Slashdot Mirror
Firefox 9 Released, JavaScript Performance Greatly Improved
MrSeb writes "Firefox 9 is now available — but unlike its previous rapid release forebears where not a lot changed, a huge feature has landed with the new version: the JavaScript engine now has type inference enabled. This simple switch has resulted in a 20-30% JS execution speed increase (PDF), putting JaegerMonkey back in line with Chrome's V8 engine, and even pulling ahead in some cases. If you switched away from Firefox to IE or Chrome for improved JS performance, now is probably the time to give Firefox another shot."
Firefox - Too little, too late?
Too little: Doesn't sound like it, given the writeup of this release.
Too late: An install of pretty much any software is one click away. No software is too late - a later version can fix the problems of earlier versions. Most users don't have any problems with memory usage, don't care about how the footprint compares with this or that version of chrome etc.
Speed was only half the issue that drove people away, the actual rapid releases and incompatibilities with add-ons with these releases among other things.
RockMelt [techcrunch.com] now. Especially RockMelt is an interesting browser - it completely abandons geeky stuff like NoScript or Adblock but instead caters to casual, normal people and how they use the internet. RockMelt has online Facebook friends directly on the site, along with recent news and updates from all social networks. It lets you easily add social bookmarks to sites like Reddit and Digg, along with sharing to Facebook and Twitter. Most people have been saying how wonderful it is compared to Firefox. It's an browser that actual people want.
I thought that sponsored "Ask Slashdot" was a bit much but now we have sponsored first posts?
And RockMeIt has much better astroturfing !
The Cloud - because you don't care if your apps and data are up in the air.
it completely abandons geeky stuff like NoScript or Adblock
Oh boy, now I can enjoy adverts featuring rotten teeth and modal popups that insist I "like" them on Facebook again!
Seriously, if I wanted to put up with this crap I'd go back to using IE.
Summation 2
...then you're probably still dealing with the fallout from that time when you switched your brain for a sponge
Who are these "most people"? I heard about RockMelt a year or two ago. I think it was Windows only which wasn't a good start, especially given that I think it was based on Firefox. It's "interesting", but I thought it was long dead. Most sites where you'd want to share something already have share buttons for social networking sites.
Adblock isn't geeky. Nobody likes ads. Apart from you perhaps, since you work in marketing.
which is totally what she said
Actually the last 3 releases all had some improvements in the memory department (I think 8 had the most improvements) and it looks to me like, there is more to come.
New things are always on the horizon
.... you don't even have a rough idea of how big the changes are , whether there will be compatability issues and so forth. I'm sure the coders have done a good job but whatever marketdroid imbecile thought that every new release must have a major version number markup should frankly be shot. And then forced to use IE 6 for the rest of his days.
Have you tried disabling 'smooth scrolling' (Options/Preferences -> Advanced -> General-tab -> Browsing: use smooth scrolling), I personally don't like it.
It is on by default, someone thinks it is a feature.
New things are always on the horizon
Is JS really that bad?
No JS is not that bad, it's just that sites are making more and more demands from it. Where once upon a time a site might have some simple functions and a few onclick handlers, now it's executing humoungous blocks of JS often tied to DOM calls. Look at apps like Emscripten for example or GWT which spew out a mass of JS code. The JS engine suddenly finds that the time it takes to parse, compile, garbage collect, execute and interact with the DOM suddenly makes a big deal of difference in performance when previously it might not have mattered so much.
The situation is bound to get even worse when tools appear which convert flash into HTML and HTML based animations with bloated JS runtimes of their own become increasingly common features on websites.
Looks like its actually just a Beta for Firefox 9?
Turns out that Javascript is expected to run faster than compiled assembly language and many people find the slowness to be an unbearable burden. So now Javascript is executed before it is downloaded to improve speed even more.
I don't think it's as bad as you make out. I get the impression that version numbers were to be depreceated and replaced with the terms Beta; Aurora and Nightly. Features would be mentioned as landing on Nightly/Aurora, appearing in users' browsers in so many weeks time. Releases themselves (every 6 weeks) aren't news in themselves. If Firefox developers communicate this clearly to reporters, then perhaps perceptions will change.
If users would still benefit from version numbers (e.g. for tech support), then I have a suggestion to make:
Next year, Firefox will be releasing version 12. On that version, there's the option of transitioning to a date-based system, with major versions following the year, and minor versions being incremented every 6 weeks. After version 11, the 1st release with this format would be 12.1; the 2nd release, 12.2; and so on. Here's how it looks like in practice:
* 10.0 January 31, 2012
* 11.0 March 13, 2012
* 12.1 April 24, 2012
* 12.2 June 5, 2012
* 12.3 July 17, 2012
* 12.4 August 28, 2012
* 12.5 October 9, 2012
* 12.6 November 20, 2012
* 13.1 January 1, 2013
Switching to a date-based system has the advantage that users will know what the current version is without having to report it, as the year corresponds to the version. Firefox in 2012 would be referred to as version 12. Reporters would focus on new and upcoming features in Firefox primarily, so that stories have a talking point and posters' comments are pertinent, primarily focused on features and improvements.
