Re:New Chrome 3 beta with 30% improvement
by ajross on Thursday August 06, 2009 @12:33PM
(#745811)
The headline is missing 22%!
(For those curious but not curious enough to read the article, it's a 30% improvement in the V8 and Sunspider javascript benchmark scores.)
Also: windows only, no OS X or Linux binary for this beta. Sigh. I want to think they're serious about being cross platform, but stuff like this tells me that "everyone" in the developer group is on windows and the other platforms are still just sometimes-updated ports.
Re:New Chrome 3 beta with 30% improvement
by icefox on Thursday August 06, 2009 @12:41PM
(#745834)
Not everyone, just the vast majority of chrome developers are on Window and develop in Visual Studio. There are a few people working hard on the Linux and Windows ports, but Windows is the important port.
Re:New Chrome 3 beta with 30% improvement
by ajross on Thursday August 06, 2009 @01:27PM
(#745945)
Right, that's why I put "everyone" in quotes. Obviously there are people on the team tasked with maintaining the secondary platforms, and they're no doubt working hard.
But the point was more architectural: clearly Chrome is a windows program. With a little porting work, you can make it run elsewhere. So that's the development model they've picked. Do the work on windows, then hand it off. There's no simultaneous work going on, and features travel in only one direction. If a windows developer commits something that breaks Linux or OS X, it's "someone else's" problem to fix during porting.
Contrast this to Firefox, for example, where (the occasional glitch notwithstanding) the assumption is that every build works on every platform. People on Linux care about regressions on windows, etc... The model is that the software is portable first, and not "a port".
Re:New Chrome 3 beta with 30% improvement
by icefox on Thursday August 06, 2009 @12:43PM
(#745838)
30% improvement in branding! :) On Linux it now defaults to not have your window manager paint the boarder and handles it all by itself (what could go wrong?) so that Chrome looks the same everywhere even if it really doesn't fit into the rest of the Linux desktops.
(For those curious but not curious enough to read the article, it's a 30% improvement in the V8 and Sunspider javascript benchmark scores.)
Also: windows only, no OS X or Linux binary for this beta. Sigh. I want to think they're serious about being cross platform, but stuff like this tells me that "everyone" in the developer group is on windows and the other platforms are still just sometimes-updated ports.
Wow. I'm guessing this is because Google really wants this on Win PCs, not because the developers are such VS fans.
But the point was more architectural: clearly Chrome is a windows program. With a little porting work, you can make it run elsewhere. So that's the development model they've picked. Do the work on windows, then hand it off. There's no simultaneous work going on, and features travel in only one direction. If a windows developer commits something that breaks Linux or OS X, it's "someone else's" problem to fix during porting.
Contrast this to Firefox, for example, where (the occasional glitch notwithstanding) the assumption is that every build works on every platform. People on Linux care about regressions on windows, etc... The model is that the software is portable first, and not "a port".
http://www.chromeexperiments.com/detail/balldroppings/
Run it on the current stable Chrome, then try the new beta.
Side note: I'm using Safari 4 and anytime I am viewing a gif my browser beachballs and crashes. Does this happen for anyone else?
Same for Ubuntu. I find it funny that I use a Google browser to view a Google site (YouTube) but can't see the videos.