Well, you appear to be heavily invested in Chrome so it's likely that nothing I can say will change your mind.
No and to the contrary. I'm not heavily invested in Chrome as I don't use it as my default browser on Windows. But I do regard it is highly interesting from a technical point of view with lots of potential for the future. And I'm quite open about discussing the various advantages and disadvantages of each browser. Unlike disciples of the Church of Apple I don't have any emotional attachment to the software I use. ;-)
Quote:
To imply that "I don't know what I am talking about" though is both rude and inaccurate.
Yes, that was rude and unnecessary and I'd like to express my sincere apologies for that. But claiming that a fundamentally different architecture could be imposed on an existing browser (be it Safari or Firefox) easily was just incongruous. You'd have to rewrite the whole code base to switch from a mostly single-threaded browser to a multi-process paradigm.
Quote:
All testing lately indicates that the next shipping version of Safari will be faster than Chrome on all benchmarks, and not just by a little bit either.
I actually never disputed this point. This also matches my experience with Chrome 1.0 and the Webkit Nighties. But this was beside the point I made. Both browsers are ridiculously fast and their differences in synthetic benchmarks are in absolute numbers so small that it doesn't have noticeable effects when using both browsers (except some Placebo effect maybe).
Quote:
Overall, Safari is known to be a very stable browser and it's in it's third iteration as I also noted.
Doing a quick search for "Safari 3.2 crash" actually shows quite a few results. But in my experience stock Safari is quite stable. It doesn't crash more or less than Firefox, which I use as well.
Quote:
It's also wildly popular on the Mac and Chrome isn't even out yet, so it's a safe bet Safari will continue to be popular for at least a little while.
This I also never disputed. It's bundled with the OS and Internet Explorer has taught us that most people are content with a limited feature set as long as it gets the job done (i.e. displaying web pages).
Quote:
Despite Chrome's technical advantage on multiple processes (and nowhere did I argue that this wasn't a great thing), overall stability is of course going to be greater in Safari. Chrome offers crash protection from errant Flash applications etc. but it's a beta browser in all new code. I find it extremely unlikely that it would be more stable than Safari except perhaps in that one tiny area that it was actually designed to be. If that's your extraordinary claim, (that it is more stable) then it's up to you to provide some kind of proof of that. But wait, we don't even have a product to compare yet do we?
"Designed to be"? It's a major architectural advantage both from a UI responsiveness as well as stability. No matter what you throw at a process, others won't be affected. The Chromium process itself (doing the IPC among others) is the only weak point and it is protected from its siblings due to the process separation. As long as Google keeps the core code small and tidy stability can be much easier ensured than in a singe process browser where each part (rendering, JavaScript Engine, plug-ins) can crash it.
Quote:
My argument would also be that the number of times someone is actually running a web application is so few at this point that the crash protection is not going to come into play very often at all in terms of stability comparisons. At least in the first round, the stability crown is going to go to the browser with the most mature code.
Quite a few million people use web mail services at least and I hear that services like Facebook, Picasa and Google Earth are very popular as well. And mature code means nothing. Case in point: Quicktime, which is quite an old framework, was obviously so riddled with problems that despite being "mature" Apple decided that it needed a rewrite. A well thought-out architecture can protect you from a lot of problems from the very beginning.
Quote:
Feel free to throw a few more insults my way though if it makes you feel better, and weave whatever conspiracy theories you like.
Well I don't have time to do all this point by point arguing and I don't think it's very productive anyway.
You are answering me line by line here, mostly arguing that my characterisation is nowhere near anything you said, but in fact it wasn't meant to be. You criticised my statements forcing me to defend them line by line, not the other way around. I wasn't intending to paraphrase what you said, only to defend what I originally said that you disagreed with.
For instance I am arguing that Safari is a very stable browser and (arguably at least) more stable than Chrome will end up being on the first iteration. But I am only saying that because of my first more general statement to the same end, and the fact that you are arguing that it isn't. Yet you mention almost in the same breath that Safari is "rock solid" in your experience. And we also both know that Chrome (for the Mac) doesn't even exist yet so it's a bit silly to argue about it one way or the other.
All in all, it seems like the argument devolves into a "he said, she said" kind of thing in which we are almost arguing the same points anyway.
I continue to stand by my general remarks which is that Chrome is a bit over-hyped for me at this point and that it is in essence a "one trick pony" that exists more as a tech demo than as a real alternative to any shipping browser. Google said as much when they released it with the exception that they also noted it's "minimalistic design approach" which as I've already noted, is at odds with what the majority of the public seem to want from a browser over the last ten years or so.
