Cool, I like how trolls disappear now. Fading to grey was kind of lame.

Incidentally, it's not entirely true that CDMA doesn't support voice and data at the same time. There is EV-DV, but it's never going to be put into use since Verizon is going to LTE instead. So on a technical matter, CDMA does support voice and data at the same time, just not EV-DO, and just not on any actual networks.

Hmm, I played Tales of Phantasia a few years back and ever since have been wanting to play another Tales game. I assume the original Symphonia is pretty good, right? I need to find a copy of it some time and play it.

@th4threat: Do you have any idea how IA64 works? There's a good reason it never took off, and that's because of VLIW. In a VLIW architecture, a bunch of instructions that can be run independently get stuffed into one word (hence Very Long Instruction Word) and then executed by the CPU at once. The problem is, this requires the compiler to make optimal chunks of instructions, and that is difficult because most tasks are not nearly parallel enough for this to be efficient.

On the other hand, in a pipelined, multiple issue (superscalar) architecture like x86, the instructions are compiled one after another, but when the CPU gets them it determines whether there are any dependencies, and if there aren't any they are executed at the same time, roughly speaking. This obviously requires more complicated hardware, but the compiler doesn't have to even know that it's happening, although the compiler can try to do things to optimize this like for example unrolling loops that don't have strict data dependencies.

Intel always claimed that with a better compiler Itanium would perform much better than x86, but the fact of the matter is that it's easier to optimize existing x86 compilers than write a very good IA64 compiler. Basically, VLIW shifts the burden of instruction level parallelism from the hardware to the compiler, and in general it seems to be harder to do it in the compiler stage than at runtime.

See http://en.wikipedia.org/wiki/VLIW , http://en.wikipedia.org/wiki/Superscalar , http://en.wikipedia.org/wiki/Instruction_level_parallelism for more info.

Yeah, the developers have been trying to figure out what the problem is but they haven't been successful thus far. They're considering disabling it for the final release if it's still buggy, although I wonder how people would react to having a feature disabled that was enabled in the beta. Still unsure whether they're going to remove the code completely or just pref it away.

Ouch, being called out as a sony fanboy by tmac of all people. I think you've reached the 8th circle of Joystiq comment hell.

It shouldn't be too hard, I've seen something similar on a really old Nokia smartphone type device and it was able to track the background moving fine.

Yeah, me too, meh. Well, even if one did come out this year (really? There's what, two months left in the year?) in Japan, it would probably be another year before it got to the US and it's not like it adds faster/more hardware like the DSi did. I mean, I'd be pissed if this were true but I think the DSi's screens are already big enough.

Yeah, everyone always does this. Note the RC1 appended to the name... But uh, I've been running Firefox 3.6 nightlies for several months and it's been pretty good so far, the nightlies almost always feel faster than the stable branch and the recent landing of Windows 7 tab previews is awesome, although currently a bit buggy.

This is a cool idea, but I have to wonder how long during the day a phone is even outside. I mean, for me it's in my pocket 99% of the day and I don't think the 10-15 minutes max I'm actually using it are enough to charge it. But I guess maybe if you had something like this you'd wear it on a lanyard thing to make sure it got sunlight.