users (because the marginal value for one more user is low). If it's a low-volume site that happens to attract FF users for some reason, complaining is likelier to succeed.
I have noticed a number of times that when sites are broken, it seems to be that anti-tracking settings, which are stronger by default on Firefox are a factor. I suspect the involvement of cross-site tracking systems like recaptcha in ill-conceived attempts to reduce automated attacks.
Firefox (unless things have changed recently) still has limited addon capability -- no key remapping (or most other stuff) until the current tab has loaded. You also can't do unsigned addons (unless you do dev mode which is a separate install that doesn't update), while Chrome allows them by toggling "dev mode".
I'm unsure what you're referring to. I use Firefox Nightly, but your prior comment implies the existence of a "dev mode which is a separate install that doesn't update"; Firefox Nightly updates just fine, and it's not a separate install, it's the only Firefox on my machine. It's perfectly stable, I haven't had a crash in years. And I didn't need to change any settings to install my add-on, I just clicked the "install from disk" option in the add-ons menu.
I mean it’s a separate install from mainline, not something you can just toggle. Plus it requires you to use an unrecommended, bleeding edge version that’s mainly for testing.
Does Firefox work on all the sites you visit, or do you have to keep a "side-piece" browser to do things like log into your bank account or use Google Docs?
Usually the culprit is an extension, and the fix is a private tab (where most of my extensions are disabled).
But no, I use Google Docs and log in to several bank accounts in FF without issues. I keep the other browsers just for compatibility checks for front-end development.
You're definitely gaining even more instability than with Chrome. And by that, I mean, insane random changes.
Just look at, for example, the recent madness with Firefox on Android. They effectively pushed a beta about 6 weeks ago, removed tonnes of features, made many menu items difficult to navigate to, removed compatibility with almost all add-ons, it's slower, it crashes for many constantly, it's just a plain, simple mess.
They had 4.5 stars at one point. They now have 3.9 stars:
Do you know how hard you have to work, to drag something down from 4.5, to 3.9, in a month, with almost 4M users?
I honestly don't think they could do worse, if it was intentional sabotage by some hostile insider.
And it's endless garbage like this from Firefox.
What firefox sees:
"OMG! Chrome is gaining ground! Quick, turf everything that is unique on our platform, that is different from Chrome. We must be just like them"
What outsiders see:
"Great, everything I cared about, unique to firefox is now gone. Plus it's slower, buggier, and they keep changing everything around on every single release. Great. Time to move on."
You don't compete with the giant, by being the giant. You etch out a unique market.
Meanwhile, Firefox did the precise opposite, and removed their uniqueness.
I've had no issues with stability or usability with new Android FF. Certainly not slower or buggier in the least, so clearly YMMV.
It's mildly annoying to re-learn the UI, but Chrome did the exact same thing to me years back when they, for example, dropped viewing every tab as a separate app in the running apps list. I refused to update for a year after that, very annoying. They all do this, I've resigned to it.
They each have their ups and downs, I've uses both on many platforms for many years for now. Generally what's an up and down depends on the person as well e.g. some consider being on the browser with the tiny usage share and hitting compatibility issues a good thing ("it helps the web") and others consider it bad ("I just want it to work"). Same thing with performance, performs how well for which sites in which use case on which system. In general, with the exception of "the browser doesn't have hardware acceleration for the platform I use", performance is probably imperceptibly different for >95% of users but each tail will be very adamant that their tail is the better performing one (and it may be for their use case).
For me personally the biggest current difference remains HDR support, which FF lacks and Safari/Chrome are continuing to grow. FF still has a big ways to go to achieve HDR so I think this one will be around for a while unfortunately. Now of course many are going to think "I don't give a shit about HDR" in which case refer to the first paragraph.
The biggest difference that finally got fixed was for the longest time Chrome had much higher quality pitch correction on sped up content e.g. YouTube 2x. It was noticeably worse at 2x and I couldn't even understand the content at 3x on Firefox but it sounded great in Chrome. This forced a lot of my usage over to Chrome until it was fixed.
