Firefox 3 In Alpha

illeism writes to note that, a mere six weeks after the launch of Firefox 2, Firefox 3 is now available in alpha. CNet reports that it is currently recommended only for software developers and testers. The big change is the upgraded Gecko rendering engine (the UI is unchanged from version 2). Because of the new Gecko code, this release will not run on Windows 95, 98, or ME, or OS X 10.2 or earlier. From the CNet article: "Firefox 3 will include some significant changes. It uses version 1.9 of the Gecko rendering engine — which itself hasn't been released yet but which includes the Cairo graphics layer. Gecko 1.9 has been in development since before the release of Firefox 2, and it provides vector-based rendering on all platforms. As the Gecko 1.9 road map explains, Cairo will 'bring modern, hardware-accelerated 2D-graphics capabilities to the whole of the Web without requiring proprietary plug-ins or rendering obsolete the broad and rich set of Web-authoring techniques developed over the past decade.'"

It's not really a Firefox alpha

[davidmcg (796487)]

It's more of an alpha of Gecko [browserden.co.uk] than of Firefox. There's no front end changes in this release, all the changes are to the backend which are shared with all Gecko browsers (Camino, Seamonkey and other Gecko apps like Thunderbird).

Development has been going on the trunk since the Gecko 1.8 was branched (sometime in 2005) - Gecko 1.8 was the basis of Firefox 2 and 1.5. So there's a lot of backend work been going on that's not been tested by a wider audience. While lots of frontend changes were made on the branch for Firefox 2, most of the backend work was restricted to the trunk.

Future alphas and betas will have more UI changes in them so can more accurately be called Firefox alphas.

Re:

[MemoryDragon (544441)]

This is more than ie7 was, ie7 was a frontend change with only a handful of bugfixes in the backend, and even the top 10 list of worst bugs was not fully fixed.

Re:

[CubicleView (910143)]

OK, the fact that all the "new" features in IE7 were implemented in the front end was news to me, thanks for the info. However would it not be better to argue how this release fits in the generally accepted definition of an alpha release http://en.wikipedia.org/wiki/Software_release [wikipedia.org] rather than flogging a dead horse, just because it happened to be in the same field?

Front End? Hardly

[shaneh0 (624603)]

Adding tabs was a huge change to the IE application model.
The rendering engine was updated for efficiency and standards compliance (which is much better now, if still not yet where you'd like it to be)
Things like anti-phishing, new security models, and a new plug-in interface are features that 'go down to the metal'

IE7 was very substantial. I'm writing this on FF2.0 and I have to say: The IE7 upgrade was far more successful than FF2. I still believe that Firefox is a better browser over all, but not by very

Re:Front End? Hardly

[Andrewkov (140579)]

You can configure Firefox 2 to have a single tab close button, like in 1.5.

Go into about:config, change browser.tabs.closeButtons to 3.

Re:

[spitzak (4019)]

Thats the "middle mouse button" for those few left who speak english.

The fact that they made the wheel also be the middle mouse button is clever, but in no way required. There are mice available that seperate the functions.

Quick Find

[Robmonster (158873)]

Put the following in your userChrome.css to revert to the old Find Bar:-

/* Use the old-style / and ' QuickFind Bar behaviour */
#FindToolbar > * {display:-moz-box}

Re:Front End? Hardly

[somersault (912633)]

They could have the phishing philter and spell checker as addons which are installed by default - with the option to not install in a 'custom' install, like you can do with the DOM inspector. Then you can remove them if you don't want them. Customisation is always fun :)

will not run..

[Threni (635302)]

> this release will not run on Windows 95, 98, or ME, or OS X 10.2 or earlier.

That's nothing. IE7 doesn't even work on Windows 2000!

Re:

[Billosaur (927319)]

When you boil it right down, anyone using one of the older versions of Windows (and I count 2000 in this, as MS doesn't support it anymore) is going to have to face up to the fact that technology advances, software changes, and no matter how much they love their old machine/OS, they're going to get left behind. Backwards compatibility leads to backwards thinking.

Re:will not run..

