Since when did memory usage become such a big deal?
Uh, you have noticed all the people crying that Firefox is a bloated memory hog, haven't you? The ones that have been demanding that Firefox use less memory? I guess you should be careful what you wish for. Personally, I agree that it's been made a much bigger deal than is necessary. I wish Firefox developers would spend more time on other fixes, as memory use is far from being an issue for me. On the other hand, it doesn't seem to have caused performance problems or a reduced feature set. It's just being more efficient with memory without much performance impact.
Yes, but those are very early development builds of those browsers. They haven't even seen an alpha release, much less a beta. The "Opera" build was actually using the WinGogi interface for Presto, and the Opera developers said not to use those builds for everyday browsing. You would want to compare those browsers to Firefox 4 nightly builds. However, I don't think work has even started on Firefox 4 yet. I opted to compare Firefox 3 to the recently released Safari 3.1 and the soon-to-be released Opera 9.5.
Now that's just silly. Mozilla developers never flamed anyone for saying leaks exist in Firefox. In fact, they said that of course Firefox has memory leaks. On the other hand, even with the leaks it seems to use less memory than other browsers. So saying it's got horrible memory leak problems, far worse than other browsers, will probably get you flamed. But simply saying it has memory leaks is an objective, verifiable fact that no reasonable person should disagree with.
Unlike most of the people in this thread, my browsing experience is in no way diminished by an inability to see the Acid2 smiley face.
Exactly. The one people who should care about browsers passing Acid tests are web developers. Users shouldn't care because very few web developers will use features that don't work in the most popular browsers. The reason you don't hear web developers cheering is because most users are using IE7 and IE6, which do not pass Acid2 or Acid3. The Acid tests are great because they will mean web developers will have an easier time developing sites and users will eventually get sites that work better, but the Acid tests are not going to have that effect for several more years. The users will be the last to see the impact of Acid tests.
You somehow missed the point of the post. If you actually used Acid2 or Acid3 features today, you would not be serving 75% of your potential customers. That's why you can't actually use those features until Microsoft has released a final version of IE that has those features, and your customers have upgraded to that new version. How fast Mozilla, Apple, and Opera release browsers that pass the Acid tests is irrelevant, as long as they are faster than Microsoft. This whole "race" is misguided and serves no purpose, except so fanboys can brag about how their browser passes Acid3.
And even after IE8 is released, IE7 will be popular for years the same way as 40% of IE users are still using IE6 today. We won't be able to use all the features tested by Acid2 until 2010 or later because of the older browsers still in use that don't pass. The only thing that could allow us to use Acid3 features faster is if Microsoft gets IE8 to pass (instead of waiting for IE9) and getting its users to upgrade to IE8 faster than they're upgrading to IE7. Unless they do, it doesn't make a bit of difference how fast Safari and Opera or even Firefox release browsers that pass.
Yes, there are programs that may or may not terminate, that no human can prove will terminate or will not terminate, except by running the program, and if he ever sees it terminate, he says, "It terminates!" But that's not the question. Does a computer program exist that a human cannot even prove whether it gives the correct result or not? That is, one that we know terminates and gives an answer, but no one can prove that the program gives the correct answer in all cases, and also no one can prove that it is possible to get an incorrect answer? I suppose we have to discount as trivial all programs that simply print an undecidable statement.
Truly mathematically speaking, yes, it is generally impossible to write a computer program that can automatically verify whether a program works. You point out that if this were possible, it would be possible to solve the halting problem. But is there really a computer program that is correct but that humans cannot prove correct?
Even if the tally was exactly right, in general you cannot prove a program correct by using only black box testing. There are simply too many possible inputs to have time to test for all but the most trivial inputs. For all we know, there's a backdoor or unintentional security vulnerability that can be used to alter election outcomes. We need to be able to examine the machine's inner workings to have any hope of verifying those are not problems with the voting machine.
