Firefox Going the Big and Bloated IE Way?

abhinav_pc writes "Wired is carrying an article pondering whether Firefox has become big and bloated, much like IE. As the browser's popularity has risen, the interest in cramming more features into the product has as well. Slowdowns and feature creep have some users asking for a return to the days of the 'slim and sexy' Firefox. 'Firefox's page-cache mechanism, for example, introduced in version 1.5, stores the last eight visited pages in the computer's memory. Caching pages in memory allows faster back browsing, but it can also leave a lot less memory for other applications to use. Less available RAM equals a less-responsive computer. Firefox addresses this issue somewhat, setting the default cache lower on computers with less than a gigabyte of RAM. Though the jury is still out on where the perfect balance between too many and too few features lies, one truth is apparent: The new web is pushing our browsers to the limit.'"

is it time

to totally rethink the browser? with broadband becoming more available could websites be built in a way that current browsers don't even let us imagine?

Re:is it time

Actually, perhaps it's time to totally rethink the internet. Browsers today are bloated partly because websites are bloated.

The majority of websites could do with a simple and less cluttered layout like google's website for instance. Compare it to yahoo and you'll see that yahoo has a bunch of "advanced features" like inpage tabs and whatnot. Lots of this extra junk you'll find around the web is javascript that chooses CSS based on browser and that displays advertisements. Lots of it is just poor use of HTML often from WYSISYG programs. More features in language means more junk on website. More junk on website means more junk in browser.

The Gecko source code is a mess.

Go look at the source code to Gecko, the rendering engine behind Firefox, Seamonkey, Thunderbird and other projects. In short, it's a mess.

Part of the problem is the foolish complexity of it. Their whole XPCOM idea sounds nice in theory. But then you actually go to implement it in C++, and it becomes a pile of crap. Soon enough, difficult tasks start to become hard, the damn near impossible tasks can't be done, and nobody really has a good idea of what large portions of the codebase actually does. That's not the way to create an efficient rendering engine. You'll end up with memory leaks galore, and excessive CPU consumption, just as we've witnessed with Firefox.

Although it's unlikely to happen now, the best thing for them to have done would have been to throw out most of the code released by Netscape, rather than rewriting a lot of it (at the same low-quality level) in the following years. Then they could have re-implemented it using a natively-compiled implementation of Standard ML. One benefit of this would have been an elimination of the memory leaks that we hear to much about today, due to the garbage collection of SML. Additionally, functional languages are well-suited to parsing (ie. of HTML, XHTML, etc.) and language implementation (ie. JavaScript), more so than C++.

Re:The Gecko source code is a mess.

I once attempted to create a page-rendering engine, starting with XHTML. Eventually, I got a decent-working rendering engine. Unfortunately, anytime there was an error (even a minuscule one), my engine would completely fail. I can't even being to imagine the hell Gecko goes through to render a site like MySpace [w3.org]. I've often thought about a better way to implement a rendering engine, but most involve fixing the web developer's crappy code before attempting to render it, which is not possible in most cases. In C++, you can't compile with an error. Perhaps development software that isn't notepad (my software of choice) should add in validation service in the same way Visual Studio 2005 does.

The internet: We have the tools to rebuild it, but we don't want to spend a lot of money.

Re:The Gecko source code is a mess.

Your renderer did what most renderers years ago should have done: failed outright upon errors. That would have been essentially the same as a C++ compiler not emitting anything upon encountering a syntax error. Unfortunately, the early browser developers, mainly at Netscape and Microsoft, decided to try to handle such shit input. And so today we have crap like MySpace.

The flip site of strict error handling

I agree -- halfway. Had early web browsers been strict about errors, we wouldn't have so much broken code out there, and cross-browser compatibility would be solely a matter of which features are supported -- not which set of error-correcting assumptions you expect.

On the other hand, the fact that those early versions of Mosaic, Netscape, IE, etc. would do something with broken code instead of refusing to display it meant that the barriers to entry were a lot lower. It vastly increased the pool of people who could create web pages, and the talent pool. Sure, some people have both artistic talent and programming ability, or have the resources to team up. But can you imagine a web built solely by programmers?

Eventually the authoring tools would have caught up. But I have to wonder if the web would be as big and diverse as it is now if it hadn't been able to pull in the casual author back in 1995.

Yes, we have crappily-coded sites like MySpace. On the other hand, 10 years ago the idea of visiting a website was inordinately dorky, and being online meant you were a social outcast. Now, it seems like being offline is considered freakish.

Re:The flip site of strict error handling

But can you imagine a web built solely by programmers?