I don't hate Chrome or wish it ill in the market, I just don't think gushing praise over a tech demo that is not even available is really warranted. My original criticism was engendered by the hyperbole of the article when it implied that Chrome was already some kind of serious competition to Safari when it's not even in Beta and further that both browsers would be involved in some kind of "smack-down" browser war later this year when it ships. It might happen, but I wouldn't count on it.
IMO this is a classic case of the tech press getting ahead of themselves and promoting an equivalency that doesn't actually exist between two products that aren't even shipping yet.
Chrome is already used more then safari, not bad for a new browser. Safari is just failing to lure new users.
I do not see safari 4.0 improving these stats.
Want to know what the most damning piece of evidence is to Chrome getting too far? The fact that according to this nearly 20% of the people on the web still use IE6.
IE6 has to be the most awful piece of crap browser ever made and yet one out of every five people still use it. This is a significant resistance to change which doesn't bode very well for any new browser.
I just hope they implement a decent RSS reader. I'm very fond of Safari's approach, but Safari's RSS tends to hang under stress (like when coming back from a vacation to find hundreds of unread articles scattered over 30-50 different feeds.)
Firefox's approach (smart bookmarks) is really fast, but it doesn't know about read vs. unread articles and doesn't retain articles that expire from the feed.
Actually, this browser should never make it to the iTunes App store as Apple forbids any applications that have the same functionality as the original iPhone apps.
Unless they change their minds for their good friends Google, but in that case it's unfair for the many apps that have been blocked before.
The only thing Chrome has going for it at the moment is the tabs running in their own memory space gimmick which is easily duplicated and could easily appear in "rival" browsers before the Mac version of Chrome is even out the door.
Is this really an advantage? Won't it require more system overhead?
Chrome is an interesting browser and I sincerely hope Google pushes its development to where it needs to be. That said, the following statement is telling: "[It can] render most Web pages pretty well". To be frank, that's not a very satisfying outcome. I know this statement was made with respect to the Mac version of Chrome, but the truth is, this statement is true of Chrome Windows version too. That is, Chromium (the browser engine) can render most pages pretty well. The flip-side of that is that there are some pages it does not render well, and there are a lot more of these pages than one could say of Safari, Firefox, or even IE.
Anyone who has participated in bug reporting with respect to Chromium knows that the browser engine has a long way to go. Even though it uses web-Kit as its core, this should not be taken to indicate that Chrome can render every page that Safari can. In fact, it lags well behind Firefox as well. Unfortunately the browser engine is lagging quite far behind these competitors, and as a developer I'm less than enthused because as it stands, Chrome is just one more weird engine I have to support.
One area that I think is very important, and very poorly done in Chrome, is in rendering images. Chromium still doesn't correctly support even basic aspects of PNG images in conjunction with CSS classes. For example, it has a significant problem rendering PNG graphics in a containing object that has an opacity of less than 100%. It also has problems rendering text in such circumstances, and has some severe ClearType problems. Beyond that, it is missing support for various CSS properties. For me, this means that I have to change the basic design of an aplication or web site because things that work just fine in Safari and Firefox and IE 7+ will not work in Chrome. So, my reaction so far is "that's nice, but don't expect me to embrace it until it is as least as good as its competitors.
To be fair, it would appear that Google IS working to fix many of these issues, but I think calling their browser "releasable" is very premature in its current state. Yet, on the PC side anyway, they have already done just that and that bugs me greatly, knowing what I do about how much work is yet to be done to get their rendering engine even up to par with browsers like Safari and Firefox. Yes, the process model is cool... but that doesn't help a lot when the browser isn't capable of actually rendering the kinds of rich user experiences that would require a process model like Chrome's. Its somewhat like a Ferrari with a four cylinder engine - it looks good, and is built to perform in every respect, but it lacks the engine to demonstrate that performance capability.
Is this really an advantage? Won't it require more system overhead?
Well, it depends on your view of things. It will require more system overhead, so if your hard drive is swapping memory before you even open your browser, it'll be a disadvantage - however its not a big memory hit. The first advantage of this architecture is that as you close your tabs for pages you've read, that memory space if freed up (for real) and the other is that if there is bad code or something crashed on a page, it doesn't take the browser down, just that tab - both things don't occur on Safari or Firefox because of their design.