On the Chrome side the biggest drawbacks are lack of extensions on mobile and lack of containers. I don't use containers as a privacy tool but more as a multiple profile tool without having to worry about which profile a window was launched in but I don't think that's the main real world use case happening. Lack of extensions on mobile is obvious.
I love Chrome's tabs and have yet to find a browser that compares.
> must perform with >50 tabs open
Testing Firefox now, it has the same behavior as Safari: when I open 16 tabs it starts a horizontal scrolling list. Absolutely unusable. Chrome shrinks every tab down to Favicon size or smaller and shows you 50+ tabs at once.
Chrome got tabs right a decade ago and has yet to mess them up, as they have with most of the rest of the browser. Very hard to switch until they do.
I usually group tabs in multiple windows for that reason, but as usual (and equally with FF and Chrome), these kinds of annoyances can potentially be addressed with add-ons like Tree Style Tab.
i actually really hate the way chrome does tabs, it's the main reason I use FF. I don't care about tab size though or horizontal scrolling, for me the thing that bothers me about chrome is that it when middle clicking multiple tabs open, they just open somewhere in the tab list. and if I middle click to open another tab and then middle click a new link in that one, it gets pushed to a completely different spot entirely.
on FF, tabs open next to the current tab, unless you previously opened a tab in this one, then it'll go to the end of the list of tabs you've opened from the current window. but as soon as you switch tabs once, it'll start placing them next to the new tab you're in again. chrome just does something entirely unpredictable, instead of "resetting your spot".
but I've found that most people don't really seem to browse the way I do/don't care about that
I remember caring about that as well back when Chrome was released, but since then have adapted to the Chrome style. For me it's a minor preference vs. the horizontal scrolling issue.
I have also made the switch. I do notice a few irritations that I struggle to remember off the top of my head, but a big one recently with the Android app is the lack of support for password managers (at least LastPass)
The two things that make me periodically think of switching from Firefox to Chrome on MacOS are (1) spell checking and (2) handling of multiple profiles.
Firefox uses the same open source spell checker, Hunspell, that Chrome, LibreOffice, MacOS, and many others use, but for some inexplicable reason have a terrible dictionary. I've ranted on this before [1] so won't do so again here.
I get the impression that the poor handling of multiple profiles is because they expect you to use multi-account containers instead.
Those are fine if you just want, say, to access multiple instances of the same site without using the same cookies for each, or want to make it harder for cross site cookie tracking, but for most of the times where I want to access multiple instances of the same site I also want separate bookmarks and add-ons and history for each instance.
Google has a huge collection of data sources (search query correction, document translation engines, the Books project) to refine a spellchecker against.
The open-source solutions have never had access to that volume of data, to my knowledge.
LibreOffice is open source, and uses Hunspell. Here is how the three word lists in the earlier comment I linked to do in that.
1. On the list of words Firefox gets wrong that I've reported but haven't yet been fixed, LibreOffice gets them all right.
2. On the list of words that I indirectly reported quite a bit earlier that Firefox now gets right, LibreOffice gets one wrong ("heliocentrism", which Chrome also gets wrong, and MacOS gets right).
3. On the list of words that Firefox gets wrong that I have not yet reported, LibreOffice gets them all right.
I don't call Firefox's terrible spelling inexplicable because Google and Apple and Microsoft do much better. I call it inexplicable because LibreOffice does it much better (about as good as Google and Apple and Microsoft do it).
Firefox mobile's current version is a lot worst the v68... They block most extension, for one... I rolled back to Fennec even if they could not import my history because I dont care about it anyways
I would sincerely love to know how to make Firefox work this well for me. For reasons I still can't decipher, the browser tends to slow any computer I run it on terribly no matter what I do or which version I run. I've noticed this across multiple machines. Yes, I often leave dozens of tabs open but Chrome (combined with Great Suspender) never gives me trouble on this.
If anyone can suggest any sort of fix, i'd love to leave the snooping, prying Google browser behind finally.