Yeah, it probably would have consisted entirely of porn.

Re:The Gecko source code is a mess.

To be fair, the w3.org validator is emitting spurious errors from the very top after failing to find the DOCTYPE declaration. It's obviously trapped in lexical lala land.

For example, it fails to recognize the head tag here:

<!-- *** VERSION 2.0 CSS *** -->
<!-- *** ELS2MWEBNET0756 *** -->
<html>
<head>

Re:The Gecko source code is a mess.

One of the principal goals of "Mozilla 2" is to subject the codebase to "deCOMtamination". Every instance of XPCOM than can be replaced with C++ exceptions will be, in order to reduce the ill effects of XPCOM that you outlined. Unfortunately, Mozilla 2 is estimated to be released as Firefox 4.0 in the first quarter of 2009--so at least a year and a half from now. This remedy may end up being too little too late.

Also see this kuro5hin story [kuro5hin.org].

Re:The Gecko source code is a mess.

Too little, too late? And who's going to come along and sink their ship in that year and a half? If Opera were going to do it, they'd have done it by now. Maybe if Konqueror could be a contender if it goes multiplatform (anywhere that runs KDE plus Windows and maybe a native Mac port) with KDE4/Qt4.

Other than that there really isn't anyone to take their place. Oon windows I highly doubt that you'll see many converts going back to IE, even if microsoft somehow makes it stop sucking with IE8, which I guarantee won't happen anyway.

Thoughts on rebuilding codebases

In a moment I'll talk about my views on rewriting large code bases, but first I'll say that I'm glad I wasn't the only one who was with the GP poster up until the SML advocacy, and then disagreed. Even given the neat way that functional languages tend to model parsing problems, web browsers do a lot more than parse HTML and CSS files. In fact, I'd go as far as to say that this is the easiest problem they solve. Systematically resolving the layout rules in arbitrarily complex cases is a somewhat difficult problem, given the way those rules are expressed in CSS. And of course, web pages are no longer the static things they used to be: today's browsers need to cope with scripts moving the goalposts arbitrarily, maintaining the integrity of the display as much as possible during lengthy downloads of large pages or after AJAXified updates, etc. etc. It's far from clear that a language like SML offers better support for these naturally concurrent operations than the many alternatives.

I do disagree with the parent post on one fundamental point, though:

In general, throwing out an existing code base is rarely a good idea. Practically speaking there's rarely a code base so bad that no part of it can be salvaged. Even when things are rewritten, it's almost always the overall structure that's just refactored by a lot of copy pasting.

I'll see your reuse dogma, and raise you my "plan to throw one away" dogma. :-)

Actually, I don't cite this as some sort of dogmatic adherence to ConceptsTryingToSoundMoreCleverThanTheyAre at all. Rather, I happen to agree with the principle based on practical experience. In general, software design is difficult, and few people are good at it. Even those who are rarely have the good fortune to know exactly what their design will be called upon to do a few years later, and will inevitably allow more flexibility (and commensurate overhead) in some places than is really needed, while making some things unnecessarily strict and thus making later changes more difficult than they might have been.

It's been my experience that in long-term projects, far too many managers aren't willing to throw out a whole module, subsystem, or even product, because of popular wisdom that anything they replace it with will just have bugs of its own. I believe this is a mistake because, again speaking only from my own experience, a high proportion of bugs originate in special or boundary cases. According to my reasoning above, a new project built from scratch with no prior experience will rarely get an overall design that automatically avoids these completely. Discipline is rarely good enough on software projects to allow for this and ensure that new requirements are integrated into a clean overall design rather than bolted on; indeed, in a commercial environment, this may not be realistic given short term deadlines and the typical management and marketing pressures. However, over time, such bolted-on special cases will tend to build up. They start to interact, they don't always get properly documented, and new people on the project team either don't know about them or at best don't know all the original reasoning behind them, making safe maintenance difficult.

Sometimes, this problem is manageable, particularly if your project leadership consistently take a long-term view and give maintenance and testing the priority they deserve. But usually, IME, the problem reaches a certain critical mass where the costs of ongoing development of a code base full of dubiously documented special cases outweigh the costs of stopping to clean things up.

As an additional, very practical concern, tools and programming techniques are always developing. Over the sort of timescales we're talking about here, it's entirely possible that more effective tools will have been created, or more effective techniques discovered, that could solve the underlying problem much more effectively in a different way.

Thus, s

Re:is it time

It's time to bring back VRML!

Re:is it time

As someone who lives out in the sticks, and pays $100/month for a 1.5MBit 802.11 connection, I say no. Keep the web as plain old HTML. Limit flash (And other plugins) to things like embedded video, NOT AS THE ACTUAL WEBPAGE.

