...PC hardware keeps getting better, why would you want to give things up, if software keeps getting better?
This is what happened in movie theatres — once digital projection started, the movies theatres started saying: "Actually, we can make the seats better." We'll see the same with software taking advantage of new PC hardware.
But the direction of HTML 5 seems to be towards giving the browser and web apps better access to PC hardware, like the GPS. Does that make a difference?
There is very little difference between what happens in a browser and what happens on the rest of your PC, if it's a hardware peripheral. All you're talking about is: can you get to it from a browser?
What about privacy issues with the next generation of browsers, where PC hardware will be accessible from the web?
Take access to sensors. Up until about a year ago, everybody said: "See! This is why using a browser is great. Because all this stuff goes on on your PC, and the browser is an island and doesn't know or have any ability to access that sensor, so you're totally protected." That was an asset.
Now everybody is saying: "But we actually want your browser to know about your location, so now we need an API or a plugin model or an extension to a standard, so the browser can get to it."
But wait a minute — just before that, you were telling me: "This is so cool. You can have a GPS in your laptop, and the browser can respect your privacy because it doesn't know." Now you're saying the browser should know, but you need to finish the privacy part. You can't just say: "Net positive, everybody wins."
So I think of it as a continuum. There's no difference with the browser knowing something about hardware — whether it can it read and write files, or get to your memory, or use any peripherals. It's the responsibility of the whole set of software to let you know.
In Windows, the sensor API warns you that something is using the GPS. If [Internet Explorer] uses the GPS, we'll use that API, and there is no way around it. The API always warns you. That's the approach that we take.
That said, time will tell. Cultures and different ways and different regions of the world will react differently to what hardware peripherals do and when you get notified of what.
There's a lot of interest in smaller devices and tablets. Do you want to bring Windows to smaller devices in a lighter-weight version?
There are many things mixed in there. Right now, the thing that's happening on x86 is our ability to go from a ridiculously mega supercomputer game machine to a laptop that has the same specs to a machine with a five-inch screen that's [touch-sensitive].
We have literally every size of PC: five-inch, seven-inch, eight-inch, 10-inch, 11-inch, 12-inch, 15-inch, 19-inch...
At some point you run out of dimensions; there are only so many ways to take a really big sheet of LCD glass and cut it. We could have a triangle PC and use the left-over corner!
But in terms of a lightweight operating system? Android claims it has advantages, like longer battery life.
The thing to keep in mind with battery life is that the number-one consumer of battery life, right now, is the screen. If you take a phone and put it underneath a really big screen, the battery life drops geometrically, because that's how screens work. That's a really big blocker today.
Even if you do everything in solid state and have the lowest-power-consumption chip, you're still lighting up a lot of light bulbs.
That's the first thing. So basically, if we go to a 10-inch screen, we geometrically increase our power consumption just for the screen.
Of course, once you got to a 10-inch screen, you need more memory to drive the screen, and so all these little things start to add up.
Plus, if you're going to use it all day... well, you don't use your phone all day. You have to compare the talk time on your phone, not the standby time.