You could spend all your budget on hardware, leaving none for software. Or you could spend it all on software, or consultants, or staplers. So what? The number of ways to screw up your business is infinite. After reading the article, I do not feel any smarter.
BTW, I thought it was called "best of breed", not "pure play".
It looks like the OS X kernel is going to remain 32-bit for the forseeable future, so that it can load 32-bit drivers. But a Core 2 Duo should allow you to run 64-bit userspace.
TFA seems to suggest that somehow AMD' hypertransport system gives it an edge over Intel's solution, however any external bus (i.e. hypertransport) is going to be slower than package-internal interconnects.
That might be true, but Intel has no package-internal interconnect. Kentsfield has a single FSB that connects the two processor dice and the northbridge; there is no on-package fast-path between the dice. 4x4 has twice as many memory channels and several HT links in addition.
Kentsfield may still end up faster, since it has a better core.
4x4 is just a different name for Opterons in Opteron motherboards. Feel free to buy a 4x4 mobo and drop in some Opteron 2212s. However, since the 2212 is almost $400 and the FX-70 is only $500, you're probably better off buying what AMD wants you to buy.
ZFS sounds great for flash, since (unlike journaling filesystems) it doesn't write anything twice during a transaction and it also spreads writes around because of copy-on-write. You'd get a lot of wear on the uberblocks, but the card's wear-leveling should take care of that. Too bad it's not available for Linux.
A lot of open standards (such as MPEG) are patented, and their owners will sue if you don't pay the licensing fee. So a covenant not to sue is even better than a regular open standard.
It does have a thermostat, but it appears to be set at a scorching temperature (like 95C). Rather than controlling the minimum fan speed directly, it would probably be better to just set the thermostat lower, but maybe the hackers haven't reverse-engineered that part yet.
Who's going to determine that the channels are empty and they don't cross over into legit stations?
There are several proposals.
There could be a map of what channels are used in what locations. Every device would contain GPS and a copy of this map, thus letting it determine what channels are empty in its location. Broadcasting map updates to all devices is left as an exercise for the reader.
All devices could use "listen-before-transmit" cognitive radio, where they listen for an ATSC TV signal on a channel and don't transmit if they can hear it. Devices could scan the channels to find a free one.
There are degrees of freedom. While OLPC may not satisfy Theo because it's only 100% free instead of 110%, it's still more free than OS X. (Besides, I can't imagine running OS X on that Geode processor anyway.)
There are already iPod knockoffs that are 20% cheaper than real iPods; people overwhelmingly buy iPods anyway. Why would this change in the (hypothetical) future where iPods are $50 and the knockoffs are $40?
The same tool was used to scan the source code of all the software, whether open-source or proprietary. Obviously this requires cooperation from proprietary software developers.
Slashdot Mirror
But why not use established systems with guaranteed update mechanisms?
Oracle Linux is basically RHEL4 and Oracle's Unbreakable Linux Network is basically up2date; these seem like established systems to me.
You could spend all your budget on hardware, leaving none for software. Or you could spend it all on software, or consultants, or staplers. So what? The number of ways to screw up your business is infinite. After reading the article, I do not feel any smarter.
BTW, I thought it was called "best of breed", not "pure play".
It looks like the OS X kernel is going to remain 32-bit for the forseeable future, so that it can load 32-bit drivers. But a Core 2 Duo should allow you to run 64-bit userspace.
TFA seems to suggest that somehow AMD' hypertransport system gives it an edge over Intel's solution, however any external bus (i.e. hypertransport) is going to be slower than package-internal interconnects.
That might be true, but Intel has no package-internal interconnect. Kentsfield has a single FSB that connects the two processor dice and the northbridge; there is no on-package fast-path between the dice. 4x4 has twice as many memory channels and several HT links in addition.
Kentsfield may still end up faster, since it has a better core.
4x4 is the Hummer of computers: designed as a military-industrial complex work machine, used by posers with more money than sense to show off.
4x4 is just a different name for Opterons in Opteron motherboards. Feel free to buy a 4x4 mobo and drop in some Opteron 2212s. However, since the 2212 is almost $400 and the FX-70 is only $500, you're probably better off buying what AMD wants you to buy.
ZFS sounds great for flash, since (unlike journaling filesystems) it doesn't write anything twice during a transaction and it also spreads writes around because of copy-on-write. You'd get a lot of wear on the uberblocks, but the card's wear-leveling should take care of that. Too bad it's not available for Linux.
We could call it Sealand 2.0.
A lot of open standards (such as MPEG) are patented, and their owners will sue if you don't pay the licensing fee. So a covenant not to sue is even better than a regular open standard.
Yes, my understanding is that ECMA is some sort of fast track to ISO. It may be faster to go through ECMA and then ISO than to go straight to ISO.
In theory you can get a DVI ADD2 card.
Hey, it worked great for .NET.
It does have a thermostat, but it appears to be set at a scorching temperature (like 95C). Rather than controlling the minimum fan speed directly, it would probably be better to just set the thermostat lower, but maybe the hackers haven't reverse-engineered that part yet.
GPUs are designed for graphics, but folding isn't graphics. That's why it's non-obvious.
Who's going to determine that the channels are empty and they don't cross over into legit stations?
There are several proposals.
There could be a map of what channels are used in what locations. Every device would contain GPS and a copy of this map, thus letting it determine what channels are empty in its location. Broadcasting map updates to all devices is left as an exercise for the reader.
All devices could use "listen-before-transmit" cognitive radio, where they listen for an ATSC TV signal on a channel and don't transmit if they can hear it. Devices could scan the channels to find a free one.
Too bad 256QAM isn't usable for wireless due to the lower SNR. Some wireless systems use 64QAM, but only at short range (1 mile?).
Imagine high-speed wireless internet with the range of broadcast TV.
It doesn't work. In a 6MHz channel you will be lucky to get 20Mbps; sharing that over hundreds of square miles would suck.
Cable companies are going all digital as soon as the box prices come down enough.
Sorry, that business plan was patented by Lemelson 30 years ago.
There's a big different IMO between voluntary IP licensing and patent trolling. Although at least Transmeta isn't threatening any graduate students.
Obviously Intel would settle, like they did with AMD and Centaur.
It basically becomes a question of how much modern userland stuff from OS X you want to throw in on top of the kernel; that's where the bloat is.
That's also the only good part of OS X. (OK, IOKit looks nice, too.)
There are degrees of freedom. While OLPC may not satisfy Theo because it's only 100% free instead of 110%, it's still more free than OS X. (Besides, I can't imagine running OS X on that Geode processor anyway.)
There are already iPod knockoffs that are 20% cheaper than real iPods; people overwhelmingly buy iPods anyway. Why would this change in the (hypothetical) future where iPods are $50 and the knockoffs are $40?
The same tool was used to scan the source code of all the software, whether open-source or proprietary. Obviously this requires cooperation from proprietary software developers.