Chrome was awesome once upon a time, but so far I've managed to avoid installing it on my 2020 MacBook Pro. I use a lot of Google services but it still scared me a bit when Chrome started offering the option (on by default) to log into the browser rather than just into a website. Colleagues at my last corporate employer rolled their eyes at me when I told them I tried to avoid it and suggested there was no serious alternative to Chrome DevTools for any serious developer.
You can disable it with SyncDisabled and BrowserSignin Chrome policies. On Windows, you just have to set the registry values. I believe there is a config file you edit for Linux and Mac OS. It’s all documented in the Chrome docs. You don’t need to pay for any Chrome enterprise nonsense.
I’m not saying this remedies this morally or, for most people, practically. But it is a practical solution for you and me and any machines we deploy.
Yes, I switched to Edge a couple months back, and honestly love it. All Chrome extensions work on it (afaik, and I've tried a lot). And the dev tools are the same as the Chrome ones, which is a must for me.
Sure, it sucks that it's Microsoft. But Microsoft is less evil than Google, imo.
I recommend anyone on the fence with Chrome to give it a shot. You won't even realize you left Chrome after about 30 minutes of using it. Even though Brave and Edge are similar, for some reason I was never able to stick with Brave.
Don't forget the part where the most recent release of Pagespeed conveniently ignores the speed impact of YouTube embeds when it didn't in prior versions.
A lot of interesting things happen, incentive-wise, when operating at massive scale. Doing something obviously unethical but "mostly-harmless" might only be worth fractions of a cent. But spread across billions of interactions the the benefit (and the harm) adds up quickly. This is the core issue of the problem with modern day mega-corporations.
Depending on the nature of the underlying bug, it would probably cause users that have their Chrome browser logged in (for configuration, history, password vault, and bookmark sharing) to be unexpectedly logged out of the browser when the cookie and local-store clearing logic fired, which is a bad UX.
Much worse UX if we're talking ChromeOS and it turns out the bug is that the logic that manages cookies and local state overlaps the logic that manages being logged into the browser-as-OS. "Navigate away from Chrome and I'm logged out of my whole computer" is no good.
What's so sad about this is that no one will ever get slapped for this. Not rank-and-file people that coded this, not the people who green-lit this, not the asshole claiming this is a "programming error", not anyone.
I know it's easy for those people to claim that they didn't know what they were doing or that they were told that it's fine. But maybe slapping a few would cause others to finally fucking think for a moment "do I wanna go to jail over this?". If you're making decisions that affect the privacy of literally millions of people then, well, maybe you need a little accountability.
1. It's next to impossible to make a mistake like that.
2. The mistake works out in Google's favour.
3. Difficult to give Google the benefit of a doubt anymore regarding anything.
4. Does anyone believe what PR dept. of a large corporation says anymore?
> It's next to impossible to make a mistake like that
It would appear that the root of this error is that Chrome is also an operating system. So in small, focused apps, yes, that's a hard mistake to make. In apps that some people are using as the main shell for their entire laptop, there's a lot of nooks and crannies for complexity to hide.
I think punishing one person (or even a group of people) is an ineffective tool. If someone gets fired for this, maybe we'll feel better, but I'm sure the cumulative tracking data would be worth it, and it wouldn't change their motivations for doing these sorts of things. It's similar to punishing an engineer for pushing a commit that causes an outage. They didn't come to work planning on an outage, and they probably didn't think they were going to cause an outage up and until the outage happened. If you fire them, you're doing nothing to prevent that outage from just happening again with someone else later.
124 comments
[ 3.3 ms ] story [ 198 ms ] threadIt's really hard to give Google the benefit of the doubt at this point.
Here's the fix, the bug is in chromium but only affects chrome because it has some "Hosted Apps" installed out of the box.
https://news.ycombinator.com/item?id=24817304 (1293 points/2 days ago/483 comments)
I call bullshit. Google is an evil company, this is nothing new.
Best context I can give is the actual CL for the fix.
Test clearing all private data -> not all private data cleared.
It /also/ wouldn't surprise me if it's exactly the way you describe.