[Threni (635302)]

> face up to the fact that technology advances, software changes, and no matter how much they love their old machine/OS, they're going
> to get left behind.

I don't love my old OS, but I have to use it (sometimes) at work because it's the OS that deployed apps use. No point in retesting huge apps on different OS's just to get a new browser. It doesn't bother me - I now use Firefox on those machines anyway. It seems a little odd, though. Aren't browsers just displaying text and graphics, and running scripts? (I don't include plugins such as Flash and Qtime as the run as seperate executable code invoked by the browser).

> Backwards compatibility leads to backwards thinking.

Hmm. You could also write "Pointlessly adopting new technology leads to the need for frequent bug fixes and faster CPUs".

Re:will not run..

[Captain Splendid (673276)]

Shame really, 2000 is a decent OS (and I'm still going to have it around for a bit.)

But I guess it's time to start getting on the horse with Linux, because it's also the last MS OS I'll be using.

Some people have to

[DrYak (748999)]

Some people HAVE to use old windows, because the old proprietary controlling software that came with a given hardware (say a robot in bio-medical lab), only runs on old OS (I've even seen spectro-photo-meters that only run on DOS. Yeah. Thank goodness FreeDOS [freedos.org] is our friend in such deprecated cases). The company has dropped support for newer OS for this peice of hardwre and is only doing hardware repairs. You either have to keep a deprecated OS for your machine, or you have to buy a newer model (Which most o

Re:

[Billosaur (927319)]

When will people learn that you have to buy a completely new machine and set of applications every couple of years!?

About the same time that people get rid of their AS/400 machines.

Too bad

[Rosco P. Coltrane (209368)]

Because of the new Gecko code, this release will not run on Windows 95, 98, or ME, or OS X 10.2 or earlier.

One of the great strengths of OSS compared to proprietary software is the ability to make use of older hardware. Not so with this new release of Firefox. But then it's the same with other "heavyweights" like KDE, so I guess there's a trend there. That's too bad...

Re:Too bad

[linuxci (3530)]

It's because cairo is not compatible with win9x they decided that it wasn't worth the effort to support this platform any more (they still support Win2k - they only dropped support because supporting win9x was holding them back). If anyone is able to contibute coding skills to make it work they have no problem accepting it. It's a technical rather than a political decision.

Re:

[swthomas55 (904301)]

Some "real world" stats:

According to the stats we collect at www.jstor.org, Win98 accounts for 1.4% of our hits (appx 2.1 million out of 151 million so far this month), but they account for only 0.6% of the Firefox users. Both Win95 and WinME are below 0.1%.

Mac OS X (all versions) is about 9% (the user-agent string, which is what we use for this analysis, doesn't distinguish versions of OS X, so I don't know how many of these are 10.2 or earlier).

Re:

[mr_mischief (456295)]

People worried about security updates for their browser shouldn't be using operating systems that get no security updates -- namely, Windows 95 and Windows 98.

There are a number of security issues in your OS that no browser can smooth over.

On the other hand, Mandriva 2006, Ubuntu 6, Xandros 3, NetBSD, and Mac OS X 10.2 all run reasonably well on 500Mhz systems with 256MB of RAM (albeit OS X on a 500Mhz Mac instead of a PC). I haven't run Solaris, OpenBSD, or FreeBSD much recently, but I'd bet you could get

Re:Too bad

[makapuf (412290)]

Why ? To each program its own target.
KDE has never been "for older hardware". However, perfectly nice & actively developed Desktop Environment exist for older hw (xfce by ex.).
Same here, OpenSource is about making use of older codebase, so nothing prevents anybody to patch FF2.x !

Re:Too bad

[Rosco P. Coltrane (209368)]

Yes, I guess you're right. There was a time when OSS software was the solution of choice for those who didn't want to throw away semi-obsolete hardware in working order to dance the Microsoft forced upgrading dance. I suppose this means OSS solutions have gained enough traction and have become credible enough that they justify requiring newer hardware to run them, which is good.

I'm aware of xfce and blackbox and the likes, they are nice, but if you want to run mainstream software that require KDE libraries, you're still hosed.

