11 ms·

This is great to read/hear. I'll put my vote in for two things. 1) Performance, performance, performance; this is what originally caused me to use Chrome a lot

by Chirael 10y ago

This is great to read/hear. I'll put my vote in for two things. 1) Performance, performance, performance; this is what originally caused me to use Chrome a lot more, the perception of quick-to-open tabs and snappy responsiveness. 2) Privacy; this is something Google almost by definition can't compete on, because their business model depends on providing something free in exchange for user data (the opposite of privacy). A third important consideration of course is "not crashing" or at least if there is a crash, containing it to the offending tab - this is why I'm very happy to read about the progress of multi-process Firefox! Keep up the great work guys.

smw 10y ago
My only serious problem with firefox is that a single tab can lock up the entire application. Otherwise, the performance _and_ memory usage is fantastic. But that one issue is enough to keep me away from it most of the time.
- cpeterso 10y ago
  Multiple tab processes (aka "e10s-multi") should be shipping in Firefox 54 (now in the Firefox Nightly channel). The initial rollout of multi-process Firefox ("e10s") had one process for the browser UI and one process for all tabs (and a separate process for sandboxing Flash), so one bad tab could lock up all your tabs but your browser UI would still be responsive so you could close the bad tab. e10s-multi will increase the number of shared tab processes to two or more, balancing tab isolation vs memory overhead. In the meantime, you can test more tab processes by increasing the "dom.ipc.processCount" pref in about:config. I set my process count limit to 999. :)
  - kristofferR 10y ago
    I've heard a lot of people say that, but I'm still having problems with a single tab locking up the whole browser for 5-40 seconds on FF 53.0a2 Dev Edition with e10 and process count at 999.
  - smeehee 10y ago
    Hmm… I'd look but there don't seem to be any builds since late November on mozilla.debian.net.
  - Jonnax 10y ago
    I had a look a week or so ago at e10s, but it seemed that because my laptop was a touch screen laptop (Dell XPS 13) the feature was disabled. The UI said it was due to accessibility features. Is there a timeline for getting it working with touch screens or has that already happened?
- fencepost 10y ago
  While that's theoretically true, I've never switched from Firefox as my daily driver and crashes are very rare in my experience (more common on mobile, but I leave a bunch of stuff open there as well). Furthermore, when it does crash the session restore does a pretty solid job of restoring and letting me easily exclude tabs from being restored.
zbraniecki 10y ago
Hey, thank you for your input. We are aware of the importance of those three things. They're all not-trivial problems to solve, but we're working on all three: 1) Performance is an insanely complex problem. We can't easily measure the "real" performance, in a scientific way. We can measure performance of so called "microbenchmarks" which can tell you that, for example, our JavaScript engine can perform X million regular expression computations per second. The X may be higher or lower than what Chakra or V8 can, but how it translates to real world experience is insanely complicated. What we know by now is that JS engine performance is not the problem. At this point all modern engines are fast. What we need to advance are things like performance of DOM, CSS, layout, painting and prioritization (ability to ensure that UI performance cannot be blocked by website's JS computations). We're working hard on several projects this year, under the umbrella of codename "quantum" that is aiming at making Firefox blazingly fast. We're taking lessons and code from Servo and applying it in Firefox. The end result, if we succeed, is that we'll have ability to perform DOM, layout, CSS and painting orders of magnitude faster (even if it's 2x it's still huge!). The other piece of performance, is what's called perceived performance. That one is even more tricky because it's all about what your brain tells you. In cognitive psychology and HCI there are tons of studies that prove that depending on how we use tricks like animation, colors, shapes and gimmicks like focus, we can trick your eye to think that things are faster or slower. Chrome is really good at that and kudos to them! I love watching their UI painting order in 120fps slow-mo just to see all the things they employed. Together with Quantum we're going to look into ways to improve the perceived performance of Firefox. I hope we'll get it right :) 2) On the privacy front we're working with Tor browser and we're also working on several projects aiming at disrupting the current way of browsing the Internet. Read about Firefox+Tor browser and Activity Streams if you're interested. 3) On the QA front, we've made major progress over the last year. That's a huge part of the technical debt that we had to remodel in 2016. We now have much better fuzzers, tests and APIs that makes it easier to write code that will not break. On top of that, multi-process helps us make sure that painting crash doesn't crash your browser, that plugins can't crash your browser and finally that content can't crash your browser. We also designed a whole new language that is significantly better suited to write security critical code than C++ is. It's called Rust and we just started moving pieces of our engine to it. If we can succeed with this, we should end up with a codebase of the unprecedented stability for a project this size and facing third-party code. If you're interested, read about Mozilla Oxidation project. It is a major bet. We're betting on JS, we're betting on Rust. We're betting on Gecko and our ability to "change the engine mid-flight". I'm excited about the opportunity, but it's going to be a huge effort to pull it off. If any of this is interesting for you, help us! We're a small team (compared to all other browser engines), but we have a culture that fosters community participation and I believe that we can compete and shape the Web together: https://whatcanidoformozilla.org/ https://whatcanidoformozilla.org/
RileyKyeden 10y ago
I finally switched to Firefox when it dawned on me why Chrome for Android has no plugins, and why they went silent about the reading mode. With Firefox, I can install uBlock Origin, and it's had a reading mode for a while. Google wants to force me to look at ads that make the page slow to load, bouncy, and hard to scroll through. Google is where Microsoft was not long ago: engaging in self-destructive user-hostile behavior to protect their profit center. It can't work forever.
- caleblloyd 10y ago
  I also switched to Firefox Mobile on Android over a year ago. Couldn't be happier with the extension support. I've got uBlock Origin and my password manager installed. Chrome is definitely smoother on mobile but using extensions is worth it for me. Does Firefox Mobile's codebase closely match that of the desktop browser? Are their release schedules similar? Wondering if we should see it speed up with Firefox's latest improvements.
  - kbrosnan 10y ago
    Firefox for Android is the exact same codebase as desktop Firefox. Many Gecko features are turned on for Firefox for Android at the same time as desktop. In other cases either Firefox for Android or desktop will have features before the other. For example Firefox for Android had HTML date/time picker support years before desktop because the OS provided widgets. In the other direction EME has been on desktop for a while and is in the works for Firefox for Android.
  - cpeterso 10y ago
    Yes. Firefox for Android uses the same Gecko engine and release schedule as Firefox for Windows, Mac, and Linux desktop. They all share most of the same code in the same source code repository [1]. Most of Mozilla's "Quantum" project to start importing Servo code into Gecko will also run on Android. Firefox for iOS has its own release schedule and shares only a little code (e.g. Firefox Accounts sync for browser history/bookmarks) with the other Firefoxes. [1] https://hg.mozilla.org/mozilla-central/file/tip/widget https://hg.mozilla.org/mozilla-central/file/tip/widget
dandare 10y ago
If I could cast my vote it would be 1) perception of quick-to-open tabs and snappy responsiveness, 2) everything else. I just could not stand how unresponsive Firefox used to be.