No, I never said that Firefox is flawless, or that it must be a problem on your computer. On the other hand, I see no serious problem in Firefox 2 or Firefox 3, despite asking over and over again how I could see these memory and CPU problems people keep complaining about. I have helped many people who have come to the MozillaZine forums complaining about experiencing problems with Firefox, and it usually turns out that a simple operation such as creating a new profile or updating their drivers fixes the problem. It looks to me like these problems are isolated problems on people's computers, not some horrible, obvious flaw in Firefox. If you think there is a problem, show us how to reproduce the problem. And it's not MY job to release software that isn't broken, as I have no affiliation with Mozilla. I was trying to help you, but now I can see is all you want is someone to argue with and complain to. Send it to/dev/null next time.
Why doesn't every site that allows you to download software warn you that it may be dangerous? And I've filed over 100 bug reports at Mozilla, and most of them (especially the bad ones) have been fixed. It's not futile at all. Stop complaining for the sake of complaining and try giving some constructive advice once in a while, and you may find that it pays off. Whinging is and exercise in futility.
It may be working fine for you, but Netscape 7.2 has tons of publicly known security holes in it. You're far safer with a browser that gets regular security updates. If you're sticking with Netscape 7 because it's a suite that includes email, the browser suite you'd probably feel most comfortable with is SeaMonkey, because it uses essentially the same code base that Netscape 7 did.
Yes, those stats are true, but only for the W3Schools server. If you want the stats for a more representative sample of web traffic, go with one that aggregates data from thousands of various servers: NetApplications, TheCounter, OneStat.
I thought it was me providing you information to help you figure out your problem. If it's occurring only on your machine, that sounds like it's the number (2) situation I describe. If you still think it's the number (1) situation, it's still up to you to write a useful bug report. You can read what folksaresayingabout Firefox 3, and I don't see any hint of it having any memory problem. In fact, they seem to be saying there isn't one.
It's true that an extension can use as much memory as it wants. The same is true of browser plugins, isn't it? Why don't you just avoid the extensions and plugins with horrible memory leaks? Complaining about Firefox not saving you from buggy extensions is almost like blaming your OS for allowing you to install a keylogger.
We're not being fanboys. We're giving you the facts of the matter. If you still don't believe us, read what a Firefox developer has to say about Firefox memory usage. There is no one major memory leak. There are lots of things that can possibly cause Firefox to use lots of memory over long periods of time: caching, fragmentation, small memory leaks adding up, etc. It looks like Firefox 3 uses less memory than Firefox 2, and both use less memory than IE (see the second graph in the page I linked to).
The point you're missing is that it's not the public who are waiting for browsers that pass Acid2. It's developers who are waiting. Even after Firefox 3 and IE 8 are released, web developers still won't be able to use all the great new features that Acid2 tests, because they won't work for all the users still using IE 7. The limiting step in what standards web developers use is still how fast Microsoft gets versions of IE out to the public and how fast they get users to upgrade to the new versions. It wouldn't make any difference if Safari, Opera, and Firefox all passed Acid2 the day after it was finished. Even if Firefox 3 is released after IE 8, Firefox users will upgrade to the latest version much faster than IE users, the same way that currently 95% of Firefox users are using the latest version of Firefox, but only 60% IE users are using the latest version of IE, according to NetApplications.
If someone found and fixed a horrible memory leak, why on Earth would it not be accepted? I've never seen any reasonable patch to fix a bug not accepted. I've seen modifications and enhancements rejected. I've seen very kludgy attempts to fix a bug rejected. But not a decent fix for an outright bug. Anyway, all anyone needs to do is point out a problem, and Mozilla developers will fix it for you. We're still waiting on that all important demonstration of THE memory leak. After we get that, we can write a bug report then just sit back and wait for the bug to be fixed.
I started Firefox on two different computers Tuesday afternoon with those exact four pages open in four different tabs. Computer A is an HP Pavilion running Windows XP Home and Firefox 3.0b5pre build 20080306. Computer B is a Dell Optiplex running Windows XP Pro and Firefox 2.0.0.12. When I first started the computers, Firefox was using 67 MB on computer A and 70 MB on computer B. After leaving Firefox running all night (more than 12 hours), Firefox was using 80 MB on computer A and 91 MB on computer B. That is an increase of 13 MB on computer A and 21 MB on computer B. It's possible that that amount of memory increase is due to a memory leak, but most often small increases are due to the effects from caching and fragmentation. I notice that the pages do have dynamic content, which would require loading new content and allocating and deallocating memory, so you would have to expect a certain amount of memory increase after hours of use. These are not static HTML pages.