In either scenario, it's equally bad.
In most fights between adding a user feature and protecting user privacy, if the feature cannot be added otherwise, feature wins. It's only "shady" if you're already assuming the feature vendor is a hostile actor.
The consistent Chrome experience enabled by logging the browser in is hells useful (especially in this era of mobile device + desktop device; no more having to remember if I bookmarked that article on my phone or my desktop or my work PC, etc.).
/Not/ protecting user privacy is always shady, regardless of who the actor is. When done knowingly, it's almost always hostile.
Once we assume that's not the decision that'll be made (because obviously), the question becomes whether the feature can be implemented at all while respecting user privacy.
And if the answer is yes, we can expect it to happen.
I don't subscribe to this school of thought, personally. I suppose that's where our differences lay.
What school of thought do you subscribe to? I had assumed that this
> /Not/ protecting user privacy is always shady
... being an absolutist statement, meant that you were open to the easiest way of protecting user privacy, which is "Don't increase the security surface area at all," i.e. add no new features.
Wouldn't we be able to tell when they push the fix?
On the other hand, a very large majority of the time that a programmer speculates on the internet about something that's obviously true about code I've written or am otherwise privy to, they're wrong
On the other other hand, most vague and fabricated statistics about anecdotal, speculative or subjective topics tend to have very little weight when making a point...
So I don't think the thought that any programmer would know this is a ridiculous edge case is one that holds any real weight here.
I think it's pretty sad that nobody here has linked to it... It's as if the people who care about this stuff aren't the same ones who write and review code for a living...
The likeliest point of introduction is where Google added account login to the browser itself. At that point, either ambiguity of specification or failure to test as though the user thinks Google is a privacy-hostile agency (while using a browser that allows them to login to Google) could have resulted in someone adding a special case to refrain from throwing the credentials away that keep the browser logged in that inadvertantly failed to throw the credentials away that keep a session logged in.
In the same way, browser makers should be decoupled from hardware and OS manufacturers, which have a similar monopoly problem.
Google bankrolled both Chrome and Firefox.
KHTML was the basis for the original webkit, and that was built basically for free.
[citation needed], as I switch from playing arcade games in my browser on archive.org to experimenting with a WebXR-driven VR experience.
- Google: obviously
- Microsoft: is a Cloud service provider
- Mozilla: VPN provider, also makes a social VR for headsets and browsers and is a cloud service provider
- Apple: iCloud, App Store, and Apple TV
- Brave: ad network wedded directly to the browser implementation
The law should be crafted in a way that allows for commercial agreements, as long as they are between equal parties, non-exclusive in nature, and time-limited if they involve default options. This way, browser companies will not be directly dependent from any single web or OS company.
Are you sure you're not just trying to rig the race so that only your favorite player can cross the finish line?
On my Android phone, I'll just go ahead and claim I'm not missing anything and it's not even close - Firefox is so much better just for being able to run extensions like uBlock Origin.
But on a workstation (macOS) - I'll keep an open mind.
Try it with a long page, something thousands of lines long. Mobile Chrome on Android chokes badly, stops responding for seconds or minutes, and page search is hideously slow. Firefox handles it fine. It's really obvious on long lists or tables, e.g. this page:
https://reitaisai.com/arts7/?p=165
Other than that, I'll admit, a large part of it is familiarity.
Even my most ardent anti-Google coworkers reluctantly switch back to Chrome when it comes time to debug.
chrome also renders UI interactions faster. maybe it's skia vs cairo? e.g. try mouse-dragging the ranging slider and compare the lag between chrome and ff:
https://leeoniya.github.io/uPlot/demos/zoom-ranger-grips.htm...
Does anyone know a Webkit browser that runs on Win - preferably with some kind of debugger?
Until recently, I had thought that browser compatibility issues were not much of a thing anymore.
It's still not worth running Chrome all the time. I only use it to complete those discrete tasks, then it's back to Firefox.
It's usually down to the remaining 10% to pick up the rest of the pieces.
