Linus Torvalds (torvalds@linux-foundation.org) on 9/19/09 wrote:
---------------------------
>I think your conclusions are naive and likely wrong.
>
>Look at the data - Apple laptops failed more. That's the
>actual hard data

I also noted that the reasons for failure / motivation to bring the system in for repair weren't disclosed. If the Apple systems are primarily brought-in due to noisy fans, and the Sony systems are brought-in because the batteries exploded and burned down a conference hall, then the naive failure rate isn't relevant.

Aren't Apple laptops the majority share holder of the college/university generation? That seems like the demographic which uses their laptops the most. What, with all the Facebooking and Twittering and MSNing that students have to do these days, perhaps Apple's failure rates are skewed by something like this. I have no hard evidence, but this is an important question.

Or we can just jump to another reliability survey which puts Apple closer to the top:

http://www.channelinsider.com/c/a/News/Asus-IBMLenovo-Top-Notebook-Reliability-Ranking/


>Then using "customer satisfaction" as a way to gauge the
>rest is totally irrelevant. For several reasons. It does
>not make the hardware any more reliable or better built,
>it just means that the Apple store experience is better
>than talking to some call center person in India (or in
>Missouri).

Valid point.


>But it might even be more slippery than that. Look up the
>psychology of satisfaction. There's actually a positive
>correlation between satisfaction and "paid a high price".
>It's true of wines (there are tons of studies that say that
>wine tasters give higher points to more expensive wines),
>it's true of cars, and it's certainly true of Apple too.

"it might be more slippery than that ... and it's certainly true of Apple too." At what point during that paragraph did you switch from this being a hypothesis to a certainty? Your point stands. Apple has 80-90% of the $1000+ laptop market. That said, I wonder how the correlation between price and satisfaction is affected by having to bring a system in for repair.


>And btw, this shouldn't surprise you. It is in no way
>unique to computers. Look at just about any technology: the
>highly priced one is almost never more reliable or even
>"better". It has other attributes that generally make up
>for it, though - status symbolism and satisfaction being
>very high on the list. Look at audio equipment. Look at
>cars. Look at clothes.

Sure, once you make it out of the bargain-basement price territory. Look at low-end, 90s-era, passenger vehicles from American automotive manufacturers, and compare them to the 25-35% higher-priced passenger vehicles from Japan. It might be hard to find the 90s-era American vehicles, because most of them have been scrapped.

An Accord is larger and roomier than a Corolla. A BMW 328i is faster, sportier, has a better suspension than the Accord, and can be leather-clad. An Audi S4 is turbocharged, all-wheel drive and faster than the BMW 328i. The Audi RS8 is another step-up from the S4. The Mercedes S550 is like riding a pillow down the street, and the Porsche 911TT will beat almost any production vehicle around the track under $200K.

The idea that the costlier vehicle isn't demonstrably "better" in some important metric is false. The two issues are (a) how much of a premium are you paying for brand, and (b) what is your minimum bang/buck ratio?