There's still a lot of people out there who are limited to dialup, satellite, or some other jerry-rigged internet connection.

rethink the OS

Is there not some way that operating systems can manage caches for applications in a way that certain datasets can be marked as opportunisitic caching. That is, make it as keep a copy of this in any free space, but you can discard it if real memory is needed.

Re:rethink the OS

The OS could send a "SIG_FREE_UP_SOME_DAMN_MEMORY". The best part would be applications that don't handle it would just crash, freeing up lots of memory :) But yeah, your point is valid; An OS managed shared cache could make cache management easier. An easier to add although not quite as elegant solution would be to have that cache be part of the desktop suite; While not fully shared, at least all the desktop apps for one user would be cooperating.

Re:is it time

I'm doubtful that there's a substantial revision to the browser that would be useful. Firefox is really not that "large". When rendering small, lightweight Web pages, Firefox is actually not the largest application I run regularly... of course, small, lightweight Web pages are a dying breed.

That's not really Firefox's fault. Eight Web pages worth of cache is nothing... when you're not visiting a site that has 6 CSS stylesheets, 8 JavaScript sources and 20 images eight pages is a breeze. But visit most Web sites today, and you'll find that that's a dying dream.

Fortunately, well-designed Web sites can take advantage of this. For example, MediaWiki has tons of CSS and JavaScript associated with each page, but it's shared in common across almost all of those pages, so keeping 8 pages in cache isn't all that much more expensive than keeping one.

Re:is it time

The original files that compose a site are not the problem. You can throw all the raw CSS you can find into memory and not make a dent. The problem is that Firefox is saving the final DOM that's parsed out of those original source files. That's a lot bigger than the raw data, and it's not something that "simple" websites do any better at.

Very nice FUD

Wow, I actually RTFA and nowhere in there does it say that Firefox is becoming as "bloated" as Internet Explorer. Nope, it says it's becoming as bloated as Seamonkey. Oh the horror. The article is also (as usual) not kind to Firefox as far as the speed and insane memory consumption it suffers from, which thousands of fanboys have spent the past three years desperately denying for some weird reason. To be fair, I use FF and I don't care about the memory problem, but that doesn't mean it's not there.

Disingenuous FUD aside, I can't for the life of me imagine how IE could be "bloated". It never had much functionality to begin with.

Kudos to Bashdot. Even the current Digg submission [digg.com] doesn't mention IE at all.

Well...

At least when you install Firefox, you don't get some version of Windows along with it :-)

I mean, talk about bloat!

Re:Very nice FUD

I suspect Firefox has a problem with the amount of time it stays open, which means it has a memory leak (or a number of them) that are slight enough not to be a problem if you have it open for 8 or nine hours and close it every day but become an issue after two or three days.

The instance I'm typing this into (2.0.x) has been up for about three days. I have no idea how many tabs I've opened and pages loaded, but the task manager shows ~300MB mem and ~120MB VM usage. Keep in mind right now I only have two tabs open.

I suppose one could say the solution to the problem is restarting Firefox at least once a day... except that restarting a web browser seems about the stupidest thing ever.

Still, I put up with it because it's far better than IE6. I don't like Opera, so I don't have a lot of choice.

FF memory leak

There's ancient bug about mozilla coming up sloooooooooow in Win32 after it's been on for a long time. We're talking about mozilla 0.8 era bug and it affects all the related products that I'm aware of (TB, FF, probably even the calendar app)

The crux of the problem is that the bug affects only win32. So the developers, one and all, refuse(d) to treat it as a codebase problem because it has to be a windows problem.

"slow" in this instance means over 1 minute and easily 2-3 minutes. I even demonstrated at the time that it's not paging issue as such as mozilla/FF sits doing absolutely nothing (from perfmon monitor tool) for long period and when it finally starts swapping pages in it happens pretty quickly (5-10sec)

It was subsequently "fixed" by making FF etc hold on to the memory they've reserved instead of releasing it back to the OS. Hence you get ridiculous 300MB memory footprint that shrinks to 50MB after restarting FF even with the same pages open. Same goes for TB and all the other apps I'm sure.

So if you've got any kind of memory leak, mozilla apps want to keep it all in ram.

Re:Very nice FUD (you too)

The Firefox/Firebird/Phoenix project was started with the intention of being a lean browser based on the Gecko engine because the Mozilla Suite (now Seamonkey) was so massively bloated that it was easier to essentially start over than it would've been to attempt to slim down the main codebase. Firefox absolutely did not start out being more bloated than Seamonkey, otherwise it would've betrayed the entire purpose of its existence.

well no wonder

That's why I never get the first post!

well

Caching pages in memory allows faster back browsing, but it can also leave a lot less memory for other applications to use.
 
