They'll only do two logic board replacements (which is what they did for me). On the third one, they give you a shiny new computer, or a $1200 (or so, depends on what you bought originally) credit towards anything you want. I got the new computer. No complaints here.
1) And I still use more than 4GB of ram regularly in my programs, and it's still a bloody pain in the ass on the x86, and even more of a pain in the ass on PPC-32. I admit, I don't need 64-bit addressing. But I use 36 bits today, would use 37 if I could... and will use 40 soon enough. Honestly, if you wanna build me a 40 bit computer, I'd be happy for about five or ten years, probably... but that would be silly.
2) 128-bit? 256-bit? Where have you ever heard anyone seriously talking about this? Okay, besides slashdot. With current process technology and memory technology (and restricting to the now-universal power-of-two bits, rather than the old nine-bit-basis), the "sweet spot" seems to be 64-bit pointer operations, with which 64-bit integer operations come "for free", and 256-bit vector engines, which are restricted to 128-bit for compatibility reasons. Maybe, just maybe, we'll see 512 bit vector engines in the near future... I suspect IBM and such will skip 256, but 512 bit is a lot of memory latency, or a wider bus than we can really make today... and maybe, even less likely, we'll see 128 or 96 bit floats, though the number of people who need that, compared to the increase in chip real estate, is pretty damn small... but I can't imagine we'll see anything wider than 64 bits for integer and load/store data, ever... and, come to think of it, Itanium has me convinced that basically all new chips will continue to use 32 bit instructions.
Re:How are we going to explaing something this sub
on
G5 vs Opteron, Finally
·
· Score: 1
Also, it's not just coincidence that the G5 is less "penalized" for 32 bit code than the Opteron. When AIM (Apple, IBM, and Motorola) sat down to design an ISA, they designed a series of chips that would logically and effectively span the gap. When Intel sat down to design an ISA... well, I guess no one brought the donuts. In any case, yes, the performances of the G5 in 32 bit mode and the performance of the G5 in 64 bit mode are much, much more similar than the performances of the Opteron in 32/64 bit modes. This is the result of conscious design and, while it suggests that in 64 bit mode the Opteron will seem significantly faster than the G5 (well, noticeable. Okay, measureably?), un under current usage, which really is mostly 32 bit on both platforms, the G5 seems to put up a more than fair fight... and I'm not sure I really get the people who are complaining "rigged test" just because AIM can design an ISA and Intel can't.
are the new big thing. To look at two totally seperate domains, check out the YS Tech TMD fan (Dan's review) and Canon's Ring USM (Photo net review). This is clearly a technology with potential for anyone working in a certain formfactor who thinks they're making a high enough quantity that they can do custom motors instead of just buying the oldfashioned barrels... and now, it seems, it scales as well. I think we'll be seeing a lot more of this. (Is it so bad of me to want a monocycle driven with this kind of motor?)
Jaysyn, while it's not normal, it's more and more the way Apple seems to be doing business. The first thing you should do is call up the apple store and ask them. There's a good chance they'll give you an honest answer. There's really three possibilities: They'll tell you the truth, and tell you they're telling the truth; they'll tell you they don't know when it'll be there, and be telling the truth; or they'll tell you it's not going to be there until after christmas, which means that there's an update (probably) coming right after christmas, and they want to give you a better item at the same price, rather than something you'll be disappointed with within the month. In any case, call and ask; that's just a hint for deciphering what they tell you. Also, it's good to register disapproval by calling, because it is (in my opinion) shitty business practice.
You know that the resolution of this device (1024x768x24) corresponds to just over a gigabit of bandwidth, yes? I think we'd all "prefer" a magic way to do gigabit-level wireless networking, even without routing it... but I also think it's rather unlikely to be cheaper than established technologies.
