ell, iirc Classic mode is basically running the complete OS 9 in a VM. But by this logic, Windows is perfectly backwards compatable because you can run any previous version inside VMware.
No, not really. First, it isn't just OS 9 in a VM. It has a virtual address space, but it is not emulated, and parts of it were rewritten to filter down properly to OS X drivers, etc. This means that it breaks stuff that tries to be really low-level, so it isn't entirely backwards compatible. Blitters in particular don't work right, because they can't lock the screen for blitting; you get artifacts of the OSX desktop.
However, I agree with your definition, more or less, and by this standard, OS X is not very backward-compatible. This is, to some degree, a good thing, because under the hood, OS 9 sort of sucked. But Apple did a very good job making it look backwards compatible, just as they did with the 68K-PPC transition, and just as MS did with the Win3.1-Win95-WinNT transitions. Not perfect by a long shot, but good enough for a casual user to get by with and for a power user to work around.
a coin-operated video game manual and tech information archive
Instructions 1) Select manuals to be read. 2) Insert coin(s) to buy time. 3) To extend time, press the red button and insert more coins...... and don't you dare post a ??? PROFIT!!! joke after this...
Yeah, I've set up a server on my OS9 box at home, but it wasn't that reliable, nor fast, and I sorely missed SSH. I also made one on my OSX box at college, and it was great. I upgraded the preinstalled apache with a Fink installation of 2.0.44, and used it to serve all sorts of stuff, forums for class discussion, etc.
I also agree that OSX is a great OS (with the possible exception of the Finder), and almost rock-solid (it did panic a few times on me during high-workload such as rendering, and some other trouble, like kernel deadlocks that made procs unkillable, even by sudo kill -9). Other than, well, basically the Finder, it's easily the best OS I've used for personal computing stuff.
But yes, I'm talking about the hardware. For most serving applications (complete with CGI), the big stats you need (other than network connection) are memory size and speed, HD size speed, and integer unit speed. Memory size is fine, as is HD size and speed on the XServe. Integer speed is not great (better clock for clock, but the clock speed is slow), but probably sufficient for most apps. A plus for the Mac is that the PPC architecture has less expensive context switches than the X86, which could help against a flood of traffic (would be most important if you were DoSed).
However, the memory bus speed is horrendous compared to just about every competing architecture. And given that if your site gets slashdotted or something, you'll be spending most of your time futzing with CGIs, and getting stuff from the buffer cache... memory speed and intop speed are what'll make or break your day.
Now, you are right that for a 1U enclosure, the server isn't that bad in most of its specs. But the dual procs don't entirely make up for the slow bus speed. Further, the system is more expensive than an equivalent X86 system would be. This is really what I meant when I said it was unsuitable: in terms of hardware price/performance, you are generally worse off with the Mac.
However, as some posters pointed out, I was wrong. If you're a small corp, can run on just a few servers, it probably would be cheaper to get XServes, since you'd save on tech support and admin costs that way. If Apple gets their file perms right this time. Also, as I pointed out in my original post, these points could all pretty much be moot, except for cost, if they end up with a 970 or other souped-up IBM chip in them.
OK. I cede the point on renderfarms; I was really talking about web/db servers. Still, there are probably better modifications to be made for rendering, ie huge video cards and shitloads of (fast!!) RAM.
I can indeed point you to a reference on Altivec, namely this page in the 970 ArsTechnia article. Quote from the article: It appears that in the 970 the Altivec unit is sort of "tacked on" to the core. While the vector register file sits alongside the general purpose and floating-point register files for the purpose of keeping LOAD and STORE latencies down, the actual vector execution hardware is off on a different portion of the die, away from the vector register file and away from the rest of the execution core. This necessitates the addition of at least two extra stages to the vector pipeline: one stage at the beginning of the execution phase to actually move the instructions out to the vector execution unit and another stage at the end of the execution phase to move the instructions from the unit back to the group completion queue.
Also note that it says on the site that while the G4e can issue any two different vector ops in one cycle, the 970 can do it only if one of them is a permute, and in addition, the pipelines are longer.