It's extremely frustrating to hear this attitude as someone who was doing front-end dev in the IE6 era.
That's about the right ratio to expect, since less than 10% of browser sessions appear to be Firefox on average.
The major difference between this era and the IE6 era is that Microsoft wasn't willing to play ball with standards. With few exceptions, Mozilla and Google are implementing to the same standards and specifications; problem is, Chrome's a better implementation, for speed and stability, in the average case and the corner case.
They have both more resources to throw at the problem and more access to data to proactively discover issues.
It is, unfortunately, a situation where the playing field is "fair" (in the sense that there's a uniform standard) and one player is just trouncing the alternative.
Corner case maybe, but average case? No way. This is a historic perspective, maybe, but it's no longer valid in 2020.
Chrome isn't consistently the best of all browsers tested, but it beats FF more often than it doesn't.
This is not the interpretation I make of the current landscape. Google is SETTING standards that directly benefit themselves, just like MS did; Chrome uses it's market share to push in directions that benefit Google before they benefit end users.
1) site owner receives complaint
2) if offered, site owner looks up the user agent
3) site owner finds that the user agent matches one that is sub-10% of their traffic and samples the complaint signal accordingly
High-volume sites don't actually optimize for 100% of potential users (because the marginal value for one more user is low). If it's a low-volume site that happens to attract FF users for some reason, complaining is likelier to succeed.
Usually when I come across a broken website, it has more to do with my ad blocker and privacy extensions than the fact I'm using Firefox.
It recently got better support (as of v80), although it's not (fully) enabled by default.
I'm also on the FF boat, though :-)
Firefox (unless things have changed recently) still has limited addon capability -- no key remapping (or most other stuff) until the current tab has loaded. You also can't do unsigned addons (unless you do dev mode which is a separate install that doesn't update), while Chrome allows them by toggling "dev mode".
But no, I use Google Docs and log in to several bank accounts in FF without issues. I keep the other browsers just for compatibility checks for front-end development.
Just look at, for example, the recent madness with Firefox on Android. They effectively pushed a beta about 6 weeks ago, removed tonnes of features, made many menu items difficult to navigate to, removed compatibility with almost all add-ons, it's slower, it crashes for many constantly, it's just a plain, simple mess.
They had 4.5 stars at one point. They now have 3.9 stars:
https://play.google.com/store/apps/details?id=org.mozilla.fi...
Do you know how hard you have to work, to drag something down from 4.5, to 3.9, in a month, with almost 4M users?
I honestly don't think they could do worse, if it was intentional sabotage by some hostile insider.
And it's endless garbage like this from Firefox.
What firefox sees:
"OMG! Chrome is gaining ground! Quick, turf everything that is unique on our platform, that is different from Chrome. We must be just like them"
What outsiders see:
"Great, everything I cared about, unique to firefox is now gone. Plus it's slower, buggier, and they keep changing everything around on every single release. Great. Time to move on."
You don't compete with the giant, by being the giant. You etch out a unique market.
Meanwhile, Firefox did the precise opposite, and removed their uniqueness.
They aren't long for this world.
It's mildly annoying to re-learn the UI, but Chrome did the exact same thing to me years back when they, for example, dropped viewing every tab as a separate app in the running apps list. I refused to update for a year after that, very annoying. They all do this, I've resigned to it.
Feel the burn, the passion, the hatred. Join the dark side. Hate. Hate with me. Maybe our collective hate will drive the madness away.
For me personally the biggest current difference remains HDR support, which FF lacks and Safari/Chrome are continuing to grow. FF still has a big ways to go to achieve HDR so I think this one will be around for a while unfortunately. Now of course many are going to think "I don't give a shit about HDR" in which case refer to the first paragraph.
The biggest difference that finally got fixed was for the longest time Chrome had much higher quality pitch correction on sped up content e.g. YouTube 2x. It was noticeably worse at 2x and I couldn't even understand the content at 3x on Firefox but it sounded great in Chrome. This forced a lot of my usage over to Chrome until it was fixed.
