OSNews

"""
Anybody using the nightlies will already be aware of the noticeable improvement in speed,
"""

Not to disagree, but I have been watching Mozilla development from the day that Netscape released their code. And I swear that for 3 years whenever anyone criticized Mozilla in any way, the response from the community was "Use the latest nightlies! They're great!".

Something horrible must happen to the "wonderful" nightly code when it is released in a stable version.

It's weird.

Edited 2007-05-18 20:04

Not only that, but Firefox became freakin' slow as time passed by. Mozilla Seamonkey, having been kept up to date by only a handful of developers - not sponsored by Mozilla Corporation, it seems - in order to "backport" some of Firefox niceties such as AdBlock+ and often accused in the past as being sluggish, is WAY faster than most of the recent FF iterations.

The only reason that I still use Firefox instead of Opera is that I prefer the way how FF rendering engine tries to work around broken websites compared to Opera's and the fact that the Opera Adblocker is not nearly as good as Adblock+ but anyone comparing FF and Opera speedwise and claiming that FF has an edge must have his or her head examinated. Not to mention that Opera's cache management knows no rivals.

Opera lacks the mindshare, yes. It could also benefit of an extensible API so that its users can build the nice things that the community around Mozilla built on top of Firefox... True. But Opera definitely surpassed Firefox some time ago.

In my sister's slower machine, the choice is a no brainer. Firefox definitely needs to go back to its roots if it wants to keep the lead over the competition since even MS seems to be waking up and getting their stuff up to the snuff!

Edited 2007-05-18 20:40

Actually, that is the case, IIRC.

Nightly builds are usually taken from an active-development tree (where devs can 'freely' make changes to the code (once meeting certain criteria), as patches).

Stable builds are taken from a fine-picked assortment of the code submitted in between releases, and resides in its own tree, making it possible to have vast differences between the two trees.

That and the difference in compiler options in stable builds, extra code (feature) patches applied, etc...

It can get fun finding the best nightly to run, though, so I'd stick to the stables if your unsure :-)

--The loon

I for one have gone back to using Safari as my browser of choice. There have been content rich pages (like the World of Warcraft Armory pages) which have caused Firefox to choke and consume 100% of my CPU speed for 5 - 10 seconds before successfully loading the pages.

Just taking this article as an example. Safari uses about 10% CPU power to display the page and the adverts. Once the animations complete, CPU usage drops to 0 - 3%. In contrast, Firefox 20 - 30% CPU usage to display that page and when the animations are done, CPU usage drops to about 5 - 10%.

This is really sad, as I much prefer Firefox since more pages render correctly.

It's probably Flash. Flash + FF is not yet a terribly friendly pairing. There's one of those stupid intel quad core ads (the one with the leopards) that causes my machine to hit 100% CPU usage everytime. I couldn't, easily, test this on IE as browser detection on the server side was returning a single GIF to IE but a .swf to FF. No idea why. I gave up at the point and just install FlashBlock!

That's because you need a quad core to see the ad.

It's not just the webpages though. I still have an old version of Phoenix installed (0.8) and it runs circles around Firefox. Sure Firefox does a lot more now, but it comes at the expense of speed, let's not pretend it doesn't.

Firefox has always been a 'Fat Elvis'.
I find it amazing that people believe it's a 'light browser'. But I do agree that web pages are getting way too bloated in general.

There are so many web pages that have javascript and I can never work out what it's being used for.
For some reason people forgot that a webpage is just a document with hyperlinks. I blame graphic designers and programmers those that never learnt real programming.

People also have to realise that Firefox runs on top of a cross-platfrom UI library, XUL. Firefox runs on OSes as diverse as OS/2, RISC OS and BeOS. Firefox cannot be as perfect as a wholly native browser, like Safari; ergo, one must cut Firefox at least a bit of slack.

People also have to realise that Firefox runs on top of a cross-platfrom UI library, XUL.

Yeah, if Opera was based on a multi-platform toolkit, it'd be slow, too!

Although admittedly Opera doesn't run on as many OSes as Firefox.

Trying to be funny? Opera runs on my phone and my Wii.

I hope you are being sarcastic because Opera is based on a multi-platform toolkit. It uses Qt.

I hope you are being sarcastic because Opera is based on a multi-platform toolkit. It uses Qt.

You're partially right, but it's not based on Qt, it's based on their own toolkit, (I can't recall the name). Most of the user interface is built on that, Qt is actually used for very little (printing is one of the major things, and only on Linux).

Also, in Skype, only the Linux version uses Qt. The UI for the Windows version is completely separate and doesn't use Qt at all. I always thought that was a hell of a strange way to use a cross-platform toolkit, especially one as nice to work with as Qt.