But in the case of FF for Windows, the problem is that Win9x users (and there are many left) will find themselves in the same situation they were with IE: they'll have to keep running the latest older version of the browser that works with their OS, which will quickly become out of date. I'm sure the FF/Gecko guys have perfectly good technical reasons to leave the old platform behind, but in a sense I hope someone will fork off a Win9x tree of FF and keep developing it, otherwise it would mean OSS is no better than Microsoft with regard to software obsolescence.

Re:Too bad

[aussie_a (778472)]

Yes, I guess you're right. There was a time when OSS software was the solution of choice for those who didn't want to throw away semi-obsolete hardware in working order to dance the Microsoft forced upgrading dance. I suppose this means OSS solutions have gained enough traction and have become credible enough that they justify requiring newer hardware to run them, which is good.
 
I'm aware of xfce and blackbox and the likes, they are nice, but if you want to run mainstream software that require KDE libraries, you're still hosed.
 
But in the case of FF for Windows, the problem is that Win9x users (and there are many left) will find themselves in the same situation they were with IE: they'll have to keep running the latest older version of the browser that works with their OS, which will quickly become out of date. I'm sure the FF/Gecko guys have perfectly good technical reasons to leave the old platform behind, but in a sense I hope someone will fork off a Win9x tree of FF and keep developing it, otherwise it would mean OSS is no better than Microsoft with regard to software obsolescence.
 

So how long are they suppose to be supporting the Win9x OSes? 2 more years? 5? 10? 20? Until there aren't any more Win9x users? But if all of the Win9x users have their OSS software continue to support Win9x, what incentive do they have to upgrade? They obviously don't care about bugs or viruses if they're still using Win9x software after all these years.

Re:

[jlarocco (851450)]

So how long are they suppose to be supporting the Win9x OSes? 2 more years? 5? 10? 20? Until there aren't any more Win9x users? But if all of the Win9x users have their OSS software continue to support Win9x, what incentive do they have to upgrade? They obviously don't care about bugs or viruses if they're still using Win9x software after all these years.

Until there's a good technical reason not too? It's not your responsibility to give people incentives to upgrade. In a lot of cases it makes more sen

Re:

[Orange Crush (934731)]

I suppose this means OSS solutions have gained enough traction and have become credible enough that they justify requiring newer hardware to run them, which is good.

Granted. The lowest specs I'd bother installing Win2k on are a 500mhz P3 w/ at least 128mb ram. This is hardly cutting edge. Grandma's 486 DX2-66 running Win95 (oh the pain . . .) is both a.) miraculous its power supply hasn't failed or a cap busted on the mobo and b.) probably much better off running Damn Small Linux anyway.

Re:

[ben there... (946946)]

But in the case of FF for Windows, the problem is that Win9x users (and there are many left) will find themselves in the same situation they were with IE: they'll have to keep running the latest older version of the browser that works with their OS, which will quickly become out of date.

If lots of people run Firefox on old PCs, there will be lots of people to develop patches for Fx 2.x.

It works the same as any open source project. The more common the scenario of your use of the project, the more likely lots

Re:Too bad

[Dr. Spork (142693)]

I have a feeling that the 2.x.x branch of Firefox will live on for a very long time, and will continue having bugfix and security updates. If you're running Win98, it will certainly not be the weak point in your system in terms of security or stability! My point is that if by your standards you consider Win98 good enough to use, there will always be a version of Firefox that far exceeds your standards. And I mean, by many miles.

Re:

[kimvette (919543)]

In some people's minds, an extensible system where you can selectively add everything and the kitchen sink by your own choosing is bloat, even if the minimal installation is extremely efficient.

I prefer to call that level of choice flexibility.

Re:Too bad

[MountainMan101 (714389)]

One of the great strengths of OSS compared to proprietary software is the ability to make use of older hardware.

This doesn't happen automagically when you license something by GPL (or similar). It takes work. The strength of OSS is that no one is stopping you from making it work on older hardware. All the code for older firefox versions, and the code for gecko 1.9 is available. Just because Mozilla team is dropping support doesn't mean they won't add patches to fix this if someone else does it. Now compare that with say Windows Vista - you have no way of patching that to run on an old 386.

