Firefox Private Browsing vs. Chrome Incognito: Which Is Faster?
blog.mozilla.org
blog.mozilla.org
What happened to the controversial cliqz program that they were testing? Last I checked, Mozilla did not respond to the privacy concerns of the integration.
This is what Microsoft also did with Edge and at least from a branding point of view, it worked quite well for them, because people do believe it's a new start from Internet Explorer. It's just that many didn't think that this new fresh start was better than Chrome for the most part (the fact that even now it only works with 70 extensions is a big reason for that belief, too).
I think Project Quantum hit the right spots for the most part. Headlines didn't just mention "Firefox 57" but "Firefox Quantum edition."
Hopefully the Mozilla team is writing down all of the things that worked well for them this time around (as a tactic checklist) so that they can repeat them for future major releases, too.
I enjoyed the read, but this is somewhat irrelevant unless a variety of people actually read it. I'm reasonably confident none of my friends read Mozilla's blog.
People find it through social media channels anyway. I'm only aware of it because of HN/reddit
It's the top bar ad on reddit for me right now (the topic, not this blog post). All the commentary is how people dislike it and are reverting to a prior version. Many are simply because it looks different. There is one person who talked tech, and in typical reddit fashion, he was called a corporate shill.
In no way am I saying the content is bad, I just struggle to view it as any significant or meaningful marketing strategy.
Edit: Not trolling, pretty serious. Would Google allow the webpages favouring Firefox over Chrome to show up on the top search results (or even the first page)? I'm not sure about it.
But, when you're engineering search results to manipulate public perception, you don't just obliterate all of the undesirable content with hamfisted obviousness. You just weight the content differently. Maybe you ensure that anyone already looking for that content can find it, but ensure that a more generic search is favorable. Maybe you just manipulate the heap to ensure the top result is the good one. Maybe you merely aim for a favorable ratio of desirable to undesirable content.
Consider: "best search engine"
The results are full of discussions about alternate search engines. Both duckduckgo and dogpile make the front page. Yet, the #1 result is an article about how Google is the most popular search engine: https://www.reliablesoft.net/top-10-search-engines-in-the-wo...
Is that true? Sure. Is the result manipulated? Probably not. But if I was google and I was going to manipulate a search result, that's exactly how I'd do it. I'd maintain the illusion of impartiality by presenting lots of alternatives, but I'd ensure the most prominent result was the one I wanted.
In any case: a quick search is insufficient to detect search bias and proves nothing.
Plus Mozilla is a non-profit, so they're probably more interested in their reputation specifically amongst HN readers and the like than they would be if they were a commercial company just trying to sell as many widgets as possible.
The point, though, is that they could. If you wanted to drop links to them with more details about why the should be excited about Firefox again, Mozilla is actually producing those.
This is one of those things that only seems unimportant until people stop doing it. Then you find out it was actually a pretty big deal.
If we do our part then everyone might be rewarded with a healthier ecosystem over the next months. (And "everyone " includes die hard Chrome users as well as people forced to use IE at work.)
Edit: what I mean is that all consumers generally benefits from healthy competition.
They are probably proceeding, sadly. I have not seen anything indicating the opposite.
The reasonable thing to expect at this point is that it is still at the phase of being an experiment for 1% of German users.
Mozilla owns parts of Cliqz, so they have insight into the business process, and they have a contract with Cliqz/Burda that says that no personal data is allowed to be stored.
So, even if Burda decided to violate the contract/law, Mozilla would notice that and could pull out by simply shutting the sending of data off in Firefox.
Mozilla also has made an official statement that no personal data will be collected, so if they did anyways, it would be misleading of consumers, so Mozilla themselves would then violate the law.
Meaning that, unless you have no faith at all into the legal system, I don't think there's a reason to be concerned.
It's just controversial, because you can easily spin a wholly different story by leaving out some facts, which is of course extremely convenient for Google/Microsoft/Apple, anyone who's for whatever reason salty about Mozilla, and last but definitely not least, clickbait journalists.
Mozilla stated "Cliqz uses several techniques to attempt to remove sensitive information from this browsing data before it is sent from Firefox"[1], which is very different from "no personal data will be collected". Even just a list of domains can be enough to reveal someone's identity.
[1]https://blog.mozilla.org/press-uk/2017/10/06/testing-cliqz-i...
