Notices


Reply
Thread Tools
Posts: 114 | Thanked: 50 times | Joined on Oct 2006
#11
Originally Posted by benny1967 View Post
And of course, Google will not use this to further gather any data or direct people to their own online services...

... as somebody wrote in an Austrian news forum today: Google is obviously trying hard in becoming the world's most hated company instead of Microsoft. :/


EDIT: Actually, I get the feeling that during the last months Apple became #1 as the world's most hated company, with MS being #2. But this could be a local phenomenon here... Anyway, I certainly agree with it, I don't think MS is all that bad.
It's open source. So if it gather data from you, you will be able to know what and how.
You can also fork it to direct people to yahoo services
 
benny1967's Avatar
Posts: 3,790 | Thanked: 5,718 times | Joined on Mar 2006 @ Vienna, Austria
#12
Originally Posted by Boke View Post
It's open source. So if it gather data from you, you will be able to know what and how.
You can also fork it to direct people to yahoo services
technically correct, but invalid in practice. do people change firefox's default search engine so they don't feed google? do they avoid google online services? do they make any effort to block google analytics? no.

of course, I myself can do it (or use a different browser in the first place), but I doubt that not having my data hurts google in any way.
 
Benson's Avatar
Posts: 4,930 | Thanked: 2,272 times | Joined on Oct 2007
#13
Originally Posted by benny1967 View Post
technically correct, but invalid in practice. do people change firefox's default search engine so they don't feed google? do they avoid google online services? do they make any effort to block google analytics? no.

of course, I myself can do it (or use a different browser in the first place), but I doubt that not having my data hurts google in any way.
Then they must not be the most-hated company yet, or people would; and it does mean any technical advances can be used in other browsers. (I was mainly impressed with the JS VM, anyway, so that's good.)

And for those concerned about multi-process vs. multi-thread, think copy-on-write; it's not as big a difference as you might suppose.
 
Posts: 2,102 | Thanked: 1,309 times | Joined on Sep 2006
#14
The strength of Google's concept is that every tab is isolated from the rest by running as an own process. On the NIT, however, you don't have that much RAM available, so that it's better to have all tabs running in once process sharing as much resources as possible. That's why Nokia introduced "browserd".
How large is the overhead in reality? The binary code will be shared anyway as presumably it will mainly be located in a shared lib (to allow just this effect), then there's the per-process data which would otherwise be shared between the tabs - stuff like the menu items, window data/theme stuff, etc.

So how large would the extra overhead really be? I always thought the main disadvantage of using processes vs. threads is in communicating across them, which in this case shouldn't be a major thing as each "tab" would not really need to do much IPC.
 

The Following 2 Users Say Thank You to lardman For This Useful Post:
Bundyo's Avatar
Posts: 4,708 | Thanked: 4,649 times | Joined on Oct 2007 @ Bulgaria
#15
Originally Posted by pycage View Post
I somehow doubt that the Google browser concept would work well on limited platforms like the NITs. The strength of Google's concept is that every tab is isolated from the rest by running as an own process. On the NIT, however, you don't have that much RAM available, so that it's better to have all tabs running in once process sharing as much resources as possible. That's why Nokia introduced "browserd".
Also, from what I have seen, the Gecko engine is currently a bit faster than Webkit (but let's see what speed improvements Google will achieve with Webkit).

But what's definitely interesting for the NITs is Google's new approach of a JavaScript VM with JIT and proper garbage collection. Since it's opensource, I'm sure it won't take long until Firefox gets the same.
FF3.1 already has JIT with the introduction of tracemonkey. That's why the current nightlies are approximately 7x faster than 3.0 (with JIT switched on that is). I'm running it and can confirm it (though it is kinda unstable yet).
 
Bundyo's Avatar
Posts: 4,708 | Thanked: 4,649 times | Joined on Oct 2007 @ Bulgaria
#16
I'm a little worried about this "tab sandboxing" thing...Does it mean you can't open two tabs (windows, popups) and control one of them from the other?
 
benny1967's Avatar
Posts: 3,790 | Thanked: 5,718 times | Joined on Mar 2006 @ Vienna, Austria
#17
Originally Posted by Benson View Post
Then they must not be the most-hated company yet, or people would;
You mean the same way the majority of consumers don't use Windows anymore because they hate Microsoft?

Last edited by benny1967; 2008-09-02 at 15:21.
 

The Following User Says Thank You to benny1967 For This Useful Post:
Benson's Avatar
Posts: 4,930 | Thanked: 2,272 times | Joined on Oct 2007
#18
Originally Posted by benny1967 View Post
You mean the same way the majority of cunsumers don't use Windows anymore because they hate Microsoft?

Well, there's a lot more investment in switching OSes than browsers, and many do stop using IE; considering that changing settings in a browser is less still, proportionately more would be doing that (i.e. changing default search) if they hated Google as much.
 
epage's Avatar
Posts: 1,684 | Thanked: 1,562 times | Joined on Jun 2008 @ Austin, TX
#19
Originally Posted by pycage View Post
I somehow doubt that the Google browser concept would work well on limited platforms like the NITs. The strength of Google's concept is that every tab is isolated from the rest by running as an own process. On the NIT, however, you don't have that much RAM available, so that it's better to have all tabs running in once process sharing as much resources as possible. That's why Nokia introduced "browserd".
An interesting point that Google makes is that it can cut down on fragmentation and you can better track what the worst offenders in websites, browser features, and plugins.

These I think could be great for the tablets. With shared libraries and "fork"s COW, I imagine there would not be too much overhead to a lot of processes.
 
Posts: 168 | Thanked: 51 times | Joined on Jun 2007
#20
Well, isn't putting each on in its own process and each one having its own interface essentially just running multiple instances, but under one main window?

I spawn multiple copies of links rather than use one process with several windows. Then when I close that "tab" all the memory is reclaimed and I don't have to worry about any cached pages.

browserd, I think, is just the opposite. It is one rendering engine with the ability to respawn the main window. Closing the window doesn't even have any effect on the browser. Am I understanding this correctly?
 
Reply

Tags
browser, chrome, google, webkit

Thread Tools

 
Forum Jump


All times are GMT. The time now is 00:23.