Moral of the story... don't be so quick to bitch about stuff.

Re:Too bad

[ari wins (1016630)]

While I understand your gripe, let me introduce you to Firefox 2.0. It was just released, and likely going to be around for a long enough time to outlast your computer with the P200 chip w/MMX technology that still has windows 98 installed.

Re:

[Runefox (905204)]

Well, Win2K will run quite nicely on a Pentium 166 with 64MB of RAM, so this point is moot.

I don't know about you, but I'd be embarassed to even be running that, let alone something older. You'd have to be running a 486 to be incompatible with Firefox at this stage of the game, and even then, Linux will still feel right at home, and - again - run Firefox.

So what's the problem?

Re:

[aussie_a (778472)]

Another one of the great strengths of OSS is that if there's enough of a demand for something, people can work on it without any legal worries. This is discouraged in businesses that concentrate on maximizing profits.

Re:

[Junior J. Junior III (192702)]

Proprietary software can support old hardware, too. They just choose not to.

OSS software can support old hardware, as well. More often, they choose to. But not always. Why can't I run Firefox on Commodore 64 or an Altair? Because I haven't downloaded the source code, written the missing parts that would enable the trunk code to be ported to $myplatform, and recompiled it.

You want legacy hardware support? If you're one of the few people still using something that old, and no one else wants to support i

Re:

[Lord Ender (156273)]

What? You mean it doesn't support MS DOS 3.0 anymore? BOO FIREFOX!

Cairo

[astralbat (828541)]

Glad to hear that the rendering will now get some hardware accerlation. Does anyone know how faster this will be? Will it lead to smoother scrolling as on my Linux machine 'smooth scrolling' is very jerky - especially so with flash adverts.

Re:Cairo

[jcupitt65 (68879)]

http://mail.gnome.org/archives/performance-list/20 06-December/msg00006.html [gnome.org]

Acid2

[savala (874118)]

Before someone else brings it up, no, this doesn't pass acid2. Purposefully, as the build from two days later does. This Gecko alpha (not Firefox alpha) was released so there'd be a good reference for people to test with before several rather major changes were landed on trunk, one of which was the reflow branch that made Gecko pass the acid2 test.

What does this mean?

[aussie_a (778472)]

Sorry, I left my marketspeak chip in the other bot. Can anyone translate this to English:

Cairo will 'bring modern, hardware-accelerated 2D-graphics capabilities to the whole of the Web without requiring proprietary plug-ins or rendering obsolete the broad and rich set of Web-authoring techniques developed over the past decade.

Without my marketspeak chip I translate that to say "We'll be doing things faster and still supporting HTML." Please tell me I got that wrong. Do they really need to specify "We support HTML" in a browser engine? And even if they did, why did they need to translate it to marketspeak? Do they have a brand new marketing droid they just couldn't wait to use?

Re:

[modeless (978411)]

It means that they're making the graphics of AJAX apps faster to better compete with Flash. That way the kinds of things that used to require Flash can move back to HTML + SVG + JavaScript. In the future we should see a lot more Google Maps style interactive HTML applications, where it becomes meaningful to talk about the "frame rate" of your web page. Firefox will achieve high frame rates using hardware graphics acceleration provided by Cairo. Today, your GeForce 8800 with X hundred million transistors

performance improvements

[diegocgteleline.es (653730)]

This version is much faster and resource friendly - opening a test google spreadsheet page went from 52 MB of RSS to 43, and almost 4 seconds less to render it.

Lots of javascript benchmarks are faster too (depending on the benchmark - other parts are slower)

Gecko 1.9 has been being developed for a long time (the "reflow branch" is 2 years old it has been said!) so I guess it's expected that it improves things so much!

Cairo is kind of slow now

[drgonzo59 (747139)]

I have tested Cairo for my project and at this moment it is slower than I needed. I was looking for a good canvas to draw a graph on, but I had to settle for something else. I like the features of Cairo, the idea of being able to render to PDF or to the screen, and so on, but it is just not fast enough. Perhaps the attention from the Gecko engine will get some more development going on the Cairo side as well...