You presuppose that it's reasonable, but is it? If you can prove that you're doing everything right to protect the user, should you not be allowed to collect the data?
If we just call every non-opt-in data collection unethical, no matter if it can actually negatively affect users or not, then you're levelling the playing field in favor of those that actually collect and use data unethically. Because you're essentially saying that it's all the same, unless you're not collecting data at all.
And for some services, you simply do need to collect data, and sometimes you even do need it to be non-opt-in, because otherwise your data sample will have a severe bias.
> Mozilla stated "Cliqz uses several techniques to attempt to remove sensitive information from this browsing data before it is sent from Firefox"[1], which is very different from "no personal data will be collected". Even just a list of domains can be enough to reveal someone's identity.
If you just so happen to read one sentence further than that, it says "Cliqz does not build browsing profiles for individual users and discards the user’s IP address once the data is collected.". That says exactly that no personal data is collected.
And yes, a list of visited domains can be used to identify a person, but it's not necessary to store the visited domains aggregated by each individual user for this. They'll just intermix the domain lists from all users.
Otherwise it's a known bug. Apparently not enough people complained about it so it is perceived to not be a blocking bug for 57.
Now - Chrome is slower vs Firefox Quantum's new privacy because of the ad networks. The ad networks have had no incentive up until now to improve their shitty code.
If people start switching to Firefox and turning on privacy more to get faster browsing - they will lose the game entirely.
All that's needed is for a modest uptick in usage of this mode and for some ad networks to get scared. They spend some of the pile of money they swim in to not ruin everyone's browser experience and bam - we get a less crappy internet.
If they don't shape up - then we get a faster internet with better privacy. Even better.
I assume these channels are a huge reason why people adopt it. I love the new Firefox and use it every day but, unfortunately, I don't think it will come close to Chrome's market share by just being slightly better.
Never. Enable. Notifications. They're just product management spam.
In contrast in Chrome the keypresses are taken into account immediately, even if the window takes a bit longer to initialize as a user I don't have to wait, I can type my query, let Chrome do it's thing and eventually I'll get my result. In Firefox I have to wait until the adress box is ready before I can type something. Even if the whole process is faster it means it still takes more of my attention time.
I would love for this to be true, but there are a number of websites that simply refuse to proceed unless you whitelist them from script/adblockers (mainly news so far but even e-commerce sites could require you to submit to their analysis tools e.g.: Bloomberg).
Someone needs to write a browser extension to block the adonis ad-blocker-blocker. You can find it in base64encoded JavaScript on some news sites (like mashable).
My experience for 3 major browsers on MacOS as follows:
- If you want noticeably longer battery life when on the travel.
1- Safari
2- Chrome/Firefox
- If you want seamless browsing experience.
1- Chrome
2- Firefox
3- Safari
Firefox still does not support VAAPI or VDPAU :(
However, I have two major tab-management issues with Firefox which make it near-unusable for me:
1. No support for "scroll wheel to switch tabs" (scroll on tabs to switch between them). This is a feature common to nearly all Linux software and Firefox removed it a few versions ago. Quantum killed "Tab Wheel Scroll" which reimplemented it.
2. No support for multi-select tab management, which Chrome is amazing with. If you didn't know, in Chrome, you can ctrl+click and shift+click tabs to select multiple at once and close, move or snap all of them at once like in a file manager. Super useful.
Wow you were not exaggerating on that.
There seems to be a new add-on for the web extension approach (https://addons.mozilla.org/en-US/firefox/addon/foxy-gestures...), but I haven't tested it yet and am not sure how robust it is at this point.
I've yet to find a gesture extension that supports the same rocker gestures (i.e. middle click + other click to change tabs).
Also mouse gestures now have the same annoying "feature" they've always had in Chrome: they don't work in "system" tabs (think about:preferences).
So you can't use mouse gesture to quickly go through opened tabs if at least one of them is the "system" one, the navigation simply dies there.
I don't think they're ever going to fix the "system tabs" issue, since it's supposedly a "feature", but at least some good news. :)
Yeah I am missing this as well. I switched to Firefox a few days ago after hearing about all the improvements.
So far, three issues:
1- On my MBP, I am having serious issues when I have tabs to recover.
2- HTML5 videos are not hardware accelerated. Neither on OSX, nor on my Arch.