I have the worst memory I know. But I use maybe 20 services regularly, all told. Of those, about 10 are what I consider "low importance"; my high-school e-mail address, my NYT account, my credit card access (guaranteed against fraud by law), so they all use one of two passwords (I've found that I need to keep two passwords to fit all the kinds of requirements for length, numbers, etc). I change those passwords every six months or so, although occasionally I'll discover one's been unchanged for a bit longer. For important stuff (banking, servers, etc.) I come up with a phrase ("here there be dragons"), chop it to password requirements ("h2BDn^^z"), and remember it, and change it every three months, at most. Seriously, think about how many phone numbers you remember. Is it that hard to remember 10 phrases, even if they do change regularly?
Yes. I've had passwords compromised, at times, in various ways. I pay my bills online, bank online, applied to grad school online... even buy stocks and bonds online. In addition, I run a handful of servers, control a few domain names, and have a bunch of e-mail accounts. It would really suck if someone got one of my passwords by sniffing when I was using telnet (yes, it happened; I was lazy, and SSH is a big speed hitwhen you're using a cell modem), and was then able to transfer all my money to their bank account, wouldn't it?
But if someone manages to somehow sniff the password for one system, you're hosed. One of the most basic rules of good passwords (in addition to having reasonable length and avoiding dictionary attacks) is to use different passwords for different things.
Re:I'm not circumcised. Will I ever get laid?
on
Real Security?
·
· Score: 1
If you have IE set to "resize images" (or whatever it calls it, when it tries to make sure your porn all fits on the screen), it will load the image until it realizes it's bloody huge, then resize it... and at that ratio, it's less than one pixel high resized, so you can't click on the "don't resize" button.
Slashdot Mirror
They'll only do two logic board replacements (which is what they did for me). On the third one, they give you a shiny new computer, or a $1200 (or so, depends on what you bought originally) credit towards anything you want. I got the new computer. No complaints here.
nothing beats that wow! factor
Try sex.
First dupe of the new year!
Warning: Parent post is long. Cliff's notes follow:
Lots of my friends are environmentalists!
I'm an environmentalist!
A bad one!
While I agree with everything you said, in combination with your sig it deeply disturbs me...
1) And I still use more than 4GB of ram regularly in my programs, and it's still a bloody pain in the ass on the x86, and even more of a pain in the ass on PPC-32. I admit, I don't need 64-bit addressing. But I use 36 bits today, would use 37 if I could... and will use 40 soon enough. Honestly, if you wanna build me a 40 bit computer, I'd be happy for about five or ten years, probably... but that would be silly.
2) 128-bit? 256-bit? Where have you ever heard anyone seriously talking about this? Okay, besides slashdot. With current process technology and memory technology (and restricting to the now-universal power-of-two bits, rather than the old nine-bit-basis), the "sweet spot" seems to be 64-bit pointer operations, with which 64-bit integer operations come "for free", and 256-bit vector engines, which are restricted to 128-bit for compatibility reasons. Maybe, just maybe, we'll see 512 bit vector engines in the near future... I suspect IBM and such will skip 256, but 512 bit is a lot of memory latency, or a wider bus than we can really make today... and maybe, even less likely, we'll see 128 or 96 bit floats, though the number of people who need that, compared to the increase in chip real estate, is pretty damn small... but I can't imagine we'll see anything wider than 64 bits for integer and load/store data, ever... and, come to think of it, Itanium has me convinced that basically all new chips will continue to use 32 bit instructions.
Also, it's not just coincidence that the G5 is less "penalized" for 32 bit code than the Opteron. When AIM (Apple, IBM, and Motorola) sat down to design an ISA, they designed a series of chips that would logically and effectively span the gap. When Intel sat down to design an ISA... well, I guess no one brought the donuts. In any case, yes, the performances of the G5 in 32 bit mode and the performance of the G5 in 64 bit mode are much, much more similar than the performances of the Opteron in 32/64 bit modes. This is the result of conscious design and, while it suggests that in 64 bit mode the Opteron will seem significantly faster than the G5 (well, noticeable. Okay, measureably?), un under current usage, which really is mostly 32 bit on both platforms, the G5 seems to put up a more than fair fight... and I'm not sure I really get the people who are complaining "rigged test" just because AIM can design an ISA and Intel can't.
