If you could actually AFFORD the phone you buy a unlocked one.. Google tried to sell you one, none of you cheap bastards bought one.

The G2 is also available for $500 from T-Mobile free and clear with no contract. It will be interesting to see what justification T-Mobile comes out with for locking down the bootloader on the G2 when it is bought outright like this.

What is needed is a compelling reason for people to get out from under NAT. Most people are so used to the Microsoft way of doing things with their desktop computer being so functionless that they can only work in conjunction with a remote server to do anything of significance. They can't host their own web pages, they can't get email delivered directly to their computer, they can't make a voice call directly to another computer etc. There are tons of application that could be written if everything was on the net directly. Anyone for syncing all their computers (files, bookmarks, etc?) without having to copy them via a remotely located server? We need to start writing these applications and tell the people hiding behind NAT boxes that they are SOL till they get themselves real, routable addresses.

Hynix just feels bad about being too young and missing the Bubble Memory Revolution.

GPS's don't use UTC for the simple reason that using a discontinuous time system at the low-level is insane. The Russian GLONASS does have leap seconds, and every time a leap second get applied the system has hiccups (as expected). The GPS system simply keeps its own true seconds-since-the-GPS-epoch counter and never steps this for leap seconds. Adding the leap seconds is left for the display routines in the individual end-user GPS devices.

As I see it, it really doesn't matter what UTC does as long as computers implement the low-level (internal) timekeeping correctly. If un*x/linux were to have a true seconds-since-the-epoch counter in the kernel with no leap seconds then time difference calculations would be trivial. Each program wouldn't have to have special (and probably largely untested) logic to deal with the time discontinuity around leap second time. Leap seconds (just like daylight-savings-time/normal-time) could be added by the display routines that map seconds-since-the-epoch into a human readable time. There is no need to muck up the low-level timekeeping for these oddities. Only programs that print out the time would need to even know that a leap second occurred, and in most cases that would all be done by the library routines.

Dan Bernstein first tried to get this low-level stuff straightened out, but folks largely didn't care to fix the problem because POSIX essentially mandated one do things in the more complicated fashion. http://cr.yp.to/proto/utctai.html

You are assuming the powering down or removing of the SIM doesn't wipe the decryption key. If I were designing a more secure cell phone that's what I'd do. I'd keep the key on an internet server and send it to the phone after it powers up. If the phone got stolen / confiscated I'd send a wipe command that wiped the key from the phone if the power were still on, and simultaneously wiped the key from the server.

The question then becomes, why is some isolated ipv6-capable router not sending an ipv6 "host unreachable" message to the host that is attempting the off-site ipv6 connection attempt? Wouldn't a correctly written application see this "host unreachable" and then try an ipv4 connection?

Is this going to be the new asbestos? Inhaling short silicon nanotubes doesn't sound like a very good idea.

If you want to route things at speed why use something based on an anemic ARM chip running a few hundred megahertz when you have a multi gigahertz cpu at your disposal? I just dual-port my main computer and have it route and nat things. Routing works at least to 700 Mbits/sec. If I ever move to a country where gigabit ethernet to the home is available at a reasonable price (like in Japan where it is ~$100), then I'll have to revisit the routing situation.

(My setup is an athlon64 at 2Ghz, two linksys pci gigabit ethernet cards, fedora and iptables doing the NAT-ing. The computer is on 24/7 anyway because it serves web pages and accepts my email, so having it do the routing doesn't really increase my power bill. If power ever becomes an issue, I'll just move the server to an old laptop which will cut my power from 80watts to 20watts. That admittedly still isn't as low as an ARM chip, but you do get quite a bit more performance for your money.)

I owned a GM car that was prone to a stuck throttle at wide-open-throttle. The damn throttle-lever was never tightened at the factory and the lever would get stuck over the top of some parts hanging on the side of the carb. The factory trained monkeys they have doing warranty repairs never found the problem. It only got fixed when I finally decided to have a look myself.