An example of an open source group who uses a similar format is Ubuntu (who base the version on the year, and the minor version on a 6 month schedule). Versions matter with this format; but there's still a sense of progression. We know what the version will be in 3 years time - even if we don't know what the features will be. Now try to imagine what Firefox's version would be with the new system, compared with the old one.
Consider that this is an issue that would involve a minor change; would benefit users and reporters (reducing confusion); and improve the quality of comments (on Firefox itself), then I think that Firefox developers will be pleasantly surprised with the results.
If they do want to focus more on development than on numbers, they would benefit by switching to a date system. I hope that some of the Firefox developers read this, as the value of changing merits the effort involved.
Firefox > help > about> "firefox 8.0 checking for updates... firefox is up to date"
www.getfirefox.com
good news your firefox is up to date
tfa
Firefox Beta Release Notes
(1.21 gigawatts) / (88 miles per hour) = 30 757 874 newtons
This particular behaviour at least is configurable: set image.mem.discardable to false. (Or, if you decide you prefer a trade-off, lengthen image.mem.min_discard_timeout_ms.)
Too late: An install of pretty much any software is one click away. No software is too late - a later version can fix the problems of earlier versions
That's only true if some other software didn't already fix their problems first. A significant number of users have already switched from Firefox, only being as good as Chrome isn't enough to get anyone to switch back.
We hope your rules and wisdom choke you / Now we are one in everlasting peace
Most people which install Chrome just install it because Google is a known brandname to them. And Google pretty much is the Internet to them, so they might as well install the Google sanctioned browser.
New things are always on the horizon
Firefox - Too little, too late?
Too little: Doesn't sound like it, given the writeup of this release.
Too late: An install of pretty much any software is one click away. No software is too late - a later version can fix the problems of earlier versions. Most users don't have any problems with memory usage, don't care about how the footprint compares with this or that version of chrome etc.
But it breaks Firefox's major original selling point: extensions. After Firefox 5 the extensions were supposed to be auto-updating in theory.
In theory, there is no difference between theory and practice. In practice, there is.
It is dangerous to be right when the government is wrong.
I just ran something I'm working on in Nightly. The first thing it does is to determine what it can about the client.
var comp; ... }
for( comp in window.navigator ) {
The code above should give values of comp for the keys in window.navigator, but comp remains undefined and the code raises an exception when window.navigator[comp] is examined in the loop. This would be the loop that should fail to execute if there are no values for comp.
I can't say anything about the release, I don't use windows so it probably isn't available to me, but Nightly is broken.
That's because of the memshrink project (earlier report on /.). You can read a weekly status report on Nicholas Nethercote's blog.
Another project that's recently started is called 'Snappy', which aims to increase the responsiveness of users' interactions with Firefox. There's a thread on Mozillazine tracking updates on Snappy.
All the people I know installed Chrome for one entirely different reason: Speed.
Chrome is so much faster than Firefox and doesn't use nearly as many resources.
- Don't do what I do, it's probably not healthy nor safe. -
Lol.. how long did the ad firm work to come up with that last line?
I don't really get all these extension complaints. Not once since FF4 have I had an extension fail on update, that I'm aware.
More importantly, some of the extensions I use have absolutely no acceptable substitute in chrome (never mind other browsers), leaving me completely baffled as to why people change just because FF changes their version number at a different pace (though I agree that it is a silly and pointless move).
I meant normal every day endusers, no geeks.
Geeks only have geeks as friends ;-)
New things are always on the horizon
Actually, he responded to me a few times the other day. I'm still unsure whether this is actually his job, or he's just an idiot. The fact that he said he works in advertising probably does mean that at least some of it is part of his job.
which is totally what she said
Yep. IE9 with the tracking protection / privacy filter lists from AdBlockPlus, and it works nearly as well as on Firefox (a little harder to configure, though still easy to turn on or off for a given site).
The fact that people are willing to put up with severely ad-laden sites always amazes me.
There's no place I could be, since I've found Serenity...
I agree that Chrome is more responsive that Firefox. (Note that I didn't say "faster"...) But to say that it uses less resources is bollocks. Chrome consumes vastly more memory than Firefox and I have less than 10 tabs open. Go ahead, browse for a day and measure it; the total memory usage of Chrome tops Firefox by quite a bit. The UI responsiveness is the only reason why I use Chrome over Firefox.
Chrome might feel snappier, but for some time Firefox uses less memory.
If he is in marketing don't discount the "just an idiot" angle.
