Google Chrome

I don't think it will take over Firefox any time soon largely due to the massive amount of customisation FF already has in place that Chrome will have to build on over the coming months and years BUT what this DOES help is existing browsers.

Since Google's Chrome is open source they have said they're allowing people to nab bits from it to use on their own work, this means Firefox builds can implement these new technologies used in Chrome as well as IE if MS decide to do so.

Google has a great vision, to make the Web surfing experience secure and open, Firefox started the whole thing and Google is building on it further.

I like it and look forward to Firefox 4 having a built in JAVA engine as well as process tree to close crashed "tabs" instead of having to restart the whole browser! \m/
 
mrk said:
Since Google's Chrome is open source they have said they're allowing people to nab bits from it to use on their own work, this means Firefox builds can implement these new technologies used in Chrome as well as IE if MS decide to do so.
Click to expand...

That'll depend on the license - Open Source doesn't mean you can just take the code and use it anywhere you like.

look forward to Firefox 4 having a built in JAVA engin
Click to expand...

It'll be more like Firefox 4000 before it bundles Java.
 
Apparently each tab will open up in a new process. Not a new thread. A whole new process.

That will work well on Unix but not on Windows. Opening processes is about 100x more expensive than opening a thread. Threads are considered lightweight in Windows.

This will flop on Windows.
 
NathanE you mean Chrome will flop on windows?... will be sweet to try out on linux on my ps3
 
I don't see how it would flop on Windows, 2-4GB RAM is standard on virtually every PC now and as long as each Chrome process does not consume Firefox-style amounts of RAM there would be no real issues.
 
There's a CPU overhead in spawning processes as well as the memory duplication (take the number of tabs you have, and multiply that by the memory consumption of your browser with a single tab open), and Firefox itself is generally quite well-behaved, in the scheme of things.
 
tolien said:
There's a CPU overhead in spawning processes as well as the memory duplication (take the number of tabs you have, and multiply that by the memory consumption of your browser with a single tab open), and Firefox itself is generally quite well-behaved.
Click to expand...

Surely you're not going not going to notice that overhead on any modern PC anyway?
 
It depends how big the overhead is - Nathan's more qualified to comment on the Windows-specific side of things than me, but even 20MB a process (which isn't by any means high for a single browser Window) * 20 tabs adds up to something that's chewing a fair bit of RAM...
 
mr.sly said:
Surely you're not going not going to notice that overhead on any modern PC anyway?
Click to expand...

The memory overhead is not really the problem. It is the process creation overhead. Obviously a fast PC that is not really busy doing anything else and you aren't going to notice. But when your PC gets a bit bogged down there is a whole world of difference between asking it to open a new process instead of a new thread.

New threads in Windows are created almost effortlessly. There is still a cost but it is so so so much less than a process. To create a process requires many many parts of the kernel to be involved in the transaction.

Given that Google had a completely clean slate I really thought they would have come up with a better design than this.

Obviously there are things that Google can do to "hide" the design flaw. Such as always having a new process standing by for when the user makes a new tab. But what if some program on the local PC decides to launch say 5 http:// URI's for you? Not uncommon. That means Chrome (and the Windows kernel) is going to have to break a sweat to deliver those 5 new tabs in an acceptable time.

Discussion of memory usage of web browsers is mostly pointless. The kernel takes care of swapping out pages of memory that aren't being used. It's no biggie.

IE7 on Vista manages to do sandboxing without creating loads of processes for each tab. Not sure why Google couldn't have used the same design.

Not saying there is going to be a "OMG Chrome is unusable because tabs open so slowly" type situation. Merely pointing out that Google is speaking crap about their rivals having an inferior design. It is they who have the inferior design. Although from their point of view it may be an advantage in terms of ease of development across multiple platforms...



I would also question whether the future of "web applications" is in HTML/CSS/JS at all. I am thinking it is more likely to be in technologies like WPF, XAML, XBAP's and Silverlight...




Despite that this browser does look interesting and I hope the IE8 team pinches a few ideas ;)
 
You must log in or register to reply here.
Back
Top Bottom