Long term, I'd expect that both Firefox and Safari will move towards the same architecture Google has here, but for the most part (other than the tab crash thing), most people won't even notice such a change. In the end, they're all capable browsers and its really just what you prefer to run in. On my Mac I actually like Camino (which is based off the same code as Firefox).
How many MB of Ram does Safari suck up on your machine? 50 - 65 ish?
I never close Safari & typically have 7 or 8 tabs open all the time. It sits around 315MB, but then I have 4GB of RAM & I never notice any slowness in anything I do. Only times I get a spinning wheel on Safari is on rare occasions as a flash app is opening, but never for even a second.
Father in law uses Safari with tons of tabs open on an XP machine running 512MB & he finds the performance to be much better than IE & Firefox. Maybe it's more of a Vista issue?
I suspect the RAM usage is a result of the typical Apple approach, leave lots cached in memory so that if you jump back to it the page comes up quickly. This is of course only practical if you have memory to spare (but of course most of the machines they are testing on do).
Someone just needs to make this a big deal to the webkit team, I don't recall ever even knowing it was an issue until it came up here.
I never close Safari & typically have 7 or 8 tabs open all the time. It sits around 315MB, but then I have 4GB of RAM & I never notice any slowness in anything I do. Only times I get a spinning wheel on Safari is on rare occasions as a flash app is opening, but never for even a second.
Father in law uses Safari with tons of tabs open on an XP machine running 512MB & he finds the performance to be much better than IE & Firefox. Maybe it's more of a Vista issue?
I suspect the RAM usage is a result of the typical Apple approach, leave lots cached in memory so that if you jump back to it the page comes up quickly. This is of course only practical if you have memory to spare (but of course most of the machines they are testing on do).
Someone just needs to make this a big deal to the webkit team, I don't recall ever even knowing it was an issue until it came up here.
My Macbook Pro, Macbook, and iMac are all maxed out with 4 GB of memory; perhaps that's why I've never noticed Safari's large memory footprint. I certainly can see how that would be a problem for people running a lot less memory.
Chrome is pretty much the best browser on my Mac. And it only runs in emulation in Parallels!! Still faster than Safari, without the 600MB memory fatness.
How many MB of Ram does Safari suck up on your machine? 50 - 65 ish?
Quote:
Originally Posted by Erunno
After a day of heavy browsing (including lots of Flash, JavaScript and pictures)? Easily over 400 MB RAM. And even after closing all tabs Safari won't release it. And before someone asks: I don't use any of the so-called unofficial extensions as I had some bad experiences with stability in the the past.
Is this really an advantage? Won't it require more system overhead?
A process comes with a fixed upfront cost in terms of memory and process creation takes longer than spawning a new thread. But since we are talking about microseconds at best there's no noticeable effect for the user. And whatever slight increase in memory usage this architecture might cause is far overshadowed by the gains, i.e. memory is released immediatly and completely, memory fragmentation is a non-issue, plug-in crashes don't even crash the tab in which they run, the interface doesn't lock up as some CPU intensive calculations are made and it scales well with a rising number of cores/processors without having to do further programming work. The Chromium core process uses 20 MB RAM currently and that's lower than a newly opened Safari. \
I suspect the RAM usage is a result of the typical Apple approach, leave lots cached in memory so that if you jump back to it the page comes up quickly. This is of course only practical if you have memory to spare (but of course most of the machines they are testing on do).
I've read this explanation repeatedly and frankly I don't buy it. Even if you close all tabs except for an empty one and wait for hours Safari won't release the memory. Other browsers have used these kind of caches for years and don't show this abnormal behaviour as they have policies to free the memory cache according to whatever conditions the freeing algorithm specifies (with varying degrees of success). Firefox, for instance, scales the memory cache depending on your amount of system memory.
Safari's atrocious memory usage has been an issue for years now. Either Apple is not willing to fix this problem or the error is not easy to fix. I'm still hoping for Safari 4 though.
... Safari's atrocious memory usage has been an issue for years now. Either Apple is not willing to fix this problem or the error is not easy to fix. I'm still hoping for Safari 4 though.
If by "years now" you mean that Safari had this problem in 2005-2006 you are definitely right.
Personally, I can find no mention of this problem after 2006 though, are you sure it still happens? How about a link to some evidence of that?
Comments
How many MB of Ram does Safari suck up on your machine? 50 - 65 ish?
On my iBook 1Ghz OS X 10.4.11 it takes roughly 10MB per thread.
As for Safari 4, I am not falling for the hype, so much was said about safari 3 and its still lacking.
No matter what happens, more users will use Chrome on OS X then they will use safari on a PC.