The Skype dev's said that they wanted the most "native" approach for every platform, which is fair enough I guess, but I think less important to normal users than we developers tend to think. Every second Windows app looks slightly different, but I've never heard any normal user complain about it.

Edited 2007-05-18 21:05

You can say the same for Opera
That takes alot more memory than Firefox, last time I looked it was 60Mb within a few minutes.

Opera uses much, much less memory than Firefox. Especially if you don't restart your browser every 15min.

At about 100 tabs Firefox easily breaks 1GB of memory and more (at least it did before I stopped using it), Opera stays below 300MB. Additionally, Opera loads faster, it's crash recovery is more reliable and somehow they managed to fit the browser and half a dozen apps I've never used in a download that's half the size of Firefox.

Opera's cookie managment is still a PITA and I miss a few features of Firefox like nuke everything (and Firesomething; the endless possibilities always put a smile on my face and I'm not even being sarcastic =).

Please don't say lies like this, it's totally false.

I'm a Opera user since early 9 version to the today's one (9.10) and I'm very unhappy with the experience on Opera. This occour to me even more than with Firefox.

I experienced a lot of bugs seeing certain sites, crashes that later deleted my bookmarks and tabs. I was unable to recovering those pages I needed to read, so again using google and wasting even some hours.

I had very bad experiences with most web browsers: Firefox and Opera both are quite buggy and corrupt bookmarks and unable to recover sites. Even both haver quite big rendering problems and slowdowns. On the other hand, Internet Explorer is even worser in most aspects.

I hope more renderers like WebKit or better are going to be developed, but that maybe cannot happen due to the web world becoming an annoying mass of bloated technologies. I think the web must evolve into a new standard with all the needed stuff but becoming simple and avoiding bloat.

Maybe the WebKit port to KDE could help the appear of kightweight browsers using toolkits like FLTK2. Sometimes I use Dillo because is quite fast, but it lacks most standards for becoming a really usable browser. It seems tkhtml2 is a dead project, it was promising at the time.

Edited 2007-05-18 22:05

I have to agree with SlackerJack, last time I *tested* it Opera 9.x did use more memory than Firefox. I did my own tests, opening a set of 'heavy' webpages in both browsers and monitored usage.

The fact is memory is irrelevent as long as you can spare it and it's managed properly by the application. *You* are correct in that Opera tends to *manage* it's memory alot better than FF, and in terms of UI, tab and webpage responsiveness I don't think Opera can be beat.

Now if only Opera would ship a 64 bit binaries for Linux (and other platforms) so I can start using it again. I am sick of dodgy fonts and an ugly UI on my otherwise beautiful pure 64-bit system. Give us a 64-bit build! We have nspluginwrapper, let us have it for the love of God...

Elvis was a bit before my time, but AFAIK, the guy was still a 'sex symbol' in his later years when he put on the pounds.

Proof there's hope for us guys with a little extra baggage.

"""
Proof there's hope for us guys with a little extra baggage.
"""

Actually, I think it was Ron Jeremy who eventually proved that indisputably. ;-)

Not great. Firefox on Mac has lagged a little, but is generally good enough for most users. I miss the lack of text services mostly. Firefox 3 promises cooca fields and decent speed improvements. To fix the fact that the theme looks like ass go to http://takebacktheweb.org

It should be noted that Firefox 3 will be using Cairo for graphics, which allows platform-specific rendering backends. Work is underway to provide a Quartz backend that should be able to provide hardware accelerated graphics through the native API. If that is delivered optimally, then we could be looking at a really smoking Firefox on Mac experience.

Whilst quality of code, as in your first sentence, is down to the team, and the engineer, you also have to bear in mind:

"By my count, the absolute most conservative view is that 100 paid people are supporting more than a dozen funded startups building extensions for Firefox, more than a half dozen funded startups building completely new applications, countless web-based application startups, and more than a million users per employee. All for less money than Adobe is putting into Apollo's launch. Way less.

That's crazy high leverage. As much leverage as any organization I know of in the world." - John Lilly

Mozilla are doing more in some areas than Microsoft, Adobe et al, with less. There hasn't [to my knowledge] been such debacle in Mozilla like Windows Vista, a product that the engineers, and especially the managers totally failed to get to grips with– Mozilla did their code reset back at the start when they inherited the absolute abomination that was Netscape 4- and that wasn't a multi billion dollar project with infinite resources.

edit: attribution

Edited 2007-05-18 18:49

Mozilla are doing more in some areas than Microsoft, Adobe et al, with less. There hasn't [to my knowledge] been such debacle in Mozilla like Windows Vista, a product that the engineers, and especially the managers totally failed to get to grips with– Mozilla did their code reset back at the start when they inherited the absolute abomination that was Netscape 4- and that wasn't a multi billion dollar project with infinite resources.