The signature of a memory leak is memory usage increasing without limit. What happens when those sites are open for a week? Does memory use continue to increase to 170 MB? If so, I wouldn't be able to explain that degree of memory use increase by caching and fragmentation, and I'd have to say it looks like a leak.
The bottom line is that the vast majority of Firefox users don't have a particular problem with memory usage. If you do, then either (1) you are experiencing a bug in Firefox, and you should find a way to reliably reproduce a problem so it can be fixed, or (2) you are having a problem on your computer, and you should follow the advice in the MozillaZine Knowledge Base to fix your problem.
You are correct that if you leave Firefox on a page with a banner ad that continually changes, and the ads cause Firefox to use more and more memory without limit, this is a problem. Give us the URL of a page we can visit to see the problem, and we can file a bug report.
Flash and JavaScript cannot really be limited to a certain amount of memory. For any limit that you try to impose, users are sure to encounter a site that needs more. In this case, I'm sure users would rather have the site work than refuse to allocate more memory. Above all, users want their browser to work properly on the sites they visit.
If you have a single tab open and load different pages in it, memory use cannot be only what is used for the last page opened. If you want to approach that ideal, you can disable the memory and bfcaches entirely. But still as memory is allocated and deallocated, memory fragmentation will cause memory use to creep up over time. There should be a maximum that is reached. If you can find a page or sequence of pages that cause Firefox to use an unbounded amount of memory as you keep loading them, please tell us what they are so we can file a bug report.
The public releases were Firefox 3 builds. No Firefox 2 builds ever passed Acid2. The point is that for over a year, Microsoft has known that Firefox 3 would be passing Acid2. My guess is that knowing this would happen (and that Opera and Safari also pass Acid2), they made sure IE8 would pass, too. And just last week they finally shipped a public version of their browser that possibly passes Acid2. It looks to me like Mozilla is still leading, even if IE8 final ships before Firefox 3 final.
Just see if you're running any of the extensions known to have bad memory leaks or one of the extensions that cause leaks in Firefox. If you're not running one of them, you'll probably be safe. For brave souls wishing to track down which extension is leaking, I suggest creating a new profile, then adding extensions one-by-one.
I don't find anecdotal evidence convincing, either. That's why without a specific set of steps to reproduce a memory problem (including especially the version of Firefox you are running and the operating system you are using), I don't believe there is one. Can you explain exactly which pages you have open (give exact URLs)? We can all try to leave the pages open all day and see how much memory Firefox is using tomorrow morning.
Slashdot Mirror
Yes, but those are very early development builds of those browsers. They haven't even seen an alpha release, much less a beta. The "Opera" build was actually using the WinGogi interface for Presto, and the Opera developers said not to use those builds for everyday browsing. You would want to compare those browsers to Firefox 4 nightly builds. However, I don't think work has even started on Firefox 4 yet. I opted to compare Firefox 3 to the recently released Safari 3.1 and the soon-to-be released Opera 9.5.
Now that's just silly. Mozilla developers never flamed anyone for saying leaks exist in Firefox. In fact, they said that of course Firefox has memory leaks. On the other hand, even with the leaks it seems to use less memory than other browsers. So saying it's got horrible memory leak problems, far worse than other browsers, will probably get you flamed. But simply saying it has memory leaks is an objective, verifiable fact that no reasonable person should disagree with.
You somehow missed the point of the post. If you actually used Acid2 or Acid3 features today, you would not be serving 75% of your potential customers. That's why you can't actually use those features until Microsoft has released a final version of IE that has those features, and your customers have upgraded to that new version. How fast Mozilla, Apple, and Opera release browsers that pass the Acid tests is irrelevant, as long as they are faster than Microsoft. This whole "race" is misguided and serves no purpose, except so fanboys can brag about how their browser passes Acid3.
And even after IE8 is released, IE7 will be popular for years the same way as 40% of IE users are still using IE6 today. We won't be able to use all the features tested by Acid2 until 2010 or later because of the older browsers still in use that don't pass. The only thing that could allow us to use Acid3 features faster is if Microsoft gets IE8 to pass (instead of waiting for IE9) and getting its users to upgrade to IE8 faster than they're upgrading to IE7. Unless they do, it doesn't make a bit of difference how fast Safari and Opera or even Firefox release browsers that pass.