by Chirael 10y ago

smw 10y ago

My only serious problem with firefox is that a single tab can lock up the entire application. Otherwise, the performance _and_ memory usage is fantastic. But that one issue is enough to keep me away from it most of the time.

cpeterso 10y ago
Multiple tab processes (aka "e10s-multi") should be shipping in Firefox 54 (now in the Firefox Nightly channel). The initial rollout of multi-process Firefox ("e10s") had one process for the browser UI and one process for all tabs (and a separate process for sandboxing Flash), so one bad tab could lock up all your tabs but your browser UI would still be responsive so you could close the bad tab. e10s-multi will increase the number of shared tab processes to two or more, balancing tab isolation vs memory overhead. In the meantime, you can test more tab processes by increasing the "dom.ipc.processCount" pref in about:config. I set my process count limit to 999. :)
- kristofferR 10y ago
  I've heard a lot of people say that, but I'm still having problems with a single tab locking up the whole browser for 5-40 seconds on FF 53.0a2 Dev Edition with e10 and process count at 999.
- smeehee 10y ago
  Hmm… I'd look but there don't seem to be any builds since late November on mozilla.debian.net.
- Jonnax 10y ago
  I had a look a week or so ago at e10s, but it seemed that because my laptop was a touch screen laptop (Dell XPS 13) the feature was disabled. The UI said it was due to accessibility features. Is there a timeline for getting it working with touch screens or has that already happened?
fencepost 10y ago
While that's theoretically true, I've never switched from Firefox as my daily driver and crashes are very rare in my experience (more common on mobile, but I leave a bunch of stuff open there as well). Furthermore, when it does crash the session restore does a pretty solid job of restoring and letting me easily exclude tabs from being restored.

