IE8 Beats Other Browsers In Laptop Battery Life 263
WARM3CH writes "AnandTech tested a laptop with an AMD CPU, a laptop with an Intel CPU, and a netbook to compare battery life while running Internet Explorer 8, Opera 10, Firefox 3.5, Safari 4, and Chrome. They tested on simple web pages and flash-infested ones. IE8 had the best battery life on both laptops (followed by FF + AdBlock), and Safari had the worst battery life. On the netbook, Chrome was slightly ahead of IE8. The report concludes: 'Overall, Internet Explorer and Firefox + AdBlock consistently place near the top, with Chrome following closely behind. Opera 10 Beta 3 didn't do as well as Opera 9.6.4, and in a couple quick tests, it doesn't appear that the final release of Opera 10 changes the situation at all. Opera in general — version 9 or 10 — looks like it doesn't do as well as the other major browsers. Safari is at the back, by a large margin, on all three test notebooks. We suspect that Safari 4 does better under OS X, however, so the poor Windows result probably won't matter to most Safari users.'"
Marketing Tag: Get infected faster... (Score:0, Funny)
I for one... (Score:4, Funny)
I for one welcome our new battery life saving overl-... wait... what?
No no no! (Score:3, Funny)
Infections last longer with IE8. Read the summary if not the article. Sheesh!
Yeah right (Score:5, Funny)
It's all about wget on single user mode.
Re:Yeah right (Score:2, Funny)
RMS, is that you?
Uhh.. whiskey tango foxtrot? (Score:2, Funny)
Slow news day, guys? I mean, seriously -- who is going to choose a browser based on how long it'll keep working in a laptop battery life test? And what's the control group for this test, anyway? In the real world, some guy decides he wants some ramen and suddenly my wifi connectivity goes to crap. What if it's really bright in the room and I have to turn the brightness up on the LCD? Well and truly, there's about a hundred things more important than which browser I'm using that affect battery life.
Now, I'm off to make some ramen and make my neighbor scream bloody murder as his high resolution download of some porn star stalls halfway through and he's stuck staring at an incomplete image for the next three minutes exactly. muwaaaaah....
New IE8 Commercial (Score:5, Funny)
Evil-Doer (played by Jerry Seinfeld): Agent Jack Bauer! How can this be? That laptop had three, maybe four minutes of battery life left on it, at most! How could you possibly have downloaded those files in time?!
Jack Bauer: Simple.
Bauer turns to face the camera, which quickly zooms in on his face.
Jack Bauer: I used Internet Explorer 8.
A giant explosion rocks the screen, and a huge Internet Explorer logo appears.
Announcer: Internet Explorer 8. Because on the Internet, seconds matter.
Of course it does... (Score:4, Funny)
Re:This is a better test. (Score:3, Funny)
Firefox has been shown to make it so under not especially exotic conditions.
That's why it's number 1.
Re:So in theory (Score:2, Funny)
Really? Loading less content is faster than loading more content???
Tell me, how did you figure that out???
Re:2% difference... big deal. (Score:1, Funny)
Actually, what you said sounds right but is complete BS (except that they did a poor job gathering the data). There's no need to do any of that. What they needed to do was make sure to randomize all their variables (which order of sites they visit, which order of browsers they test etc). Then run it multiple times. This is far easier in this scenario than controlling a lot of variables (& far more representative of real-world since in your scenario you're significantly changing the environment by how long the network card is active, how quickly the data is fetched on typical websites etc).
Then, having done this, run the tests multiple times. Like 5-10 would be a good start. This was their main mistake. They noted variation at 3% & then ran their tests twice and claimed that somehow they were within 1% indicating a significant lack of understanding of statistics. What the multiple tests give them is an average with a confidence interval - typically, one would use a 95% confidence interval (two standard deviations) and then graph the results displaying the errors bars which allows us to see whether or not there's a lot of variability in the results and whether or not the bars overlap significantly. Single numbers are useless metrics.
I am not a statistician, so perhaps someone who is can point out the mistakes in my methodology and how to improve it.
Re:So in theory (Score:4, Funny)
Aren't you looking back right now?
No.. If I was, I wouldn't be able to see the monitor.
Re:So in theory (Score:5, Funny)
It's like Ron Paul in the politics tab during the election.
Yeah, well, if you were using Firefox, you could get the RonPaulBlock plugin... ~
Re:So in theory (Score:3, Funny)