I'll agree that Microsoft really made some mistakes with Vista, but I think that primarily has to do with some bad technology bets that the company made (ie. using managed code in the Shell) -- and management led by Jim Allchin that was wholly disinterested and incapable of charting a course away from disaster until the cost was exorbitantly high.

Microsoft has a lot of capital -- but it doesn't have infinite resources, btw. They are notorious tightwads when it comes to team composition, hardware, etc. Part of the problem is that they have gotten so BIG over the past decade that they are increasingly incapable of prioritizing resource allocation across teams. For example, a less important team may have the same resources as a more important team -- and, if you try to change things, you will get tremendous political pushback from dueling management chains. Thus, Microsoft's biggest enemy is its own bureaucracy.

That's not true. OSS is generally higher quality eventually (perhaps not so POLISHED but higher quality internally, less bugs etc.), because of peer review.

There's an easy comparison:
Closed source is like a group of scientists working secretly for a company on something no one else will see until it's done, and even afterwards no one will be able to do any change on it.

Open source is like if those scientists put all their work, theories etc. in publications and many others joined in on improving it or criticizing (which is also important).

OSS has 2 problems one of which doesn't relate to firefox:

1. If the popularity of the project is low, peer review can be 0, thus the advantage goes away (not the Firefox case for sure).

2. Polish is missing because there's no one to shout "easy clickeys" at the devs, who find the interface etc. "good enough" while concentrating on the internals mostly. I think Firefox is past this as well.

Hope that explains why OSS is better in many cases.

That's not true. OSS is generally higher quality eventually (perhaps not so POLISHED but higher quality internally, less bugs etc.), because of peer review.

Well, that assumes that the many eyeballs are actually looking at the code -- which is by no means certain. I think that what tends to happen is that code which is most interesting and sexy (ie. Linux kernel) tends to get a lot of attention, while less interesting code languishes without any peer review at all.

1. If the popularity of the project is low, peer review can be 0, thus the advantage goes away (not the Firefox case for sure).

Yes, but the granularity is a lot finer than simply by project. There are parts of code in any project which are simply tedious and uninteresting to read; hence, they get a lot less scrutiny.

2. Polish is missing because there's no one to shout "easy clickeys" at the devs, who find the interface etc. "good enough" while concentrating on the internals mostly. I think Firefox is past this as well.

I dunno. Most software have really crappy interfaces. Firefox is no exception.

Many, many people say this, but there are a great deal of other factors that go into good software which arent taken into account.

First of all, there is the boring stuff. Bug fixing (intermittant bugs especially) is the worst part of any programmers job. Since there is zero financial motivation for the vast majority of open source devs, bugs which do not directly effect a core developer will sometimes take years to be fixed.

Secondly, you mentioned how unpopular projects will often not have any real peer review. The thing is, popular projects will frequently have the same problem! When it comes to a massive codebase that you have to learn, it takes a significant portion of your life to become involved. OO.o for example is one of the most widely used pieces of open software, but 99% of the work is done by people on SUNs payroll. Most OSS devs pick their poney, a kernel hacker will probably not be a regular in GNOME for example.

Third of all, the other thing developers hate is stupid users. Fixing bugs that can be avoided with common sense, making good interfaces, writing good documentation are all things developers hate doing, and usually will only do if they are getting paid. This isnt as big a problem as it once was, for the sole fact that we have so many big companies paying for the work to be done now. (if you have been using linux long enough, im sure you will remember this http://www.fh-muenchen.de/home/ze/rz/services/projects/xcdroast/e_0...)

Lastly, your analogy is a little off. Open source is like academia, closed source is big business. Lots of cool stuff comes out of the universities, and they are very important to our society as a whole, but to get the job done the quickest and cheapest way possible, the answer almost alwas comes from industry. This is why the GIMP is more then good enough for hobbyest use, but is still not an adiquet tool for a professional (even the artists hired by OSS companies still use photoshop). Or while audacity is good and all, ALSA is really not good enough for serious recording (i have found up to an 80ms latency difference between vista and feisty.). Or how OGL/SDL has improved tremendously over the years, it really doesnt come close to directX. I have been around long enough to have written more then my share of DOS and bash scripts, but I mean come on guys, its 2007. Lets get with the program and upgrade to what has been the standard interface for everything since about 1980. As much as apache is so much greater in so many ways then IIS, why is it that I MUCH perfer IIS while developing any sort of web stuff? It is because it is because its a few clicks to configure anything I want. Sure, I know how to use apache, I also know how to use slack, but just because I *can* deal with the added pain of working with text files and commandline interfaces, doesnt mean I *want* to.