Seriously. I'm as big a Mac fan as the next guy, but the Mac is just not a server computer. Maybe if there're 970s at this WWDC that'll change a bit, but the current Moto PPC is just too underpowered, and its FSB is just too slow to be competitive. You can get more racks, but the dang things cost too much to be price-competitive.
The best thing about the PPC is its vector unit, and that's not all that useful on a server (POWER architecture doesn't even have it; it had to be hacked on in the 970, and is worse in some ways than the MOTO Altivec). You could use it to speed up your crypto, but it's really a bigger help with RC5 than say RC4 for obvious reasons, and doesn't give that much an advantage for AES or 3DES or just about any other code. Not unless they add a GF(2^32) vector multiply or something crazy to keep those weird benches up. I suppose it could be useful with RSA or DH, but that could take a good deal of assembly to optimize.
And you can run apache httpd on them (I run it on my eMac). Big whoop, you can run it on an AMD or whatever for half the price.
The OS is not designed to be a server, it's designed to be a personal use OS. Really, Apple will not do well in that market without major IBM help.
There is spam and there is spam. While all spam is annoying, I mind the type that suggests I visit some online embroidery shop (actually got such a spam once) or the like, much less than those with penis enlargements, women screwed by horses, and mortgage refinancing scams.
I wouldn't really care if my hypothetical kid were to see an ad for an embroidery shop. I'd care a bit if it were for some toy, because ads like that do seriously corrupt little kids, but then they get enough of that on TV. But bestiality porn? That would piss me off.
I wouldn't be so sure he's paranoid. His "coup" scenario is where someone gets a hold of his opponents' TIA logs and embarrasses them out of any hope for re-election, or even gets them to resign.
This sort of thing has happened many times before, minus the TIA. So with that much data available â" and yes, it's not unreasonable that a Congressman could do this, they have access to a whole lot of classified data â" you don't have to be paranoid to imagine a major "coup."
The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.
It doesn't take a terribly "liberal interpretation" of the 4th amendment to see a "right to privacy" here. I mean, it would be pretty extreme to claim this prevents surveillence in a public place, but I think TIA, Carnivore, etc. constitute unreasonable searches against people's papers and effects and are certainly done without probable cause.
This isn't as broad a "right to privacy" as some might like, but it's not a stretch at all to claim that it rules out trying to spy on as much of the country as you can manage.
Um. You know, it really doesn't matter what your keysize is after a certain amount, once the cheapest attack is to sneak into your house and boot in single-user or attach a keylogger. Or do you keep a smart-card under your pillow/in your pocket for that purpose?
Remember that for cheaper than the feds could break your key, they could subpoena you.
Most of my AIM friends fall into fairly large cliques. I mean, what's to stop them from using a PGP-like (non-)structure, where you just get introduced by your friends? Or are the masses too stupid to make this model practical (check fingerprints etc). Maybe those who actually need secure messaging will do this...
Your reasoning is off-base. You're correct that if you choose a "co-Sophie Germain prime" (you want p=2q+1, not q), the discrete logs problem becomes much nastier, because you have a group of order q, instead of smaller groups for each factor to break.
But there are much faster ways to do discrete logs than by brute force. They use quite advanced math, but the current best attacks on factoring use methods which can be applied to discrete logs with very little modification. Google "index calculus."
For pretty much exactly the reason you mentioned, DH is harder to break than RSA at the same keysize, but not nearly as hard to break as RSA at twice the keysize. And remember how quickly RSA-129 (a 426-bit or so key) was factored by distributed.
You could break DH-128 on a PC if you had to, or in a lab more quickly with some clever programming. It might take them a while, but even if it took weeks running as their PC screensaver, they'd have your whole conversation if they needed it. It'd be enough to keep the jackass down the hall from listening in on your cybersex, but not enough to hide your plan/idea/whatever confidential data you might send over IM/? from a hacker.
Easy to deal with. Put in the default comment block for that key, "This signature is for authentication purposes only. It is not legally binding."