The amount of RAM used for caching pages could be set by the user in the options. I think most Firefox users could handle that.

To turn off the cache

A quick look finds the option to turn off the cache:

browse to about:config
search for the browser.cache.memory.enable setting
set it to false
restart the browser

On my machine, that lowers the memory footprint from 125MB to just under 50MB.

Re:well

So I've disabled page caching.
My FF has not been up to 800 MB like it has before, but it still gets up around 200-300 MB.
For a damn browser, that is retarded. I mean, entire servers used to have less RAM that that and could do far more than process and display html/images.

Re:well

The amount of RAM used for caching pages could be set by the user in the options. I think most Firefox users could handle that.

Sure, for geeks. But if we want people to stop using IE we must provide a credible alternative.

There should definitely be an option to tell Firefox to use less than n megabytes of memory, and let firefox figure it out, instead of setting the memory limit through the number of undo levels per tab.

Re:well

That's not what I'm saying at all. Right now, if you want to reduce the amount of memory Firefox uses to enable you to click the back button faster, you have to set a per-tab limit. This is dumb. I mean, it's nice to have that option, but for it to be the only means of controlling that limit? Stupid. There should be a nice fat "never use more than x megabytes of memory" option, and you can fill in the value of x. If the browser starts to run out of memory, it can prompt you and ask if you'd like to increase the memory limit. It needs to be as easy as possible for average users, who do not feel the same joy that you or I might when we figure out some computer-related esoterica. They just want to surf myspace and look at pictures of half-naked teenage girls.

Firefox=Mozilla?

More than anything it's reminding me of Mozilla, now known as SeaMonkey. The reason I switched from Mozilla to Firefox was because I wanted a smaller, more nimble browser. I didn't want a RSS reader, e-mail, IRC, etc. packaged together. Firefox hasn't integrated all of those yet but it's moving towards it and I don't like it.

Re:Firefox=Mozilla?

That's really the problem, I think. Firefox was originally supposed to be just a great browser, as opposed to the bloat that was Mozilla, with additional functionality being provided by Add-ons. Now, though, the development direction seems to be to take the best of the extensions and incorporate them into the main product. It might be better to keep the browser as it is, and then release a separate bundle with Firefox + the most popular add-ons. That way, people that want the slim browser they switched to Firefox for in the first place can have it, while the Firefox team can still have a download that will allow them to crow about all of the great features Firefox has.

And yet, there's Opera

What's odd is that Opera packs all that stuff in and more (even a BitTorrent client!), and it's faster and more lightweight in terms of resource requirements. Even the download size is amazingly small. What is it that makes Firefox worse in that regard? The XUL stuff? Convoluted codebase making improvement difficult?

Re:And yet, few use Opera

The number and severity of its security problems certainly qualify as "horrible." Obvious? Depends on who you ask. I'm in IT; I think so. Most people in IT would probably agree with that. End users? Maybe not. They have a different idea of "obvious" than most /. readers.

To put this into perspective, I haven't had a security problem on a windows box in over four years. All you need is to follow some good practices and you're perfectly safe. Vista with its limited user powers should help a lot in reducing the effort involved in those "good practices".

I love my Mac, but WTF do I have to reboot after updates?

Because the kernel has been updated. After you update itunes a reboot is not needed.

Is it horrible that an OS designed in the late 1960s, when the industry was still so young and inexperienced with security, is better-designed than NT and its descendants, which were designed twenty years later?

If you're referring to multics, no commercial operating systems have caught up to that yet. If you're referring to unix I have to disappoint you, unix was not designed to be inherently secure. On early versions there were many security issues, because the concept of limited user powers took a while to gain a foothold, and even when they did the system's design was still full of security holes. The first internet worm (the morris worm) specifically targeted unix systems, in case you've forgotten. Just look at the security track record of commercial unices. It's pretty poor.

Linux is not inherently more secure than other commonly available operating systems, it's just too much of a moving target because there is no binary stability, so a worm can't target large swaths of systems at the same time. That binary instability is a strength, but it's also the reason why there are so few binary drivers and commercial applications on linux.

Opera!

I guess it is the time now for people to look into Opera, which seems to be able to keep the balance. I think software should not be discriminated on the basis of not being FOSS.

Re:Opera!

I think software should not be discriminated on the basis of not being FOSS.

And I think it should. Guess that's why different things matter to different people.