Yes, there are programs that may or may not terminate, that no human can prove will terminate or will not terminate, except by running the program, and if he ever sees it terminate, he says, "It terminates!" But that's not the question. Does a computer program exist that a human cannot even prove whether it gives the correct result or not? That is, one that we know terminates and gives an answer, but no one can prove that the program gives the correct answer in all cases, and also no one can prove that it is possible to get an incorrect answer? I suppose we have to discount as trivial all programs that simply print an undecidable statement.
Truly mathematically speaking, yes, it is generally impossible to write a computer program that can automatically verify whether a program works. You point out that if this were possible, it would be possible to solve the halting problem. But is there really a computer program that is correct but that humans cannot prove correct?
Even if the tally was exactly right, in general you cannot prove a program correct by using only black box testing. There are simply too many possible inputs to have time to test for all but the most trivial inputs. For all we know, there's a backdoor or unintentional security vulnerability that can be used to alter election outcomes. We need to be able to examine the machine's inner workings to have any hope of verifying those are not problems with the voting machine.
No, I never said that Firefox is flawless, or that it must be a problem on your computer. On the other hand, I see no serious problem in Firefox 2 or Firefox 3, despite asking over and over again how I could see these memory and CPU problems people keep complaining about. I have helped many people who have come to the MozillaZine forums complaining about experiencing problems with Firefox, and it usually turns out that a simple operation such as creating a new profile or updating their drivers fixes the problem. It looks to me like these problems are isolated problems on people's computers, not some horrible, obvious flaw in Firefox. If you think there is a problem, show us how to reproduce the problem. And it's not MY job to release software that isn't broken, as I have no affiliation with Mozilla. I was trying to help you, but now I can see is all you want is someone to argue with and complain to. Send it to /dev/null next time.
Why doesn't every site that allows you to download software warn you that it may be dangerous? And I've filed over 100 bug reports at Mozilla, and most of them (especially the bad ones) have been fixed. It's not futile at all. Stop complaining for the sake of complaining and try giving some constructive advice once in a while, and you may find that it pays off. Whinging is and exercise in futility.
And that would prevent plugins from using all your computer memory how exactly, oh Firefox FUD thrower?
It may be working fine for you, but Netscape 7.2 has tons of publicly known security holes in it. You're far safer with a browser that gets regular security updates. If you're sticking with Netscape 7 because it's a suite that includes email, the browser suite you'd probably feel most comfortable with is SeaMonkey, because it uses essentially the same code base that Netscape 7 did.
W3Schools server traffic is not a good representative sample of web traffic. According to NetApplications Netscape is used more than IE 5.
Yes, those stats are true, but only for the W3Schools server. If you want the stats for a more representative sample of web traffic, go with one that aggregates data from thousands of various servers: NetApplications, TheCounter, OneStat.
I thought it was me providing you information to help you figure out your problem. If it's occurring only on your machine, that sounds like it's the number (2) situation I describe. If you still think it's the number (1) situation, it's still up to you to write a useful bug report. You can read what folks are saying about Firefox 3, and I don't see any hint of it having any memory problem. In fact, they seem to be saying there isn't one.
It's true that an extension can use as much memory as it wants. The same is true of browser plugins, isn't it? Why don't you just avoid the extensions and plugins with horrible memory leaks? Complaining about Firefox not saving you from buggy extensions is almost like blaming your OS for allowing you to install a keylogger.
We're not being fanboys. We're giving you the facts of the matter. If you still don't believe us, read what a Firefox developer has to say about Firefox memory usage. There is no one major memory leak. There are lots of things that can possibly cause Firefox to use lots of memory over long periods of time: caching, fragmentation, small memory leaks adding up, etc. It looks like Firefox 3 uses less memory than Firefox 2, and both use less memory than IE (see the second graph in the page I linked to).