are the new big thing. To look at two totally seperate domains, check out the YS Tech TMD fan (Dan's review) and Canon's Ring USM (Photo net review). This is clearly a technology with potential for anyone working in a certain formfactor who thinks they're making a high enough quantity that they can do custom motors instead of just buying the oldfashioned barrels... and now, it seems, it scales as well. I think we'll be seeing a lot more of this. (Is it so bad of me to want a monocycle driven with this kind of motor?)
The technology is out there, it's just expensive. Check out some of Kurzweil's toys.
Yes, but has your seeing eye dog?
You are no longer welcome here. Please fix, thanks.
Jaysyn, while it's not normal, it's more and more the way Apple seems to be doing business. The first thing you should do is call up the apple store and ask them. There's a good chance they'll give you an honest answer. There's really three possibilities: They'll tell you the truth, and tell you they're telling the truth; they'll tell you they don't know when it'll be there, and be telling the truth; or they'll tell you it's not going to be there until after christmas, which means that there's an update (probably) coming right after christmas, and they want to give you a better item at the same price, rather than something you'll be disappointed with within the month. In any case, call and ask; that's just a hint for deciphering what they tell you. Also, it's good to register disapproval by calling, because it is (in my opinion) shitty business practice.
Um, it's called parody. Or satire, if you prefer.
You know that the resolution of this device (1024x768x24) corresponds to just over a gigabit of bandwidth, yes? I think we'd all "prefer" a magic way to do gigabit-level wireless networking, even without routing it... but I also think it's rather unlikely to be cheaper than established technologies.
Canola. What a great name. Like chilean sea bass.
(Hint: They're both marketing names. Why did the marketroids think we'd not buy rapeseed, and argentinian toothfish?)
Some things are very likely beyond our capability, forever. Simulating a few tin cans and a couple tons of hot air ain't one of them.
Which greeks would those be? The ones that lived in sumeria?
You make the common mistake of assuming that "can't yet" is the same as "can't".
I have the worst memory I know. But I use maybe 20 services regularly, all told. Of those, about 10 are what I consider "low importance"; my high-school e-mail address, my NYT account, my credit card access (guaranteed against fraud by law), so they all use one of two passwords (I've found that I need to keep two passwords to fit all the kinds of requirements for length, numbers, etc). I change those passwords every six months or so, although occasionally I'll discover one's been unchanged for a bit longer. For important stuff (banking, servers, etc.) I come up with a phrase ("here there be dragons"), chop it to password requirements ("h2BDn^^z"), and remember it, and change it every three months, at most. Seriously, think about how many phone numbers you remember. Is it that hard to remember 10 phrases, even if they do change regularly?
Yes. I've had passwords compromised, at times, in various ways. I pay my bills online, bank online, applied to grad school online... even buy stocks and bonds online. In addition, I run a handful of servers, control a few domain names, and have a bunch of e-mail accounts. It would really suck if someone got one of my passwords by sniffing when I was using telnet (yes, it happened; I was lazy, and SSH is a big speed hitwhen you're using a cell modem), and was then able to transfer all my money to their bank account, wouldn't it?
But if someone manages to somehow sniff the password for one system, you're hosed. One of the most basic rules of good passwords (in addition to having reasonable length and avoiding dictionary attacks) is to use different passwords for different things.
Dude, just buy some scissors.
The last digit is a four!
Wait, no, it just got slashdotted before it fully loaded...
somewhere the pointillists giggle to themselves.
If you have IE set to "resize images" (or whatever it calls it, when it tries to make sure your porn all fits on the screen), it will load the image until it realizes it's bloody huge, then resize it... and at that ratio, it's less than one pixel high resized, so you can't click on the "don't resize" button.