Re:Opera!

Yes, some of us are prepared to use the best tool for the job rather than blindly follow FOSS.

Re:Opera!

Proprietary software is not practical, especially in the long term.

That's got to be the dumbest argument I've heard in favor of Free/Open Source Software in a long time. Look at IBM's mainframes. I don't think you could get more closed or proprietary. Yet, many businesses have stuck by them due to the fact that backwards compatibility is never broken on those machines. Heck, you could run OS360 packages from the 70's on the modern zOS machines without a problem.

I think its just a bit extreme to say that no proprietary solution is practical in the long term. It depends on the application, the stability of the provider, and the relationship that you have with the provider. These three criteria apply whether the provider is the open-source community or a private company.

Re:Opera!

I call bullsh1t on this. You've reviewed all the source of all the pgms you use? Stop this argument, please, it's not a real reason to choose one over the other unless you're actually willing to go through the source of every one of them

You can save your specious arguments for an audience that will buy them. We don't have to review the source of all the programs we use to gain the "transparency" benefit of Open Source or Free Software. The idea is "many eyes", not "my eyes".

Re:Opera!

I can watch whats going in and out of my ethernet and wireless card, but even so opera could be using some undocumented "feature" of a closed-source operating system to make sure I don't see it.

That's possibly the most insane statement I've ever read by someone who claims to be savvy enough to fork a development project on his own. If you really think it's possible for an OS to pass information over a network in such a manner as to make it undetectable, then I have a very special invisible firewall to sell you. Don't worry, it's open source.. I only charge for the distribution costs.

"Less available RAM"

Ignoring the poor grammar for a moment: "Less available RAM equals a less-responsive computer" is a bit simplistic. Unused memory is wasted memory, this is similar to the arguments about top(1) on Linux reporting all your memory being used in buffers etc.

Streamlined Version

Firefox has an awesome ability to add-on things very effectively. I don't understand why they don't keep fx slim with with all the proposed additional features as external (and hence optional) add-ons. Perhaps the not-so-computer-literate can use the bloated-up version of fx so they don't have to figure out how to use add-ons (I'm still amazed at how computer illiterate people can be), but leave a streamlined version for us techies to add-on options as we choose.

Re:Memory And Performance Rot

It seems that there is a magical responsiveness threshold which humans tolerate, and as the processing power and memory sizes grow, the applications follow along, staying just below that threshold. Usually the reasons are increasing amounts of shared libraries and scripting languages, which allow us to build more application per unit programmer time. We get more features and modern applications, at the expense of a sluggish environment.

This performance penalty is perhaps hard to notice. The easiest way to experience it is to run some old applications; they absolutely scream on modern hardware, to the point that the instant response becomes almost worth the loss of extra features. This is probably why things like xfce prosper.

Besides the cache

So other than the memory cache, what features could be stripped from FF to make it leaner and faster? I know nothing of its internals, but without any extensions it doesn't seem to have many wasteful features.

Re:Besides the cache

So other than the memory cache, what features could be stripped from FF to make it leaner and faster? I know nothing of its internals, but without any extensions it doesn't seem to have many wasteful features.

A quick glance at the Firefox features page lists these things, which as far as I'm concerned are bloat as they are not fundamental to a web browser:

• Spell Checking
• Search Suggestions
• Session Restore
• Web Feeds (RSS)
• Live Titles
• Integrated Search
• Live Bookmarks
• Pop-up Blocker
• Accessibility
• Phishing Protection
• Automated Update

I don't see any reason why all of those things are integrated and not seperate addons. And that list gets bigger with each new version.

And then you're nothing but IE...

Face it, folks - Phoenix was about slim, slim, slim. Firefox is NOT - it's about best-of-breed. For all the talk of bloat, take one look at things like the default toolbar layout (simple and streamlined), preferences (compare it to Mozilla or IE), and menu layout (again, dare to compare). Internal things like memory management and feature support may have increased, but the user interface has remained streamlined and efficient. THAT is what is important to the vast majority of users.

As to the parent post, let's see now:

RSS Support:

• Web Feeds (RSS)
  
• Live Titles
  
• Live Bookmarks

I could easily see removing RSS support. Firefox's implementation is nothing an extension couldn't do, and do much better. It's a joke for handling more than a handful of feeds, and stifles development of third-party extensions. Gee, and we used to complain about competing against built-in programs...

Security:

• Pop-up Blocker
  
• Phishing Protection
  
• Automated Update

Can you honestly say a browser should be shipped without these, or even an option to not install them? Especially for the popup blocker - are you insane, or have you simply forgotten what the popup-infested web was like? Phishing protection is unobtrusive and useful, as is auto-update.