In fact, you probably don't have to worry about it anyway. You aren't legally bound to stuff you write in say a postcard, even if you sign it, unless you put in a bunch of legalese.
- encryption algorithm: one-time pad should just about do it. - random key selection process: come on, if you have this quantum crypto thing running, you surely have enough money for a decent RNG. And anyway, you can use a quantum one with all the equipment you'd have. - "check if tapped" procedure (that quantum stuff): right. According to the guys doing this, 1/4 per bit, which means 1/2 per bit actually usable as key... you could probably mess with it a bit to get even better guarantees by adding a few extra bits.
So... this would be pretty much unbreakable. Not that ssh isn't also pretty much unbreakable (though you never know... those NSA...)
If so, 30 years from now, all communications might be so secure that we wouldn't have to worry about eavesdroppers
Nope. I mean, it wouldn't be so expensive today to encrypt point-to-point links with a stream cipher. But the problem is, it has to go through a router at some point. And you just have to put a bug in the router, have it copying traffic... this stuff is multi-stage, there's no way you could tell if the router were hacked/bugged from the timing.
I think if you're going to fantasize about a future with no eavesdroppers, you may as well fantasize about IPSec.
While POV-Ray doesn't benefit from dual procs (normally, I think there's a patch), faster procs would be a HUGE IMPROVEMENT when it takes some of my scenes 8 hours or more to render at screen resolution.
Hm. I assume that means their servers are down. I haven't been to that site in months, and then it was working properly (I went to get the floorplan of Random Hall).
Not just Bonzai, but all of those ads that try to 'trick' users into clicking through: fake 'X' buttons, bogus radio buttons, etc... How many people, once fooled into the company's web site, decide "Hey, I'm already here. Why not give them my credit card number for some useless piece of shit I don't need."
Well, the Bonzai people made it look like a button, so that you didn't know you were being tricked. It just looked like clicking through a regular dialog...
Interesting point on the music store... but I disagree with that about Linux. Windows may be more intuitive (though not so much as MacOS), but for programming I prefer Linux to Windows, and find Windows' lack of a decent command line a damning flaw. I mean, sure, X11 sucks, but Windows doesn't have anything which comes even close to bash.
Slashdot Mirror
ell, iirc Classic mode is basically running the complete OS 9 in a VM. But by this logic, Windows is perfectly backwards compatable because you can run any previous version inside VMware.
No, not really. First, it isn't just OS 9 in a VM. It has a virtual address space, but it is not emulated, and parts of it were rewritten to filter down properly to OS X drivers, etc. This means that it breaks stuff that tries to be really low-level, so it isn't entirely backwards compatible. Blitters in particular don't work right, because they can't lock the screen for blitting; you get artifacts of the OSX desktop.
However, I agree with your definition, more or less, and by this standard, OS X is not very backward-compatible. This is, to some degree, a good thing, because under the hood, OS 9 sort of sucked. But Apple did a very good job making it look backwards compatible, just as they did with the 68K-PPC transition, and just as MS did with the Win3.1-Win95-WinNT transitions. Not perfect by a long shot, but good enough for a casual user to get by with and for a power user to work around.
a coin-operated video game manual and tech information archive
... and don't you dare post a ??? PROFIT!!! joke after this...
Instructions
1) Select manuals to be read.
2) Insert coin(s) to buy time.
3) To extend time, press the red button and insert more coins...
Yeah, I've set up a server on my OS9 box at home, but it wasn't that reliable, nor fast, and I sorely missed SSH. I also made one on my OSX box at college, and it was great. I upgraded the preinstalled apache with a Fink installation of 2.0.44, and used it to serve all sorts of stuff, forums for class discussion, etc.
I also agree that OSX is a great OS (with the possible exception of the Finder), and almost rock-solid (it did panic a few times on me during high-workload such as rendering, and some other trouble, like kernel deadlocks that made procs unkillable, even by sudo kill -9). Other than, well, basically the Finder, it's easily the best OS I've used for personal computing stuff.