Actually looking at stats
http://www.w3schools.com/browsers/browsers_stats.asp
Chrome is already used more then safari, not bad for a new browser. Safari is just failing to lure new users.
I do not see safari 4.0 improving these stats.
This is good news, gives people options.
As for Safari 4, I am not falling for the hype, so much was said about safari 3 and its still lacking.
No matter what happens, more users will use Chrome on OS X then they will use safari on a PC.
Actually looking at stats
http://www.w3schools.com/browsers/browsers_stats.asp
Chrome is already used more then safari, not bad for a new browser. Safari is just failing to lure new users.
I do not see safari 4.0 improving these stats.
Certainly looks different here:
http://en.wikipedia.org/wiki/Usage_s...f_web_browsers
Well, you appear to be heavily invested in Chrome so it's likely that nothing I can say will change your mind.
No and to the contrary. I'm not heavily invested in Chrome as I don't use it as my default browser on Windows. But I do regard it is highly interesting from a technical point of view with lots of potential for the future. And I'm quite open about discussing the various advantages and disadvantages of each browser. Unlike disciples of the Church of Apple I don't have any emotional attachment to the software I use. ;-)
To imply that "I don't know what I am talking about" though is both rude and inaccurate.
Yes, that was rude and unnecessary and I'd like to express my sincere apologies for that. But claiming that a fundamentally different architecture could be imposed on an existing browser (be it Safari or Firefox) easily was just incongruous. You'd have to rewrite the whole code base to switch from a mostly single-threaded browser to a multi-process paradigm.
All testing lately indicates that the next shipping version of Safari will be faster than Chrome on all benchmarks, and not just by a little bit either.
I actually never disputed this point. This also matches my experience with Chrome 1.0 and the Webkit Nighties. But this was beside the point I made. Both browsers are ridiculously fast and their differences in synthetic benchmarks are in absolute numbers so small that it doesn't have noticeable effects when using both browsers (except some Placebo effect maybe).
Overall, Safari is known to be a very stable browser and it's in it's third iteration as I also noted.
Doing a quick search for "Safari 3.2 crash" actually shows quite a few results. But in my experience stock Safari is quite stable. It doesn't crash more or less than Firefox, which I use as well.
It's also wildly popular on the Mac and Chrome isn't even out yet, so it's a safe bet Safari will continue to be popular for at least a little while.
This I also never disputed. It's bundled with the OS and Internet Explorer has taught us that most people are content with a limited feature set as long as it gets the job done (i.e. displaying web pages).
Despite Chrome's technical advantage on multiple processes (and nowhere did I argue that this wasn't a great thing), overall stability is of course going to be greater in Safari. Chrome offers crash protection from errant Flash applications etc. but it's a beta browser in all new code. I find it extremely unlikely that it would be more stable than Safari except perhaps in that one tiny area that it was actually designed to be. If that's your extraordinary claim, (that it is more stable) then it's up to you to provide some kind of proof of that. But wait, we don't even have a product to compare yet do we?
"Designed to be"? It's a major architectural advantage both from a UI responsiveness as well as stability. No matter what you throw at a process, others won't be affected. The Chromium process itself (doing the IPC among others) is the only weak point and it is protected from its siblings due to the process separation. As long as Google keeps the core code small and tidy stability can be much easier ensured than in a singe process browser where each part (rendering, JavaScript Engine, plug-ins) can crash it.
My argument would also be that the number of times someone is actually running a web application is so few at this point that the crash protection is not going to come into play very often at all in terms of stability comparisons. At least in the first round, the stability crown is going to go to the browser with the most mature code.
Quite a few million people use web mail services at least and I hear that services like Facebook, Picasa and Google Earth are very popular as well. And mature code means nothing. Case in point: Quicktime, which is quite an old framework, was obviously so riddled with problems that despite being "mature" Apple decided that it needed a rewrite. A well thought-out architecture can protect you from a lot of problems from the very beginning.
Feel free to throw a few more insults my way though if it makes you feel better, and weave whatever conspiracy theories you like.
This won't be necessary. :-)
No and to the contrary. ...
Well I don't have time to do all this point by point arguing and I don't think it's very productive anyway.
You are answering me line by line here, mostly arguing that my characterisation is nowhere near anything you said, but in fact it wasn't meant to be. You criticised my statements forcing me to defend them line by line, not the other way around. I wasn't intending to paraphrase what you said, only to defend what I originally said that you disagreed with.