The point you're missing is that it's not the public who are waiting for browsers that pass Acid2. It's developers who are waiting. Even after Firefox 3 and IE 8 are released, web developers still won't be able to use all the great new features that Acid2 tests, because they won't work for all the users still using IE 7. The limiting step in what standards web developers use is still how fast Microsoft gets versions of IE out to the public and how fast they get users to upgrade to the new versions. It wouldn't make any difference if Safari, Opera, and Firefox all passed Acid2 the day after it was finished. Even if Firefox 3 is released after IE 8, Firefox users will upgrade to the latest version much faster than IE users, the same way that currently 95% of Firefox users are using the latest version of Firefox, but only 60% IE users are using the latest version of IE, according to NetApplications.
If someone found and fixed a horrible memory leak, why on Earth would it not be accepted? I've never seen any reasonable patch to fix a bug not accepted. I've seen modifications and enhancements rejected. I've seen very kludgy attempts to fix a bug rejected. But not a decent fix for an outright bug. Anyway, all anyone needs to do is point out a problem, and Mozilla developers will fix it for you. We're still waiting on that all important demonstration of THE memory leak. After we get that, we can write a bug report then just sit back and wait for the bug to be fixed.
I started Firefox on two different computers Tuesday afternoon with those exact four pages open in four different tabs. Computer A is an HP Pavilion running Windows XP Home and Firefox 3.0b5pre build 20080306. Computer B is a Dell Optiplex running Windows XP Pro and Firefox 2.0.0.12. When I first started the computers, Firefox was using 67 MB on computer A and 70 MB on computer B. After leaving Firefox running all night (more than 12 hours), Firefox was using 80 MB on computer A and 91 MB on computer B. That is an increase of 13 MB on computer A and 21 MB on computer B. It's possible that that amount of memory increase is due to a memory leak, but most often small increases are due to the effects from caching and fragmentation. I notice that the pages do have dynamic content, which would require loading new content and allocating and deallocating memory, so you would have to expect a certain amount of memory increase after hours of use. These are not static HTML pages.
The signature of a memory leak is memory usage increasing without limit. What happens when those sites are open for a week? Does memory use continue to increase to 170 MB? If so, I wouldn't be able to explain that degree of memory use increase by caching and fragmentation, and I'd have to say it looks like a leak.
The bottom line is that the vast majority of Firefox users don't have a particular problem with memory usage. If you do, then either (1) you are experiencing a bug in Firefox, and you should find a way to reliably reproduce a problem so it can be fixed, or (2) you are having a problem on your computer, and you should follow the advice in the MozillaZine Knowledge Base to fix your problem.
You are correct that if you leave Firefox on a page with a banner ad that continually changes, and the ads cause Firefox to use more and more memory without limit, this is a problem. Give us the URL of a page we can visit to see the problem, and we can file a bug report.
Flash and JavaScript cannot really be limited to a certain amount of memory. For any limit that you try to impose, users are sure to encounter a site that needs more. In this case, I'm sure users would rather have the site work than refuse to allocate more memory. Above all, users want their browser to work properly on the sites they visit.
If you have a single tab open and load different pages in it, memory use cannot be only what is used for the last page opened. If you want to approach that ideal, you can disable the memory and bfcaches entirely. But still as memory is allocated and deallocated, memory fragmentation will cause memory use to creep up over time. There should be a maximum that is reached. If you can find a page or sequence of pages that cause Firefox to use an unbounded amount of memory as you keep loading them, please tell us what they are so we can file a bug report.
The public releases were Firefox 3 builds. No Firefox 2 builds ever passed Acid2. The point is that for over a year, Microsoft has known that Firefox 3 would be passing Acid2. My guess is that knowing this would happen (and that Opera and Safari also pass Acid2), they made sure IE8 would pass, too. And just last week they finally shipped a public version of their browser that possibly passes Acid2. It looks to me like Mozilla is still leading, even if IE8 final ships before Firefox 3 final.
Just see if you're running any of the extensions known to have bad memory leaks or one of the extensions that cause leaks in Firefox. If you're not running one of them, you'll probably be safe. For brave souls wishing to track down which extension is leaking, I suggest creating a new profile, then adding extensions one-by-one.
I don't find anecdotal evidence convincing, either. That's why without a specific set of steps to reproduce a memory problem (including especially the version of Firefox you are running and the operating system you are using), I don't believe there is one. Can you explain exactly which pages you have open (give exact URLs)? We can all try to leave the pages open all day and see how much memory Firefox is using tomorrow morning.