Miscellaneous:

• Spell Checking
  
• Integrated Search
  
• Search Suggestions
  
• Session Restore
  
• Accessibility

Integrated search was one of the highlights of Mozilla ages ago, and is now a standard feature in every single browser. Firefox/Mozilla did a particularly good job by adopting an existing open format (from Apple's Sherlock) rather than reinventing the wheel. Search suggestions are the latest evolution of that (primarily thanks to Google Suggestions, if I'm not mistaken). Spell check is marginal - many operating systems offer their own - but I don't see how a third-party extension could improve upon it. Accessibility is just critical for those who need it. Session Restore I'm torn on, as many extensions handled it, but not necessarily well. I see that as the Firefox team deciding to take all of the lessons learned from the third parties, and do it right (much like Apple did with iTunes 1.0).

Bloat is only a problem if it hinders program development, maintenance, execution, or usability. The examples given here don't generally meet those criteria. Most of the features here are simple, self-contained, unobtrusive, and likely have low code and memory footprints.

Re:Besides the cache

Why does slimming Firefox down necessarily mean removing features? Opera can do pretty much all of the things you quotes and much, much more besides (email client, bittorrent client, customizable to the extent that would need about 15 different FF extensions to emulate, etc.) -- and it still manages to be slimmer than Firefox -- a smaller download (4.7 vs 5.7MB), faster to start, more responsive, a smaller memory footprint, etc.

Shared Javascript Namespaces

Aside from that, the ongoing issue with Web 2.0 apps and javascript with multiple tabs using the same shared namespace and overwriting variable names still hasn't been highlighted by the security community and as AJAX and web based applications become more prominent, the end user will find more and more applications breaking other applications.

Not my experience

I spend all day^H^H^H^H^H^H^H a few momentes when I would not otherwise be productive, pimping my music round myspace (surely the biggest resource hog on the net) and firefox holds up fine on my 256MB Thinkpad (running ubuntu).

Phoenix user since day 1

I used one of the very first phoenix builds. It impressed me because at the time i was using mozilla. Phoenix was literally just gecko + some ui and it was really really light and fast. There was no installer, no control panel (well it was blank), etc.

I'm very happy with firefox so far. I run half a dozen extensions to give me features like "session saving" etc. Ram usage is not too much of a concern with me. I would like it if the default was to not cache 8 pages back. And on disk cache should be fast enough to retrieve and render. 90% of the time i only go back 1 click anyways.

Firefox 3 is implementing major changes. Under the hood they are switching to garbage collection and cairo (vector rendering) just to name a few. Cairo is a great abstraction that hasn't fully realized its performance capability. I don't suppose glitz will be out anytime soon. The sql-lite bookmarking looks neat. Epiphany has something similar. But i must admit that i've fully switched to del.icio.us and the extension v1.5.29. That's quite fully featured and it syncs across computers.

The rss reading capability i do not like at all. That should be implemented as an extension. I prefer to use liferea. There are plenty of firefox features that should be implemented as extensions. That way you can disable them if you wish.

Oh! Oh! I know!

Let's make a new and smaller browser, based on the same rendering engine! We'll call it Phoenix or something like that. You know, like it's brand new! It comes from the ashes, it must be good! And we won't bloat it, no, no. We'll make it speedy!

Where did I hear that before?

FireFox is a huge resource hog

My MacBook Pro had 512 megabytes when I bought it. That ought to be enough memory for anyone. But I found that running Parallels (a virtual machine that can host Windows or Linux) at the same time as FireFox was completely intolerable, even if I set Parallel's memory allocation to a minimum level.

Whenever I clicked from one window to the other, I'd get the Spinning Pizza of Death for a minute or so while the other task's memory was paged in. I had to add another gig of RAM before I could switch windows quickly.

That made this old coder wanna cry. My first Mac had only 512 kilobytes (kilo - not mega) but that was enough for me to write GUI applications with.

Kids these days don't know how to write code.

Re:FireFox is a huge resource hog

My MacBook Pro had 512 megabytes when I bought it. That ought to be enough memory for anyone.

With 512MB, your system was probably already swapping, albeit not very much. OSX uses more RAM than there is any possible justification for, and I don't mean for buffers.

Windows XP is pretty much useless without at the very least 256MB RAM. Oh yeah, you can use it, but you're not going to do anything quickly. You will be constantly swapping. OSX is useless without at least 512MB RAM. You had 256MB too little ram to even play, let alone to have things be efficient.

Kids these days don't know how to write code.