For instance I am arguing that Safari is a very stable browser and (arguably at least) more stable than Chrome will end up being on the first iteration. But I am only saying that because of my first more general statement to the same end, and the fact that you are arguing that it isn't. Yet you mention almost in the same breath that Safari is "rock solid" in your experience. And we also both know that Chrome (for the Mac) doesn't even exist yet so it's a bit silly to argue about it one way or the other.
All in all, it seems like the argument devolves into a "he said, she said" kind of thing in which we are almost arguing the same points anyway.
I continue to stand by my general remarks which is that Chrome is a bit over-hyped for me at this point and that it is in essence a "one trick pony" that exists more as a tech demo than as a real alternative to any shipping browser. Google said as much when they released it with the exception that they also noted it's "minimalistic design approach" which as I've already noted, is at odds with what the majority of the public seem to want from a browser over the last ten years or so.
I don't hate Chrome or wish it ill in the market, I just don't think gushing praise over a tech demo that is not even available is really warranted. My original criticism was engendered by the hyperbole of the article when it implied that Chrome was already some kind of serious competition to Safari when it's not even in Beta and further that both browsers would be involved in some kind of "smack-down" browser war later this year when it ships. It might happen, but I wouldn't count on it.
IMO this is a classic case of the tech press getting ahead of themselves and promoting an equivalency that doesn't actually exist between two products that aren't even shipping yet.
This is good news, gives people options.
As for Safari 4, I am not falling for the hype, so much was said about safari 3 and its still lacking.
No matter what happens, more users will use Chrome on OS X then they will use safari on a PC.
Actually looking at stats
http://www.w3schools.com/browsers/browsers_stats.asp
Chrome is already used more then safari, not bad for a new browser. Safari is just failing to lure new users.
I do not see safari 4.0 improving these stats.
Want to know what the most damning piece of evidence is to Chrome getting too far? The fact that according to this nearly 20% of the people on the web still use IE6.
IE6 has to be the most awful piece of crap browser ever made and yet one out of every five people still use it. This is a significant resistance to change which doesn't bode very well for any new browser.
Firefox's approach (smart bookmarks) is really fast, but it doesn't know about read vs. unread articles and doesn't retain articles that expire from the feed.
Unless they change their minds for their good friends Google, but in that case it's unfair for the many apps that have been blocked before.
The only thing Chrome has going for it at the moment is the tabs running in their own memory space gimmick which is easily duplicated and could easily appear in "rival" browsers before the Mac version of Chrome is even out the door.
Is this really an advantage? Won't it require more system overhead?
Anyone who has participated in bug reporting with respect to Chromium knows that the browser engine has a long way to go. Even though it uses web-Kit as its core, this should not be taken to indicate that Chrome can render every page that Safari can. In fact, it lags well behind Firefox as well. Unfortunately the browser engine is lagging quite far behind these competitors, and as a developer I'm less than enthused because as it stands, Chrome is just one more weird engine I have to support.
One area that I think is very important, and very poorly done in Chrome, is in rendering images. Chromium still doesn't correctly support even basic aspects of PNG images in conjunction with CSS classes. For example, it has a significant problem rendering PNG graphics in a containing object that has an opacity of less than 100%. It also has problems rendering text in such circumstances, and has some severe ClearType problems. Beyond that, it is missing support for various CSS properties. For me, this means that I have to change the basic design of an aplication or web site because things that work just fine in Safari and Firefox and IE 7+ will not work in Chrome. So, my reaction so far is "that's nice, but don't expect me to embrace it until it is as least as good as its competitors.
To be fair, it would appear that Google IS working to fix many of these issues, but I think calling their browser "releasable" is very premature in its current state. Yet, on the PC side anyway, they have already done just that and that bugs me greatly, knowing what I do about how much work is yet to be done to get their rendering engine even up to par with browsers like Safari and Firefox. Yes, the process model is cool... but that doesn't help a lot when the browser isn't capable of actually rendering the kinds of rich user experiences that would require a process model like Chrome's. Its somewhat like a Ferrari with a four cylinder engine - it looks good, and is built to perform in every respect, but it lacks the engine to demonstrate that performance capability.
Anyway, this is just my perspective on things.
Have a blessed day.
Is this really an advantage? Won't it require more system overhead?
Well, it depends on your view of things. It will require more system overhead, so if your hard drive is swapping memory before you even open your browser, it'll be a disadvantage - however its not a big memory hit. The first advantage of this architecture is that as you close your tabs for pages you've read, that memory space if freed up (for real) and the other is that if there is bad code or something crashed on a page, it doesn't take the browser down, just that tab - both things don't occur on Safari or Firefox because of their design.
