Chrome 27, Firefox 22, IE10, And Opera Next, Benchmarked
tomshardware.com
tomshardware.com
1. Destructuring references is not yet supported on Chrome
[a,b,c]=[c,d,f]
[a,b]=[b,a+b]
[1,2,3,4].map(function(x) x*x)) // [1,4,9,16]
Granted CoffeeScript makes some of these easy, but they are too paranoid in my opinion to support new and useful syntaxes (e.g. yield keyword for generators, getters/setters).
[1,2,3,4].map(x => x * x) // [1,4,9,16]
This blog post explains the differences between these functions and current functions: http://javascriptweblog.wordpress.com/2012/04/09/javascript-...; i wish the unbound "->" alternative is included, so that functions used as methods can also use the new syntax.
Eh. Yea, Firefox got support for that (and `let` and generators/iterators), because they added it in 2006 (Firefox 2.0) as an experimental feature.
https://developer.mozilla.org/en-US/docs/Web/JavaScript/New_...
[1,2,3,4].map(x => x*x)) // [1,4,9,16]
Also, destructuring objects:
let obj = { foo: 42, bar: { baz: 32 } }
let { foo, bar } = obj;
foo // 42
bar.baz // 32
List of all the goodies: https://developer.mozilla.org/en-US/docs/Web/JavaScript/New_...
edit: formatting
It just doesn't feel like a piece of software that I enjoy using, like it used to some long years ago. Fix that, and you win me back.
PS: I think the memory score is not so relevant as a few years ago. People talked a lot about FF memory, but that was because it was leaking like crazy and ultimately lead to crashes of your browser. I don't care how much ram my browser is using as long as it doesn't start swapping nor crashing.
1) Start-up time.
2) Performance with 25+ tabs open.
3) Performance on older devices.
4) Performance with a "basic" set of extensions.
I don't really care about benchmarks these days; I am much more interested in whether a browser supports HTML5, WebGL, etc. I'm currently bummed that Opera 13 doesn't display shapefiles on GitHub. :(
(This is a tough one to debug, perhaps if Firefox had a pixel perfect Chrome theme it'd work for you?)
Firefox feels like Eclipse to me.
So yes, I'm hand waving a comparison between two pieces of software who have great technical features but feel painful to use, because this is a comment on a benchmark post and not me doing a bug report.
However, on a more limited machine (Since when is a 4GB machine not good enough to browse the web???), Firefox definitely wins and is much more enjoyable.
I was a long time Chrome user ever since it came out. However, recent versions (on OS X and Windows) are unusable with a large number of tabs. Firefox handles many tabs much better.
Frankly the memory usage isn't as infuriating as the "tabs extend past the right edge of the window" bug, which has been open for forever -- I'd link to a bug report but it's a nightmare to find with a search engine because of how hard it is to describe. Firefox has a simple fix for this -- they let you scroll the "row" of tabs -- but for some reason Google refuses to fix it this way. It's probably too rare of a use case; you have to have a relatively small resolution monitor and at least ~70-80+ tabs open in a single window, IIRC.
> It just doesn't feel like a piece of software that I enjoy using
> Fix that, and you win me back.
Fix what? Those are extremely vague complaints. Why does it feel clunky? Why don't you enjoy using it?
FWIW, firefox still took more than twice the memory to show that site than IE did.
1) look around on http://planet.mozilla.org https://wiki.mozilla.org/Performance etc. people are working on UI-specific performance issues (commonly called "jank")
2) try the UX branch builds for the latest UI which is coming soon - https://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/late...
Progress is being made, but it takes a while to hit the beta/release channels.
It's pretty critical that such improvements don't cause stability/security/standards problems too so they're not the simplest of changes!
The memory efficiency doesn't explain much to me. It's like complaining that linux is holding memory hostage in buffers/cache or that Chrome should lose points for its process model. The test surely doesn't tell the whole story; not to mention, I can't be bothered to worry about a couple 100M when I can get 160 of them for the price of dinner. For all I know, the chunking I experience with FF could very well be a relic of their process and/or memory model.
And security is just a smudge on the score? No thanks.
This only leads to one conclusion: unless your product is largely better than the competition, its all about who's the cool kid on the block and who has the most financial interests.
"news at 11" i guess
Last time Opera crashed on me? Hmmm…
Anything specific?
Feel free to file bugs, we want to make it better :)
https://bugzilla.mozilla.org/enter_bug.cgi?product=Firefox&c...
I see a lot of these things landing in the FFs landing in the last week, so I can't commend those efforts enough. Literally, I don't want to have to ever install Firebug; it makes the browsing experience even more sluggish. And I would have probably never started using FF, a decade ago, if it wasn't for the tooling, and I would have never had a chance to convince as many friends and family to use FF. The same reasons I will just recommend Chrome, today.
Frankly, you will have to have something that Chrome doesn't to bring us back, and it's been a long time coming for the tooling to be even on par. Again, I commend your efforts, because I think your community involvement and dev blogs are really great.
(I keep development separate from regular browsing, either through Incognito or separate Firefox profiles. This way, my regular browser addons don't distort my view of how-my-website-looks-to-most-people, and Firebug speed becomes irrelevant.)
Just a note, we use bugzilla for everything, so if there is ever a specific feature you want, you can make a bug for that feature. For example, "integrate source maps with the debugger" was the title of a bug that I worked on.
One thing that irks me is inspecting returned JS values - I almost never want to see [object Array]. I vastly prefer the Chrome dev tools output: inline, expandable, and a meaningful representation for objects.