zbraniecki 10y ago

Hey, thank you for your input. We are aware of the importance of those three things. They're all not-trivial problems to solve, but we're working on all three: 1) Performance is an insanely complex problem. We can't easily measure the "real" performance, in a scientific way. We can measure performance of so called "microbenchmarks" which can tell you that, for example, our JavaScript engine can perform X million regular expression computations per second. The X may be higher or lower than what Chakra or V8 can, but how it translates to real world experience is insanely complicated. What we know by now is that JS engine performance is not the problem. At this point all modern engines are fast. What we need to advance are things like performance of DOM, CSS, layout, painting and prioritization (ability to ensure that UI performance cannot be blocked by website's JS computations). We're working hard on several projects this year, under the umbrella of codename "quantum" that is aiming at making Firefox blazingly fast. We're taking lessons and code from Servo and applying it in Firefox. The end result, if we succeed, is that we'll have ability to perform DOM, layout, CSS and painting orders of magnitude faster (even if it's 2x it's still huge!). The other piece of performance, is what's called perceived performance. That one is even more tricky because it's all about what your brain tells you. In cognitive psychology and HCI there are tons of studies that prove that depending on how we use tricks like animation, colors, shapes and gimmicks like focus, we can trick your eye to think that things are faster or slower. Chrome is really good at that and kudos to them! I love watching their UI painting order in 120fps slow-mo just to see all the things they employed. Together with Quantum we're going to look into ways to improve the perceived performance of Firefox. I hope we'll get it right :) 2) On the privacy front we're working with Tor browser and we're also working on several projects aiming at disrupting the current way of browsing the Internet. Read about Firefox+Tor browser and Activity Streams if you're interested. 3) On the QA front, we've made major progress over the last year. That's a huge part of the technical debt that we had to remodel in 2016. We now have much better fuzzers, tests and APIs that makes it easier to write code that will not break. On top of that, multi-process helps us make sure that painting crash doesn't crash your browser, that plugins can't crash your browser and finally that content can't crash your browser. We also designed a whole new language that is significantly better suited to write security critical code than C++ is. It's called Rust and we just started moving pieces of our engine to it. If we can succeed with this, we should end up with a codebase of the unprecedented stability for a project this size and facing third-party code. If you're interested, read about Mozilla Oxidation project. It is a major bet. We're betting on JS, we're betting on Rust. We're betting on Gecko and our ability to "change the engine mid-flight". I'm excited about the opportunity, but it's going to be a huge effort to pull it off. If any of this is interesting for you, help us! We're a small team (compared to all other browser engines), but we have a culture that fosters community participation and I believe that we can compete and shape the Web together: https://whatcanidoformozilla.org/ https://whatcanidoformozilla.org/

RileyKyeden 10y ago

I finally switched to Firefox when it dawned on me why Chrome for Android has no plugins, and why they went silent about the reading mode. With Firefox, I can install uBlock Origin, and it's had a reading mode for a while. Google wants to force me to look at ads that make the page slow to load, bouncy, and hard to scroll through. Google is where Microsoft was not long ago: engaging in self-destructive user-hostile behavior to protect their profit center. It can't work forever.

caleblloyd 10y ago
I also switched to Firefox Mobile on Android over a year ago. Couldn't be happier with the extension support. I've got uBlock Origin and my password manager installed. Chrome is definitely smoother on mobile but using extensions is worth it for me. Does Firefox Mobile's codebase closely match that of the desktop browser? Are their release schedules similar? Wondering if we should see it speed up with Firefox's latest improvements.
- kbrosnan 10y ago
  Firefox for Android is the exact same codebase as desktop Firefox. Many Gecko features are turned on for Firefox for Android at the same time as desktop. In other cases either Firefox for Android or desktop will have features before the other. For example Firefox for Android had HTML date/time picker support years before desktop because the OS provided widgets. In the other direction EME has been on desktop for a while and is in the works for Firefox for Android.
- cpeterso 10y ago
  Yes. Firefox for Android uses the same Gecko engine and release schedule as Firefox for Windows, Mac, and Linux desktop. They all share most of the same code in the same source code repository [1]. Most of Mozilla's "Quantum" project to start importing Servo code into Gecko will also run on Android. Firefox for iOS has its own release schedule and shares only a little code (e.g. Firefox Accounts sync for browser history/bookmarks) with the other Firefoxes. [1] https://hg.mozilla.org/mozilla-central/file/tip/widget https://hg.mozilla.org/mozilla-central/file/tip/widget

dandare 10y ago

If I could cast my vote it would be 1) perception of quick-to-open tabs and snappy responsiveness, 2) everything else. I just could not stand how unresponsive Firefox used to be.