Long term, I'd expect that both Firefox and Safari will move towards the same architecture Google has here, but for the most part (other than the tab crash thing), most people won't even notice such a change. In the end, they're all capable browsers and its really just what you prefer to run in. On my Mac I actually like Camino (which is based off the same code as Firefox).
over the top imo
How many MB of Ram does Safari suck up on your machine? 50 - 65 ish?
I never close Safari & typically have 7 or 8 tabs open all the time. It sits around 315MB, but then I have 4GB of RAM & I never notice any slowness in anything I do. Only times I get a spinning wheel on Safari is on rare occasions as a flash app is opening, but never for even a second.
Father in law uses Safari with tons of tabs open on an XP machine running 512MB & he finds the performance to be much better than IE & Firefox. Maybe it's more of a Vista issue?
I suspect the RAM usage is a result of the typical Apple approach, leave lots cached in memory so that if you jump back to it the page comes up quickly. This is of course only practical if you have memory to spare (but of course most of the machines they are testing on do).
Someone just needs to make this a big deal to the webkit team, I don't recall ever even knowing it was an issue until it came up here.
I never close Safari & typically have 7 or 8 tabs open all the time. It sits around 315MB, but then I have 4GB of RAM & I never notice any slowness in anything I do. Only times I get a spinning wheel on Safari is on rare occasions as a flash app is opening, but never for even a second.
Father in law uses Safari with tons of tabs open on an XP machine running 512MB & he finds the performance to be much better than IE & Firefox. Maybe it's more of a Vista issue?
I suspect the RAM usage is a result of the typical Apple approach, leave lots cached in memory so that if you jump back to it the page comes up quickly. This is of course only practical if you have memory to spare (but of course most of the machines they are testing on do).
Someone just needs to make this a big deal to the webkit team, I don't recall ever even knowing it was an issue until it came up here.
My Macbook Pro, Macbook, and iMac are all maxed out with 4 GB of memory; perhaps that's why I've never noticed Safari's large memory footprint. I certainly can see how that would be a problem for people running a lot less memory.
How many MB of Ram does Safari suck up on your machine? 50 - 65 ish?
After a day of heavy browsing (including lots of Flash, JavaScript and pictures)? Easily over 400 MB RAM. And even after closing all tabs Safari won't release it. And before someone asks: I don't use any of the so-called unofficial extensions as I had some bad experiences with stability in the the past.
My current numbers:
Firefox 3: 340MB 15 tabs open
Safari 3: 770MB 5 tabs open
Is this really an advantage? Won't it require more system overhead?
A process comes with a fixed upfront cost in terms of memory and process creation takes longer than spawning a new thread. But since we are talking about microseconds at best there's no noticeable effect for the user. And whatever slight increase in memory usage this architecture might cause is far overshadowed by the gains, i.e. memory is released immediatly and completely, memory fragmentation is a non-issue, plug-in crashes don't even crash the tab in which they run, the interface doesn't lock up as some CPU intensive calculations are made and it scales well with a rising number of cores/processors without having to do further programming work. The Chromium core process uses 20 MB RAM currently and that's lower than a newly opened Safari.
I suspect the RAM usage is a result of the typical Apple approach, leave lots cached in memory so that if you jump back to it the page comes up quickly. This is of course only practical if you have memory to spare (but of course most of the machines they are testing on do).
I've read this explanation repeatedly and frankly I don't buy it. Even if you close all tabs except for an empty one and wait for hours Safari won't release the memory. Other browsers have used these kind of caches for years and don't show this abnormal behaviour as they have policies to free the memory cache according to whatever conditions the freeing algorithm specifies (with varying degrees of success). Firefox, for instance, scales the memory cache depending on your amount of system memory.
Safari's atrocious memory usage has been an issue for years now. Either Apple is not willing to fix this problem or the error is not easy to fix. I'm still hoping for Safari 4 though.
Certainly looks different here:
http://en.wikipedia.org/wiki/Usage_s...f_web_browsers
Exactly.
The truth is, Chrome's share is only about 60% of where Safari's was over five years ago.
... Safari's atrocious memory usage has been an issue for years now. Either Apple is not willing to fix this problem or the error is not easy to fix. I'm still hoping for Safari 4 though.
If by "years now" you mean that Safari had this problem in 2005-2006 you are definitely right.
Personally, I can find no mention of this problem after 2006 though, are you sure it still happens? How about a link to some evidence of that?