3- Cannot move multiple tabs from one window to another but that actually seems to work better with i3's drag and drop.
Hi; I work on scrolling/rendering in Chrome. Thanks for bringing this to my attention, that is some super janky scrolling in Chrome >_<. I've filed http://crbug.com/786991 for us to look into it.
Some technical background (since this is HN after all); at a quick glance there are two things going wrong here in Chrome.
1. We would normally want to 'composite' the scrolling region here, which means we paint the pixels only once into a saved layer and then just slide that up and down as you scroll. I believe that is not happening on the Stripe page because of the unusual way that they are doing the split-panel background. Looks like we're currently unable to tell that the background is an opaque color, so we can't apply sub-pixel font AA[0] - and on low DPI we will refuse to composite in that case.
2. Chrome's repainting of this page is really slow. :(. Firefox Nightly is also repainting the page when you scroll (as far as I can tell), but they're definitely doing it much faster than Chrome!
So our goal will be to try and figure out how to composite this page, and if that's not possible then we'll have to dig into the slow painting.
Note: Because we composite more readily on high DPI devices (where sub-pixel font AA[0] doesn't matter), this site should scroll nice and fast on a high resolution display. This includes most new laptops, all (?) phones/tablets, etc.
> Note: Because we composite more readily on high DPI devices (where sub-pixel font AA[0] doesn't matter), this site should scroll nice and fast on a high resolution display. This includes most new laptops, all (?) phones/tablets, etc.
Not the case for me. Touchscreen display on a UX asus laptop, 3.2k resolution on linux with intel card. Scrolling using the touchscreen is snappy, scrolling using the scroll wheel (synaptics emulation) is even worse than on my desktop.
I’m about to go mad because I’ve been reporting bugs in subpixel aliasing under hi-dpi displays with non-standard DPIs (!= 96dpi) for a while but the Chromium QA team in India has been having a hard time triaging them, possibly because of the hi-dpi display requirement.
I’ve also got a fully reproducible issue involving CSS transforms I haven’t bugged yet as I’ve just been struggling to get someone to triage #765848 correctly as of yet:
https://bugs.chromium.org/p/chromium/issues/detail?id=765848
Can you send me an email, pretty please? mqudsi [at] NeoSmart [dot] net
I was experiencing this particular bug on all my hi-dpi machines but it magically went away on one of them (perhaps after an OS update, as it’s running Windows 10 RS4 seeds) but it’s fully reproducible on all my other machines with a clean profile on all channels.
Looks like https://bugzilla.mozilla.org/show_bug.cgi?id=566510 tracks the multiselect thing in Firefox and has seen some recent movement... That would indeed be very nice to have!
I find my web browsing experience on Os X quite bad overall.
Not sure if this is something YouTube does or it's a general video playback problem.
I observed it before FF 57 was released, on different Windows computers but especially catastrophic on the less powerful ones ( https://news.ycombinator.com/item?id=15688195 or https://news.ycombinator.com/item?id=15687830, I've mentioned that I also discovered that there are complaints at least since February this year, https://www.reddit.com/r/firefox/comments/5sikxt/firefox_unb... only to be downvoted heavily, however let's ignore that at the moment) but now on FF 57 it is provably the same. Just produce a page with more videos autostarting and marvel how much (unbelievably lot!) CPU Firefox uses.
8 videos playing on the page are enough to demonstrate it. Fascinatingly, even if they play somewhere on the page outside of the part which is currently displayed, and even if they are not producing any sound (marked as such) they are going to eat immense amount of CPU.
The thing that bugged me about the mobile app was that coming from Chrome it didn't seem to have the same level of integration with my phone. For instance I'd google search something, get the result which included a map, and then the map was not clickable. I expected to be able to tap it and have the map open in Google Maps but no. It was a static image.
I had similar experiences with other apps like Twitter, Instagram, and Redfin. I expected a link to the mobile site to launch the installed app on my phone but it never did. I looked around a bit in the settings and couldn't find anything to do it. I'm sure there's a solution for it but I got tired of poking around and went back to what just worked.
https://bugzilla.mozilla.org/show_bug.cgi?id=1418510
VP9 is disabled by default in Firefox on Mac because H.264 can be hardware-accelerated on Mac but VP9 can't. Firefox supports hardware accelerated VP9 on some Windows machines, though.