The stuck throttle was pretty exciting. There is nothing like accelerating down city streets watching the slow traffic in front of you getting larger in your windshield. I can see why some people might panic. There is a part of you that just screams NOOOOOOO! On the other hand there is hopefully also the part of you that says, time to turn off the ignition, which is what I did.

The inclination to anyone that is used to driving a stick and likes their engine is to never hit the clutch when the engine is at WOT. It will almost certainly over-rev and destroy itself. I just turned the ignition off, coasted to a stop and then waited for my blood pressure to return to normal.

I find these claims of failed brakes etc very hard to take. Just how is the brake supposed to fail? The brakes are also much stronger than the engine, so would always win a tug of war. Which production car can do 0-60 faster than it can do 60-0?

I've never had a problem with the lid closed as long as I put the laptop on end with the exhaust port facing up. You do want to allow convection to cool the thing. Keeping it flat with the lid closed is going to be the worst possible orientation.

I second the laptop idea. My athlon-64 laptop running 64-bit linux sitting around lightly loaded takes about 15 watts as measured by the "kill-a-watt" power meter. I've often thought it would make more sense to run it as the server and save ~$200/year.

Watts
    47 compaq dv5000z laptop (bios, booting)
    22 compaq dv5000z laptop (fc5, yum update, login screen, 1Ghz w. ath0 ubiquiti card)
    18 compaq dv5000z laptop (fc5, idle, login screen, 1Ghz, ath0 ubiquiti card)
    15 compaq dv5000z laptop (fc5, idle, login screen, 1Ghz, screen blanked, ath0 ubiquiti card)

(Yes, these measurements are from many years ago. I expect similar or slightly better numbers for fedora-11/12.

I feel sorry for the folks putting this release together. They got handed a hot potato and it isn't clear how they are going to stabilize this distribution in time for the real release in a month. I installed the 64-bit version on two AMD machines (one laptop and one desktop) and both of them have issues with random lockups after 10 minutes or so. In addition nfs4 with autofs stopped working. Ironically default nfs4 was supposed to be one of the new features of this release. Last time this happened it was some weird nfs4 interaction with IPv6. I guess not even Red Hat is taking IPv6 seriously and making the developers eat their own dogfood.

In any case, I'll probably wait a while after the real release before I roll it out to the rest of the machines. It reminds me a bit of the old X11 releases where the word on the street was always to wait a month or two before installing it and let some other poor slob work out what is broken and how to work around it.

Blocking root is a bit of an over reaction. Blocking password, pam, kerberos, onetime and all that other rot is good enough for stopping the brute force attacks. Nobody is going to brute force root's 1k (or 2k) key in any usable timeframe. While one is editing the file, might as well put an ssh-level keepalive in there for 30 minutes to clean up dead connections and keep lame NAT boxes primed. Protocol 1 should probably also be turned off. It has a few known failure modes. Cranking down logingracetime keeps the bruteforce attacks from tying up too many resources. One should never see RSA login take 10 seconds on a modern system with the rsa key unlocked and stored in keyserver. Even typing the passowrd in realtime shouldn't take 10 seconds.

Protocol 2
LoginGraceTime 10
PermitRootLogin without-password
PubkeyAuthentication yes
PasswordAuthentication no
ChallengeResponseAuthentication no
KerberosAuthentication no
GSSAPIAuthentication no
UsePAM no
X11Forwarding yes
ClientAliveInterval 60
ClientAliveCountMax 30
AllowUsers root myaccount

Give them a couple of years and the digital ballot stuffing software will get better. The voter numbers should be waaaay up.

Whatever happened to the claim that the ads paid for the hardcopy papers and the cover price was just a token fee to assure the advertisers that the papers weren't likely to just be taken and thrown out unread? If that is true, why would they need to charge anything in order to deliver the paper electronically? Don't the ads more than cover the delivery costs?