Not a bug, but I think it would be a big improvement :)
BTW, we use bugzilla bugs for features too, so if you feel there is a specific feature missing, just open a bug for it.
I like what Mozilla is doing with regards to privacy and so on though...
Try out a UX Nightly http://people.mozilla.org/~jwein/ux-nightly/ to get a feel for the work-in-progress version.
I switched from Chrome a few weeks back and had the same opinion about firefox's native UI. FxChrome and Close-Tabs-to-the-Right [https://addons.mozilla.org/en-US/firefox/addon/close-tabs-to...] made the switch completely worth it.
Edit: I noticed a difference in the default scrolling as well. Try disabling smooth scrolling (Options -> Advanced -> General.)
All in all, I can't applaud Mozilla enough for pulling prolongated and difficult efforts (the new 3-layered releases process, memory compaction, js jit) with only a tiny percent of Chrome's ressources. This is really impressive.
Pinned tabs are not included in any benchmarks I've seen but have a drastic positive impact on memory usage and speed. I typically have 20 or so pinned tabs and only 1-3 unpinned tabs when I quit Firefox. A cold start takes just a few seconds and about the only time I restart the browser is when I restart my system (2-3x per week).
Chrome loads all tabs on start so it can take around a minute to load the same 21-23 tabs. And furthermore Chrome tends to operate sluggishly after a couple hours of use with so many open tabs, so when using Chrome I tend to close then reopen the browser every couple hours.
Chrome 5 had a similar feature called phantom tabs but it was removed in Chrome 6. I loved that feature and was very unhappy when it was removed.
FF 'won' this benchmark despite a unfair memory score against it.
Pages that were recently visited are stored in memory in such a way that they don't have to be re-parsed
I guess the same applies for closed tabs, since those can be re-opened as well.
What does Tom think caching is, or what unused RAM is good for? To be a pretty and useless number in the task manager? That not all users understand this is one thing, but a so-called reviewer? Just wow. Thanks for actively making people worse informed, Mr. Expert.
I use suckless.org's surf now, and it "feels" faster than any of the other browsers, despite no doubt faring worse in benchmarks (and not being generally accessible, lagging tabbing and a url bar)
I gave up on Chrome because of Flash crashes, general crashes (two or more per day), and pages showing up as plain, pale blue pages (instead of the web page content). Maybe I had too many tabs open and not enough RAM (3GB). Either way, the tabs were too small to show what the pages were, which wasn't helpful.
Firefox has worked perfectly, and it doesn't feel any slower, so I've no complaints....
Congratulations to the memshrink team for fixing all the memory leaks in FF 4. Now if they could only do something about the UI responsiveness when multiple intensive tabs are open, I might consider switching back from Chrome on my desktop.
I really do hope FF catches up.
> all downloaded images are decoded and live forever regardless of if they are shown.
Nope. Before the backed-out-of-FF22 patch landed, all images in a page were decoded and kept in memory as long as that page was in the foreground. But if you switched to another tab, the decoded images would be discarded.
With the backed-out-of-FF22 patch in place, most images that are downloaded are decoded temporarily, but those that aren't visible are discarded quickly.
When bug 847223 is done, the not-visible images won't even be decoded on page loading.
So, to summarize: in FF24, on image-heavy pages you'll see a memory spike on first load, but it'll quickly drop down. And once bug 847223 is done (hopefully soon, though these things are hard to predict), you won't even see that spike.
Also, "discard images when you scroll them offscreen" was disabled in (1) because it broke sites with lots of images (like Pinterest and Tumblr Archives, according to the bug).
These two issues alone make Firefox pretty painful for a lot of common use cases on real-world machines (like my old web browsing machine with only 2GB of RAM).
(0): https://bugzilla.mozilla.org/show_bug.cgi?id=847223 (1): https://bugzilla.mozilla.org/show_bug.cgi?id=872235
They fixed the memory leaks a long time ago guys... One of the big one was each extension have a copy of the DOM tree I believe.
And firefox is the only webbrowser I know that can handle 100+ tabs. >___<.
Aurora is very very stable btw.
First it was the memory leak issues that were there for a long time, then it was the feature creep of things I didn't care about, then it was Mozilla's stances on formats and junk that seemed a bit to self-interested, and then Mozilla making their own OS instead of just making webOS or Android more awesome.
It's weird, but I get the vibe that Mozilla is sort of selfish and is becoming less relevant over time. That is a stark contrast to Firefox making the web awesome back in the pre 1.0 days. Maybe they haven't changed and I have, but it's kind of odd
Their experiment with FxOS will repay itself as their API will make way into a W3C standard for mobile connectivity.
No thank you.
I like Firefox more and more. Their memory leak issues have been fixed. Their memory leaks in add-on have been contained. Mozilla make other tech besides browsers to advance the state of computing - Persona, Rust, Thunderbird, PDF.js, asm.js, FirefoxOS, etc. All incredibly useful and innovative. All these benefit most people and advance the web.
Funny, I'm starting to swing the other way. I keep wondering how we let a web advertiser on all our boxes.
The dominant browser shapes standards. It was true with Navigator, it was true with IE. Decisions on cookies, tracking, even some protocol designs over the next ten years will occasionally pit the interests of users against the interests of advertisers.
I like Google, a lot. But I do not trust Google to lobby on my behalf on these issues.
(Also, I really miss Ubiquity.)