Many of us would love it if the entire system were rewritten in tight, efficient code. I suggest you get right on that.

Slightly ot... a nit pick about the file cache

Why is the FF file cache so obscure? (kept in hex named files that appear to be indirectly referenced by other map files...)

One think IE does right is a true file-for-file cache of what you have browsed.

Sometimes I like to troll thru my "Temporary Internet Files" folder and pick out a few bits for posterity. Especially large .swf or .flv files that I might have watched. The worst is when I watch one of those in FF, then want to grab the file... the easiest thing to do is to watch it AGAIN in IE so that I can go cache-picking later...

maybe it's just me.

Re:Slightly ot... a nit pick about the file cache

Automatically storing files locally with contents and names that are defined remotely is a security risk. It would not be a security breach in itself, but it could create an opportunity to exploit unrelated bugs which would otherwise not be remotely exploitable.

become?

I don't think Firefox ever was such a lean or efficient browser. It's also buggy and the developers don't seem to care much about Linux or MacOS (bad profile support, inefficient graphics, etc.). Opera and Konqueror both seem better written and better designed.

I still use Firefox. Why? Because Firefox works well enough, it's up-to-date, compatible, and, most importantly, has tons of useful extensions.

I hope the Firefox developers will be able to clean up their act, but unless it gets a lot worse, I'm sticking with Firefox, because, on balance, it's still the best browser there is.

Firefox stopped being lean a long time ago

Firefox was only leaner than Mozilla back when it was called Phoenix and had only the bare minimum UI necessary to be a web browser.

Mozilla never was slow (at least not after it reached the point that it was good enough to consider using as your standard browser) and really wasn't a memory hog. That perception came about from the people who really didn't want an integrated email program, but absolutely refused to choose "Browser only" when the installer asked what they wanted.

Around the time of the name changed from Phoenix to Firebird, the two browsers were about on par. By the time the name changed to Firefox, it was already more bloated than Mozilla. The project goals moved more towards grabbing attention than being lean.

If Mozilla had just made a theme that blended in to the OS (Classic doesn't do a good enough job of it) and put a link on the download page to an installer that only had the browser included, there never would have been a need for Firefox.

Firefox, the new EMACS

To paraphrase an often heard comments about EMACS way back when, "EMACS isn't an editor, it's a lifestyle". Hopefully Firefox isn't headed down the same path.

IE is bloated?

There are many, many things you can criticize IE for... but being bloated doesn't really seem like one of them. If you RTFA, they compare the growing bloat not with IE, but with Mozilla.

True, 3rd party add-ons for IE can bring it to a crawl, but that's not IE's fault. The same problem exists in any browser that supports extensibility via a plugin model.

I use Firefox on XP because it's safer than IE, certainly not because it's less bloated. Firefox consistently uses far more ram (I have several screen shots of Firefox using 1.5GB+ of ram with *no* plugins enabled and just one tab open), dies a painful death due to poor integration with things like Flash (100% CPU Flash advertisements, anyone?), or simply just crashes.

On Vista I use IE 7 w/Protected Mode. Why? Well, again, because it's safer. But it also has the benefit of returning me to the days when a browser didn't use 2x the RAM of Photoshop. Imagine that.

The Wrong Question

All over the web today there are stories about FireFox's (supposed) bloat, but no actual facts on whether it is or is not actually "bloated." Since "bloat," to most people, apparently means the state of a program having more features than is necessary, it's hard to see how the average user would ever be able to definitively answer this question. The question is probably better phrased as "Are you having major performance problems with FireFox 2.0?"

I don't know how the file size (the other definition of "bloat"), of a FireFox installation compares with other browsers but it doesn't seem like an overly large file to download. It also seems to me that when I check my FireFox preferences it actually has a very basic, simple feature set similar to what's available in almost every other browser. If the feature set is roughly the same as other browsers, how can it be rightly called "bloated?"

I think the problem with FireFox is one of performance, not "bloat" per se. I run FireFox on a Mac with only a single extension and a single theme. My computer is relatively new, the OS is up to date, it has a Gig and a half of RAM and a fast video card. On this machine FireFox is as slow as molasses. It takes ages to start and ages to load a page. It also crashes (a lot!).

I use FireFox because of AdBlocker and because as bad as it is, it's still the best there is on the Mac right now. This will likely change in October when the new Safari comes out so this summer's FireFox 3.0 release will have to be extremely, extremely good just to keep the same market share IMO.

OS Level Control?