That's a setback for royalty-free media formats on the web. Is there a plan for VP9 to be enabled by default again in the future?
Unfortunately 3rd party browsers are at a disadvantage because it is harder to automatically detect and use video acceleration hardware capabilities.
https://developer.apple.com/library/content/technotes/tn2267...
The VDADecoderCreate function will return an error if HW video decoding isn't available.
Equivalent features are provided in the Video Toolbox framework from OS X 10.8 and later.
Firefox only supports 10.9 and later from now on, and no longer uses VDA.
Unlike VDA, when you create a decoder, you just get one. You then query if it's a hardware decoder or not, but the behaviour is fundamentally the same between SW and HW.
For some reasons, the VT frame gives us a software decoder most of the time for some users.
Until now, I had only ever seen Mac Pro (including 2013 trashcan model) to not provide hardware decoding support.
If you attempt to decode a video that is less than 144 pixels high, you always get a software decoder.
The only major complaint I have with Firefox (as of 57) is with openh264. It grabs a whole CPU core for a video that Safari handles at a fourth of that or less.
There shouldn't be much difference for h264 videos between Safari and Firefox, except in Full Screen. For full screen, Safari uses the OS to render the images and can have a very efficient code path. Firefox doesn't support it at all (for now)
Firefox doesn't use a blacklist of drivers / GPU like it does on window to enable HW decoding or not. Typically mac hardware is far more reliable.
the media.hardware-video-decoding.force-enabled is only of use on Windows in the event a card is blacklisted and you want to override the blacklist
With the amount of tracking going on these days, a FF Private window is literally the only way I'll sign into either Facebook or Twitter. I wish they would also enable Tracking Protection and Do Not Track by default on non-Private browsing as well, but that was an easy enough change for me to make.
Perhaps Firefox should display a "privacy start screen" the first few times the browser is launched after downloading it, explaining the features to the user, and allowing them to opt-in on their own decision?
For things like DNT or Tracking Protection, I imagine most common users do not know about these things so the usage rate is much lower than it "should be".
Fun fact: in Firefox you can actually turn Tracking Protection on to use it all the time, not just when you're in Private Browsing mode: Options > Privacy & Security > Tracking Protection > Always
And Tree Tabs [1] seems to be doing something functional into the direction of tab groups. Maybe that works for you...
[1]: https://addons.mozilla.org/en-US/firefox/addon/tree-tabs/
Let me explain: I've used Firefox since the beginning (and Navigator before that). I like Mozilla and I agree that most of the changes that they've done internally in quantum were kind of necessary (although many could have been handled better).
However, along the way, they've broken the way many people use Firefox and the only answer the frustrated users seem to get is "but it's fast".
They've changed the UI to be similar to Chrome, to the point where the reload button is not where it used to be. But it's fast.
They've broken the extensibility, which was (at least for me) the main reason to use Firefox instead of Chrome, Chromium or any of the others. You could do things in Firefox that the others couldn't even dream of. But it's fast.
While doing this, they've alienated plenty of plugin developers, people who loved the Firefox ecosystem, some of which have had to rewrite their plugins thrice already (with some being told to, essentially, "bugger off"). But it's fast.
As far as I can tell, it's using more resources in MacOS than the old version, at least the processor runs hotter and the fans kick in more often. I've had issues with the CPU use sky-rocketing, but I believe that has more to do with opening multiple tabs quickly rather than with one specific page. But it's fast.
Honestly, I haven't noticed any major changes in speed in Quantum (it's a bit snappier, that's it) and was happy enough with it's previous performance. IMHO what we've lost is not worth what we've gained.
I'd move to other browser, but which one? None of the main ones have the things I liked about Firefox and I don't really see most of the forks being able to keep up the pace. I guess that I'll stay, hoping that the things I love about Firefox will eventually come back.
While it's possible to re-implement Tab Groups with the new API, extensions like SQLite manager [1] will never work again.
I wrote instructions [2] on how to restore lost tab groups after update (involves downgrading and disabling updates).
[1] https://addons.mozilla.org/en-US/firefox/addon/sqlite-manage... [2] https://github.com/Quicksaver/Tab-Groups/issues/560#issuecom...
It's mainly more secure by having most things sandboxed now. That requires Firefox to be executed in the new multiprocess architecture and for that they would have had to kill off at least those legacy extensions that are not multiprocess-compatible, which was essentially all of them as well.