Re:

[Dr. Spork (142693)]

Yeah, I totally agree. I'm also hoping that Gecko coders will turn their attention towards optimizing Cairo, because its current performance is unacceptable. According to this benchmark [blogspot.com], Cairo's rendering performance isn't just somewhat slower than its open-source rival Qt. It's something like 700% slower. If that doesn't improve dramatically before Mozilla's 3.0 release, it will account for dreadfully many wasted CPU cycles.

I understand the decision to go with Cairo, but like you said, I hope it's couple

Re:

[ThisNukes4u (752508)]

You should try the 1.3.x preview release series on cairographics.org. There are a whole bunch of performance improvements, including a new tessellator. Also, cairo's performance on linux is heavily dictated by how well your video card driver supports XRender. I have found that r200 radeons with the new EXA driver acceleration mechanism accelerates cairo, among other things, quite nicely. If you can't use a driver that supports EXA, you can try rendering to a image backend first(which forces software fal

Mac version faster

[shaper (88544)]

Recent nightly builds for Mac OS X feel much snappier than Firefox 2.0. One of the obvious culprits is that Cocoa widgets [mozilla.org] are now used on Mac OS X builds. I don't know if there are other changes affecting the performance on Mac OS X, but the difference is fairly dramatic.

I love Firefox on Windows, but I have stuck with Safari on the Mac because Firefox has always felt porky and slow compared to Safari on the same hardware. The newer builds of Firefox 3 for the Mac are much better: windows, tabs, menus and other user interface elements have a nice immediate feel to them. And the page rendering is more performant than Safari on certain Web 2.0 type sites like digg and Slashdot's new discussion system. It's buggy alpha code, but early indications seem to be good for a nice improvement on the Mac when Firefox 3 comes out.

Re:

[linuxci (3530)]

It will in alpha 2.

Re:

[linuxci (3530)]

There'll still be firefox 2 which will have updates a while after 3.0 is released. That's still a major improvement over IE6 which is their only option.

Re:

[by Anonymous Coward]

>> It sounds like something is wrong with that.
Yes. What's wrong is having users who still scream for compatibility with their old OS. XP was out in 2001. Win2000 was out in '99. That's 7 years. I really doubt much software when Win2000 was RTMd was still compatible with Windows 3.0 of 1992...
For how many years should we cripple innovation in open source projects just to support DOS 3.3 on 286 ?

Re:

[Lisandro (799651)]

Unfortuantely Opera 9 is too unstable.

I keep hearing this... don't know why, but in my Linux box, Opera 9.02 is rock solid - it haven't crashed once since i installed it. I experienced a couple of crashes back then with O8 though, but the session management (restores your session completely after a crash) rendered them relatively painless.

I must say all versions of FF i've tried were perfectly stable aswell, but the insane memory requierements (among other peeves) prevents it from being my main browser.

Re:

[Max Romantschuk (132276)]

Even though I have 2 gigs of RAM, FF routinely escalates to over 1 GB of usage (I have a lot of tabs open, but c'mon), and hogs so many resources that the system slows to a crawl that requires a reboot.

IMO, this issue should have been the #1 priority in the move from 1.5 to 2.0. I am losing hope it will be fixed in the next release either (which is a shame, since I would prefer to run FF). IE7 and Opera9 simply do not have this problem to the extent FF does.

While I wholeheartedly agree I've found that using

Re:

[matlhDam (149229)]

But will it pass Acid 2?

My understanding is that this alpha won't, but the next alpha should. The reflow refactoring branch [mozilla.org] was merged back onto trunk [mozillazine.org] recently -- this is a rationalisation of the layout code that fixes a lot of bugs, which also gets Acid 2 rendering properly.

Re:

[vally_manea (911530)]

No, it didn't RTFA you linked.

Also, the recently released Firefox 3 Alpha 1 does not pass the test

Re:If it means decent Zoom...

[Dan Ost (415913)]

Go into your FF preferences. Select the 'Content' preferences and look at the advanced font settings. You'll see a minimum font size setting. Change this to be however big you need the font to be and behold, your problem is solved.

You're very welcome.