Why can't the OS, when it sees that it is running out of memory, send a signal/message/henchman to applications and tell them that if they have the ability to give up some memory (i.e., caches, etc), then do so, to keep the system happy. There could be several levels of urgency in the request as well, like "yeah, dude, just thinking here, yeah, could you ease up a little on the memory, cheers!" through to "Sieg Heil! Deine Memory, SCHNELL!!".

If firefox becomes bloated

If firefox becomes bloated I will eat the internet with a fork.

Get it? 'Fork'? *wink*

I wouldn't worry to much.

Oh and give epiphany a try.

Extensions are a nightmare

Firefox needs to stop cramming information about all your extensions into a couple of Registry-Hell-ridden configuration files. And then they cross-link by hard-to-remember GUIDs rammed into hard-to-read RDF? wtf?

Look at how extensions are done for Eclipse or JBuilder. It's much cleaner. Don't want an extension anymore? Just delete the JAR or folder. That's it. And it's clean.

Not just Firefox.

Most of my apps are using obscene amounts of RAM these days. Gaim/Pidgin for example, going by the RSS value, is using 32MB even when minimised to systray with no active conversations. The XFCE settings daemon is another 20, and that doesn't even have a GUI. Doesn't help much when I dumped KDE for it in the first place to try and fix exactly this...

Re:Firefox 2.x crashes all the time

I'm running 3/4 of a gig, and I've never had Firefox crash. And I have BOINC running all the time. My CPU is spinning pretty high all the time, and I tend to have a good bit of my RAM being used all the time. So I don't know what you're doing wrong dude.

Re:Firefox 2.x crashes all the time

My attempt to get modded up (any positive mod) by only quoting grandparent, parent, and the summary(and in that order). . . here goes:

Most annoying thing are the crashes of Firefox 2.x! I don't care if it eats a lot of memory (I've got 2GB - who wouldn't these days?) or is bloated, but I can't stand the crashes!

I'm running 3/4 of a gig, and I've never had Firefox crash. And I have BOINC running all the time. My CPU is spinning pretty high all the time, and I tend to have a good bit of my RAM being used all the time. So I don't know what you're doing wrong dude.

Firefox addresses this issue somewhat, setting the default cache lower on computers with less than a gigabyte of RAM.

Re:Firefox 2.x crashes all the time

256MB ? Holy crap! I don't understand why people think we need at least 512MB to run anything decently! In 2001 i was running on 64MB and i can remember i could run a web browser (granted it was IE, but nevertheless!) Winamp and some other stuff. And people _expected_ it to run smoothly with only 64MB ! I know it 6 years from that time, Moores law and such, but i still wonder - why this insane amount of hardware requirements? Notice that Opera for Symbian must run with 8MB of RAM and it has to share. And there's no virtual ram, so swapping is not an option. This of course doesn't count Flash. Right now both of my boxes have 1GB of Ram, and i don't plan on upgrading that number anytime soon - I don't play games (consoles are for that, and my Gamecube has about 48MB combined too!), i don't run VMs and i don't even have a swap partition - it never got touched anyway.

Re:Firefox 2.x crashes all the time

Most annoying thing are the crashes of Firefox 2.x! I don't care if it eats a lot of memory (I've got 2GB - who wouldn't these days?) or is bloated, but I can't stand the crashes!

Startup firefox like so: `firefox -p`. Add a new profile, use that one instead. Profile corruption is usually the cause of firefox crashing. All you have to do is move things over you still want like your bookmarks and go from there. It's a good chance that this'll fix your problem.

A user of mine had this problem yesterday. 5 minutes on Mozilla's knowledgebase does wonders.

Re:Firefox is written in J#, IE is written in D

Yeah, they're currently addressing that problem by rewriting Gecko in PHP (the UI will, however, continue being written in their flavor of eXtended ML). It will run on Mozilla's webservers and you will be able to run Firefox via any compatible web browser.

Apple, OTOH, will rewrite Safari, KHTML, Konqueror and most of KDE in Objective Ruby, which will run on their iNternet iServers, accessible via iTCP/iP (compatible with Mac OS 10.9 and up). Right after they switch their kernel to Hurd.


Yes, that's exactly how the future is going to be or my name is not Sir Reginold Frankbarrister O'Fritzebolt-Tooley the Thirteenth!

Re:FF2 is a horrible disappointment

I refuse to use FF2 anymore. It takes FOREVER to save images, and even worse, while it's taking forever, it essentially locks up FF, leaving it unusable for up to 2 minutes at a time. This is completely unacceptable, IE3 with a 56k connection saved images faster.

Delete the downloads.rdf from whatever directory your profile is stored in. Instead of complaining you could have just googled for "firefox slow downloads".

Enjoy,