I could go on, but I will stop here.

Open source development has alot of strengths, but it also has alot of weaknesses. The strengths are remarkably well articulated (ESR did a great job with the CatB articles), and are expressed regularily. The weaknesses however are rarely articulated by anyone except for outside competitors, and are typically then dismmissed as FUD.

First of all, there is the boring stuff. Bug fixing (intermittant bugs especially) is the worst part of any programmers job. Since there is zero financial motivation for the vast majority of open source devs, bugs which do not directly effect a core developer will sometimes take years to be fixed.

This depends on "itching". If the bug itches someone with the skill there's a quite good probability it'll get fixed unless it's some deep-shit kind of bug.


When it comes to a massive codebase that you have to learn, it takes a significant portion of your life to become involved. OO.o for example is one of the most widely used pieces of open software, but 99% of the work is done by people on SUNs payroll. Most OSS devs pick their poney, a kernel hacker will probably not be a regular in GNOME for example.

True but OO.o is a piece of bloated... you know what.

Third of all, the other thing developers hate is stupid users. Fixing bugs that can be avoided with common sense, making good interfaces, writing good documentation are all things developers hate doing, and usually will only do if they are getting paid. This isnt as big a problem as it once was, for the sole fact that we have so many big companies paying for the work to be done now.

Not necessarily true. It boils down to personal pride and feeling of accomplishment. I for one hate improper GUIs and wouldn't release anything like that out (at best in beta or such but never as the 1.0).

Lastly, your analogy is a little off. Open source is like academia, closed source is big business. Lots of cool stuff comes out of the universities, and they are very important to our society as a whole, but to get the job done the quickest and cheapest way possible, the answer almost alwas comes from industry. This is why the GIMP is more then good enough for hobbyest use, but is still not an adiquet tool for a professional (even the artists hired by OSS companies still use photoshop). Or while audacity is good and all, ALSA is really not good enough for serious recording (i have found up to an 80ms latency difference between vista and feisty.). Or how OGL/SDL has improved tremendously over the years, it really doesnt come close to directX. I have been around long enough to have written more then my share of DOS and bash scripts, but I mean come on guys, its 2007. Lets get with the program and upgrade to what has been the standard interface for everything since about 1980. As much as apache is so much greater in so many ways then IIS, why is it that I MUCH perfer IIS while developing any sort of web stuff? It is because it is because its a few clicks to configure anything I want. Sure, I know how to use apache, I also know how to use slack, but just because I *can* deal with the added pain of working with text files and commandline interfaces, doesnt mean I *want* to.

SDL/OGL is more than adequate if you also use OpenAL for audio and some other SDL minor libs together. It's not as nice package as Dx but it's about as powerful, it's that you have to mix it up. Not much problem for anyone who knows what he's doing (OpenGL is not less powerful than D3d, if you think that you have a problem).

As for the polish well it's as I said. There's no one to shout "bells and whistles" at the guys. But if I compare Windows XP programs I use to Linux versions, the Linux versions are less annoying, more to the point, and lately, easier to use. For comparison.. "WinRAR/WinZip/Winwhateverarchiver" vs the default gnome front-end archiver. Or Firestarter vs some windows firewalls (not the default one tho that one is simple ). It's not all black and white.

Professional level software is a special case in which OSS has problems and always will. It could penetrate (see Blender, not top but good enough for many professional things) but commercial backing is a must in this case.

I agree with you to some extent, and as a programmer who is working on an OSS project AND has to get some money, I am more or less a supporter of the mixup group (those who shout "coexist!" ). It's just that overall I feel (as a user) that OSS software is proving to be of higher quality, albeit somewhat frustrating at times. (I could sing about my windows/$$ - software troubles which were sadly unfixable)

I think you are misunderstanding features and bloat. Bloat is something that is over-engineered, unexplainably inefficient or not of any practical use for most users.

XHTML & CSS are necesary, required, basic features. Having your hardware supported is not 'bloat'.

See, to slightly quote Chris Rock, I'm "TIRED, TIRED, TIRED" of people like yourself who incessantly reply "buy a new PC if your current one can't handle Firefox".

I'm using a "new" PC right now, a 2006 HP Pavilion dv6000 notebook with AMD Turion, 1.61 GHz, 992 MB of RAM, 14.8 out of 80.5 GB of HD space, with WinXP MCE '05 using Firefox 2.0.0.3.