Also, yes, multiprocess did also most definitely play into performance. But we're not talking about the difference between 56 to 57, it's rather the difference between 47 to 57. Because 48 already introduced multiprocess, even though as a user you were still perfectly capable of installing a multiprocess-incompatible extension, which then caused Firefox to drop out of multiprocess whenever that extension became active. A completely intransparent way for users to completely kill off their performance.
Lastly, you don't claim that what Mozilla did on a whole is wrong, but yeah, just to point it out anyways: They are somewhat throwing power users under the bus right now. They don't enjoy doing it (essentially all Mozilla devs just as well are Firefox power users), but they do want to save their market share in order to still have a word when it comes to web standards and the market share just isn't made out of power users.
Around 40% of Firefox users in Sep. 2015 did not have any extensions installed. Another sizeable chunk will have their ad blocker and nothing else. This release won't affect these users negatively at all. They'll just get the better performance and the better security.
The only thing I heard from my parents was my mum asking me, if it's normal that Firefox looks different now. Maybe there'll be some confusion with the new Library-button, but other than that, it's just yet another browser-UI. They know this.
However when comparing private/incognito browsing, my main concern wouldn't be the speed, but privacy.
There are no errors, but I notice a LOT of GC Pauses in the perf dump
Chromium works just fine on these sites.
This is with fully clean install and several attempts at cleaning profile. I notice it seems to get stuck on "waiting for site" and "Performing TLS Handshake", which is strange because im on a 100gbit line and I can get chrome to load those sites full speed just fine simultaneously.
X1 Carbon Fedora 26 Intel Core i7-6600U 16 GB RAM
Mesa DRI Intel(R) HD Graphics 520 (Skylake GT2)
Do you have a good place to send these profile dumps to?
Posting the videos of Chrome vs Firefox Quantum somewhere like YouTube might be an interesting thing to watch.
How'd you get a 100 GbE NIC in that X1 Carbon!? :)
My situation is the opposite: workstation with a 10gbe connection the the corporate lan but upstream is capped to 2gbe.
After doing that I've run it on my laptop's 4k screena nd a 4k external monitor on integrated graphics with no performance problems whatsoever.
That way the omnibar suggestions show me content that I'm more likely to revisit and doesn't get clogged up with random results I don't care about; and YouTube gives me more interesting recommendations.
Keep in mind the algorithms for session management are complicated and a session might persist past when you thought it would.
A marriage of private browsing + containers would also do the trick.
[0] https://addons.mozilla.org/en-US/firefox/addon/containers-on...
However, I experience severe speed problems loading Google apps (drive, sheets, slides, inbox) along with occasional freezes and CPU spikes.
Additionally, I experience a bug on iOS where certain pages will only render partially and turn completely blank when I scroll down. Refresh doesn't help, which makes me open Chrome again to load the page.
Firefox is still my default browser as I'm still trying to convert myself, but it turns out I'm typing this message in Chrome again.
Note: I'm on Linux
Other than that though, Firefox is my favorite browser.
it's nice that private mode is more private and faster by default though.
I work on a University campus, and when I'm on the Uni wifi (eduroam), google search does not work in FF. However, it works just fine in Chrome.
Super weird and I haven't been able to debug the problem.
Changed my system's wifi settings on the network to use Google's DNS (8.8.8.8, 8.8.4.4) and it worked a treat. Does Safari also have it's own DNS resolution? That is why I missed this - googling worked just fine in Safari.
They're also planning [] to send more detailed browsing history as telemetry, so there's that.
[] https://groups.google.com/d/msg/mozilla.governance/81gMQeMEL...
Extensions operate on requests before the platform-level channel classifier.
In theory, the platform-level classifier should be faster than the webRequest API, but we've not measured it specifically.
Running both uBO and and Tracking Protection can give some extra coverage; I've noticed sometimes the Tracking Protection shield pops up even though uBO (and Privacy Badger) are running.
However, it's not obvious to me that the protections offered by incognito are a superset of the protections offered by UO and I wonder why this is desirable ?
Do I want incognito mode with UO running on it or is that unnecessary ?