Some of what I say is fact, some is conjecture, the rest I'm just blowing out my ass...you guess.
Didn't make it, though something like 99% of the stuff on AMO is compatible. The problem will still be third party plugins (better called "malware" IMHO)
because FF changes their version number at a different pace (though I agree that it is a silly and pointless move).
The version number changes quicker because the releases happen quicker. That's not hard to understand.
I don't use windows so it probably isn't available to me
I don't get what you mean here. Firefox is most certainly not a "Windows-only" product.
but Nightly is broken
It breaks often. That's why there's an Aurora and Beta in between before you even get to a release.
Well, know me: I have switched to Chrome because FF still hasn't implemented a minimal tabs-on-title interface on Linux. That wasn't relevant back when I had a 4:3 monitor, but with a little 16:9, every tiny row makes a difference. Plus I switch tabs a lot and everything sitting on the edge of the screen is simply easier to activate, especially since I like my mice fast. It's not a huge issue, but then again FF and Chromium aren't that different right now - they're both fast enough and use similar resources, so a tiny edge is all one of them can win by.
How is it possible that the default installation of Firefox does not include a setting to disable plugins? Opera and Chrome have this nifty feature where you can start plugins individually by request.
Sure, then leave it open overnight. Firefox leaks like a sieve.
I never close firefox on my living room computer. It's been up for two weeks since I last did a reboot, and it's currently using up 256.6 mb of RAM. Doesn't seem like it's leaking.
I am seeing that it spawned 26 threads, which looks excessive. I assume it's a thread pool for when they actually need it, and that they're not really active right now.
Warning: Opinions known to be heavily biased.
No: in fact, very few runtime-typed languages support type inference. What happens instead is that the value gets tagged with a type. E.g., in Python, when I type "x = 3", the variable x has no type attached to it, but the value 3 has the type 'int' attached to it. When the system needs to know type information, it queries the value.
Type inference is a little bit of a hard thing to do in runtime-typed languages. Not impossible, but ... interestingly wacky. Basically, the runtime environment has to be able to make a formal mathematical proof that "there is no code path in which this variable can point to any type of value except an 'int'", or what-have-you. If it's able to do that, then it might be able to optimize access to that value in ways that normal runtime typing can't. E.g., if you know something is always going to be a 32-bit integer, why not store it as a native primitive, rather than wrap it up in an object and have all the associated object overhead? That sort of thing.
Hope this helps!
Type Inference in 9.0 wasn't a security fix. Reducing memory usage by 30-50% in 7.0 wasn't either. (I'm sure the other releases had improvements other than security fixes, just nothing that interested me personally).
In general, anytime code has to be generated by a machine, I'm skeptical. That stuff is going to be absolutely unmaintainable in 10 years. And the tools that generated that code will likely be long gone.
I agree that JS is being used for purposes for which it was never intended. For robust, maintainable applications, you need a strongly-typed language. It seems to me that Java Applets would have eventually filled this space perfectly (a strongly-typed language capable of interacting with the DOM). Unfortunately, Applets were introduced about 15 years too soon (and pretty much killed off by Microsoft), and thus suffered from perceptions of poor performance. And now that Java is owned by Oracle, that opportunity is gone forever.
New releases usually are though.
Bollocks are they, check the FTP server ftp://ftp.mozilla.org/pub/. I've just downloaded the 64 bit version of Firefox 9.0 for Linux from there.
It's often due to buggy plugins. The problem with Firefox's single process model is if a plugin or Firefox leaks, you often can't free up memory by closing windows or tabs. You have to close everything.
In contrast even though Google Chrome might actually leak more, you can usually just close the offending tab, and the memory is freed up. You can even reopen the tab without having to log in again. So if a page gradually leaks memory, you can close it once it gets to big and reopen it again. All without losing sessions in your other tabs/windows.
One might say "Don't use those plugins then". But without those plugins there might be no other reason to use Firefox instead of some other browser.
...can I just expect all the sites already bloated with slow, broken JavaScript to just increase said cruft by 20-30% to take advantage of this?
Liberty in your lifetime
It's about time, the performance of Firefox has suffered greatly the last few version. even the right click on a link to show the drop down menu takes several seconds on a 2 GB mac mini (2009 model). At the moment Chrome is the best choice performance wise, but I prefer Firefox.
dude, there is somehing seriously wrong with your pc.
Wealth is the gift that keeps on giving.
This is why I love Tree Style Tabs. You get the tab bar on the left (or wherever else you like it), tabs structured hierarchically, collapsible trees and all that fancy stuff, including vertical screen estate.
Ignore this signature. By order.
Do you happen to use any plug-ins with it? The problem may be with a third party software and not Firefox.
Rules of Conduct:
#1 - The DM is always right.
#2 - If the DM is wrong, see rule #1