Currently, it is using 165,000 K, and can shoot up to 500,000 or 600,000 K in a single session. It requires going to Task Manager > Processes to manually strangle the life out of the firefox.exe-beast.

Oh, and Windows XP is supposed to be the one on which Firefox works best?

Dude, crack is wack.

"""
If the code grows bigger and bigger, it's impossible to avoid the so called "bloat".

By the above definition, some people who complain about anything being "bloated" shouldn't even use computers, because hardware which computers are built upon is complex ("bloated").

All operating systems and good software are "bloated", because they take a lot of memory and hardware resources.

If you don't want "bloated" software, stick to DOS or buy better PC.
"""

Yes. But to help people understand, perhaps an analogy would be useful.

It's like diet.

If I keep on eating and eating more and more high calorie food every day, it is impossible to avoid so called bloat. There's is nothing I can do to avoid getting fat. It's just part of modern life.

So anyone who dislikes being fat should (by definition, as you put it), not eat at all, or eat only green salads without dressing (and perhaps a bit of plankton), or simply buy larger clothes.

Does that clarify matters for everyone?

Edited 2007-05-18 20:36

And lust like diet: different people react differently on the same food. Some people can eat a lot while not becomming fat. Some need more workouts, some less.
Just like the different browsers render the same sites in a different way. Some do it more efficiently without becoming fat.

"""
Some do it more efficiently without becoming fat.
"""

And as with people, some light is less flattering than other light.

Here's some unflattering light for modern applications:

http://tinyurl.com/3cjm22

"Here's some unflattering light for modern applications: "

Dude, you tinyurl'd a PDF document...at least give us some warning!

My memory was vague, but wasn't there thoughts floating around of creating XML-node hardware; being able to store and access an XML tree in hardware would be great!

>> "today's furious, high-maintenance web surfer, who can open as much as (I repeat this from a comment on Digg) 180 tabs"

I wonder what all those 180 tabs are used for. Porn?

I can name a few:

Wikipedia (both for reading and editing)
Digg
Google Search
YouTube
Blackboard (online classes for web design diploma)
More Wikipedia
Apple
More Google
etc.

I don't usually go much further than 65 tabs in a session, but I can and have done so.

Plus, I put the laptop to sleep overnight with the browser on, wake up, get back on, and start back to the browsing routine.

-----------------------

Who in their right mind would have that many tabs open? At most I will have maybe 10, normally 3-4. I close tabs when I'm done getting the information I want from the web page, and it would be very rare that I would have any reason to have more open than 20 maximum. If I want to go back to something, I will bookmark it. I can't imagine how anyone could possibly keep track of 180 open tabs. Firefox's memory usage has never bothered me. I rarely have more than 100MB of memory usage with Firefox, and with 2GB of RAM, that is a very small percentage of my available memory (5% in fact).

You can always close tabs, that is what bookmarks are for people...

Wasn't 10 tabs considered insane back in 1999? :-)

(Oh crap, now I'm doing the same thing that I accused someone else of earlier up in the thread).

To be honest, I don't think browser-side bookmarking is as necessary as it once was. Why should I have to treat a single web document with such status that I would have to close and save it for later, for fear that it will take up my memory?

I just want to open the webpage, put it aside for when I need it, and go to the next webpage in the meantime.

Oh, and if you couldn't tell, I "open link in new tab" alot. Almost all the time that I open a link from Google or some other site.

So maybe there is an issue with Firefox on the side of the OS or hardware; we won't be stuck with just 10 tabs for web surfing in the future.

Edited 2007-05-19 02:59

-----------------------

I have no fear of it taking up memory, it is simply that I am not able to keep track of more than a dozen or so open tabs at a time. I tend to read one thing at a time, and usually I will have multiple tabs open when I am comparing products from a range of retailers for example. I also turn off my computer and close my browser, so bookmarking useful links, or just googling to find them again is more useful to me than keeping the browser running with hundreds of tabs open.

Not needing more than 10 - 20 tabs has nothing to do with any limitations of computers or browser software per se, but rather my brain - I'm just not that good at multi-tasking. I don't have more than a couple of windows or apps open at a time for the same reason. That's just me though, I have no problem with anyone wanting to have lots of tabs open, though unless there is a large market for it, I don't see it as something Mozilla devs should be focusing on.

So.... basically... kill the user experience.

There are so many poorly written sites out there that such a set up would constantly have the user bringing into memory, the bad renderer... and if you must have your least bloated running, that means every time you render something with it, you then kill the instance and remove it from memory so that you can run at peak system efficiency rather than peak user experience.