Last time I checked chrome does not by itself offer any tracking protection while in incognito mode (supposedly since this would make it easier to detect when a user is in incognito mode). So you probably want UO enabled for that.
chrome://extensions -> uBlock Origin -> allow in incognito
The tradeoff is that many extensions are not safe and could spy on your incognito activity if allowed to work incognito
Try Bugzilla [0] or leaving a comment on the blog post instead.
It might be biased since it's from µblock developer.
On a modern Mac with plenty of RAM the difference isn't really noticeable.
As far as I know, it works with all of these now except for Xmarks. I am impatiently waiting until then.
Works, but very buggy with multiple sync sources, might be server-side, I've had trouble with failed syncs losing entire folders for a long time. I only use the Upload/Download functionality and make sure not to have conflicting syncs. The product has been on minimal life support for years, so can't expect much I guess. But remains the only solution I've seen that works cross-browser.
[1]: https://addons.mozilla.org/en-US/firefox/addon/multi-account...
You can also use command line flags to do the same. For example, you can run `firefox --profile $(mktemp -d) --new-instance` to create a temporary, throwaway profile. (Change --new-instance to --no-remote on Windows).
b) if they didn't make noise about this no one would try to switch to this fairly big release. It's a bit like the hype around Firefox 3.
One thing I've learned in life is, you can do great work, but if people don't know about it or appreciate it, it simply won't help you.
More info at https://wiki.mozilla.org/Security/Fingerprinting
Is there a reason that setting does so many different things?
I've been trying it for a few days but it's a real problem when I want some of those settings but not others. For example the protection against canvas image extraction needs work. It's stuck on 'ask' with constant popups, and once you pick yes/no on a domain it's impossible to see the permission or change it ever again. Until it forgets on browser restart and starts bugging you all over again.
Every percent we make the browser faster makes it more usable for people (and libraries!) with older, slower computers. A lot of the world can’t afford to upgrade. When we make it faster for everyone, we make it faster for them too.
If we save 0.3 seconds per page view, how many millions of hours do we save humanity from waiting for their browser to do something it could do faster if we just worked hard enough? That’s why faster is good, to me, and so this campaign aligns well with my personal reasons for working here.
Can you explain in a just a bit more details how Firefox "save humanity" 3 seconds if I and nine more people around shave off 0.3 seconds from some web page load and what exactly that mean?
Practically, I'm not sure how I feel about this metric on the 0.3 second level, but I sure do understand it at the 5 second level. I've stared at useless pages that are stuck on some tracking JS so many times I can't count, so if those stop making me stuck, that's invaluable once and multiply that win by a million.
But that's just time. If you also consider the resources that are needed to execute the page JavaScript, we're probably saving megawatts of power per interval (totaled across all Firefox users) with each small efficiency gain. The changes that are "parallelize all the work" may cost more, less, or same amount of power; the ones that are "use fewer CPU cycles" will definitely cost less power. And so there's a very real theory that we could probably measure how many gigawatt^-hours Quantum has saved the world to date, which appeals to me greatly :)
^ I arbitrarily picked "a million" and "gigawatts". This does not offer any guidance whatsoever on specific or approximate user counts.
EDIT: My UPS says that my entire desk has a load of ~75 watts @ 120V, times a million^ users = 75 megawatts. I haven't the faintest idea how much power it takes to evaluate Javascript inefficiently, but saving 0.3 seconds of any amount of power times a million^ users per page view quickly balloons into reality-impacting figures.
Personally it got me to start running Firefox again. I probably wouldn't have bothered if they hadn't been making such a hubbub about it.
I remained a loyal FF user myself and actually wanted to take this opportunity to promote Firefox Focus which makes it real easy to browse sensitive things like porn on a tablet shared by the family or SO.
It's oriented around privacy so it saves nothing and provides easy to access shortcuts for clearing browser history when you're done.
And all that wasn't enough since they also shadily (based on user agent for no good reason) blocked features for other browsers in their products.
For what it's worth, "millions of dollars" is probably understating things by 2-3 orders of magnitude. The ad budget for Chrome was likely comparable to Mozilla's entire budget for a few years there, even if you ignore the free product placement on google.com...
Step 1 is "Get people to (re)try the product".
Unless there's a magic wand waved that gets people unaware of a milestone release to (re)install and (re)launch FF57, Mozilla needs to put out some marketing just to even seed the basis for word-of-mouth of the product's merits.