On the Chrome side the biggest drawbacks are lack of extensions on mobile and lack of containers. I don't use containers as a privacy tool but more as a multiple profile tool without having to worry about which profile a window was launched in but I don't think that's the main real world use case happening. Lack of extensions on mobile is obvious.
> must perform with >50 tabs open
Testing Firefox now, it has the same behavior as Safari: when I open 16 tabs it starts a horizontal scrolling list. Absolutely unusable. Chrome shrinks every tab down to Favicon size or smaller and shows you 50+ tabs at once.
Chrome got tabs right a decade ago and has yet to mess them up, as they have with most of the rest of the browser. Very hard to switch until they do.
on FF, tabs open next to the current tab, unless you previously opened a tab in this one, then it'll go to the end of the list of tabs you've opened from the current window. but as soon as you switch tabs once, it'll start placing them next to the new tab you're in again. chrome just does something entirely unpredictable, instead of "resetting your spot".
but I've found that most people don't really seem to browse the way I do/don't care about that
Firefox uses the same open source spell checker, Hunspell, that Chrome, LibreOffice, MacOS, and many others use, but for some inexplicable reason have a terrible dictionary. I've ranted on this before [1] so won't do so again here.
I get the impression that the poor handling of multiple profiles is because they expect you to use multi-account containers instead.
Those are fine if you just want, say, to access multiple instances of the same site without using the same cookies for each, or want to make it harder for cross site cookie tracking, but for most of the times where I want to access multiple instances of the same site I also want separate bookmarks and add-ons and history for each instance.
[1] https://news.ycombinator.com/item?id=24819714
Google has a huge collection of data sources (search query correction, document translation engines, the Books project) to refine a spellchecker against.
The open-source solutions have never had access to that volume of data, to my knowledge.
1. On the list of words Firefox gets wrong that I've reported but haven't yet been fixed, LibreOffice gets them all right.
2. On the list of words that I indirectly reported quite a bit earlier that Firefox now gets right, LibreOffice gets one wrong ("heliocentrism", which Chrome also gets wrong, and MacOS gets right).
3. On the list of words that Firefox gets wrong that I have not yet reported, LibreOffice gets them all right.
I don't call Firefox's terrible spelling inexplicable because Google and Apple and Microsoft do much better. I call it inexplicable because LibreOffice does it much better (about as good as Google and Apple and Microsoft do it).
If anyone can suggest any sort of fix, i'd love to leave the snooping, prying Google browser behind finally.
I’m not saying this remedies this morally or, for most people, practically. But it is a practical solution for you and me and any machines we deploy.
Sure, it sucks that it's Microsoft. But Microsoft is less evil than Google, imo.
I recommend anyone on the fence with Chrome to give it a shot. You won't even realize you left Chrome after about 30 minutes of using it. Even though Brave and Edge are similar, for some reason I was never able to stick with Brave.
https://github.com/GoogleChrome/lighthouse/issues/10874
I don’t really understand it though, would it hurt Google to just delete data for their own domains? Do they really need it that badly?
Much worse UX if we're talking ChromeOS and it turns out the bug is that the logic that manages cookies and local state overlaps the logic that manages being logged into the browser-as-OS. "Navigate away from Chrome and I'm logged out of my whole computer" is no good.
I know it's easy for those people to claim that they didn't know what they were doing or that they were told that it's fine. But maybe slapping a few would cause others to finally fucking think for a moment "do I wanna go to jail over this?". If you're making decisions that affect the privacy of literally millions of people then, well, maybe you need a little accountability.
Here is the fix and the link the cl for the fix with a link to the cl that created the issue as well: https://chromium.googlesource.com/chromium/src.git/+/f800edb...
I'll let you judge for yourself, but this is not coming from PR.
It would appear that the root of this error is that Chrome is also an operating system. So in small, focused apps, yes, that's a hard mistake to make. In apps that some people are using as the main shell for their entire laptop, there's a lot of nooks and crannies for complexity to hide.
https://lapcatsoftware.com/articles/chrome-google.html