But yes, I'm talking about the hardware. For most serving applications (complete with CGI), the big stats you need (other than network connection) are memory size and speed, HD size speed, and integer unit speed. Memory size is fine, as is HD size and speed on the XServe. Integer speed is not great (better clock for clock, but the clock speed is slow), but probably sufficient for most apps. A plus for the Mac is that the PPC architecture has less expensive context switches than the X86, which could help against a flood of traffic (would be most important if you were DoSed).
However, the memory bus speed is horrendous compared to just about every competing architecture. And given that if your site gets slashdotted or something, you'll be spending most of your time futzing with CGIs, and getting stuff from the buffer cache... memory speed and intop speed are what'll make or break your day.
Now, you are right that for a 1U enclosure, the server isn't that bad in most of its specs. But the dual procs don't entirely make up for the slow bus speed. Further, the system is more expensive than an equivalent X86 system would be. This is really what I meant when I said it was unsuitable: in terms of hardware price/performance, you are generally worse off with the Mac.
However, as some posters pointed out, I was wrong. If you're a small corp, can run on just a few servers, it probably would be cheaper to get XServes, since you'd save on tech support and admin costs that way. If Apple gets their file perms right this time. Also, as I pointed out in my original post, these points could all pretty much be moot, except for cost, if they end up with a 970 or other souped-up IBM chip in them.
OK. I cede the point on renderfarms; I was really talking about web/db servers. Still, there are probably better modifications to be made for rendering, ie huge video cards and shitloads of (fast!!) RAM.
I can indeed point you to a reference on Altivec, namely this page in the 970 ArsTechnia article. Quote from the article: It appears that in the 970 the Altivec unit is sort of "tacked on" to the core. While the vector register file sits alongside the general purpose and floating-point register files for the purpose of keeping LOAD and STORE latencies down, the actual vector execution hardware is off on a different portion of the die, away from the vector register file and away from the rest of the execution core. This necessitates the addition of at least two extra stages to the vector pipeline: one stage at the beginning of the execution phase to actually move the instructions out to the vector execution unit and another stage at the end of the execution phase to move the instructions from the unit back to the group completion queue.
Also note that it says on the site that while the G4e can issue any two different vector ops in one cycle, the 970 can do it only if one of them is a permute, and in addition, the pipelines are longer.
Seriously. I'm as big a Mac fan as the next guy, but the Mac is just not a server computer. Maybe if there're 970s at this WWDC that'll change a bit, but the current Moto PPC is just too underpowered, and its FSB is just too slow to be competitive. You can get more racks, but the dang things cost too much to be price-competitive.
The best thing about the PPC is its vector unit, and that's not all that useful on a server (POWER architecture doesn't even have it; it had to be hacked on in the 970, and is worse in some ways than the MOTO Altivec). You could use it to speed up your crypto, but it's really a bigger help with RC5 than say RC4 for obvious reasons, and doesn't give that much an advantage for AES or 3DES or just about any other code. Not unless they add a GF(2^32) vector multiply or something crazy to keep those weird benches up. I suppose it could be useful with RSA or DH, but that could take a good deal of assembly to optimize.
And you can run apache httpd on them (I run it on my eMac). Big whoop, you can run it on an AMD or whatever for half the price.
The OS is not designed to be a server, it's designed to be a personal use OS. Really, Apple will not do well in that market without major IBM help.
There is spam and there is spam. While all spam is annoying, I mind the type that suggests I visit some online embroidery shop (actually got such a spam once) or the like, much less than those with penis enlargements, women screwed by horses, and mortgage refinancing scams.
I wouldn't really care if my hypothetical kid were to see an ad for an embroidery shop. I'd care a bit if it were for some toy, because ads like that do seriously corrupt little kids, but then they get enough of that on TV. But bestiality porn? That would piss me off.
They say that they use plasma to "deflect high-energy beams." So that would be a yes, at least for the phasers.
I wouldn't be so sure he's paranoid. His "coup" scenario is where someone gets a hold of his opponents' TIA logs and embarrasses them out of any hope for re-election, or even gets them to resign.
This sort of thing has happened many times before, minus the TIA. So with that much data available â" and yes, it's not unreasonable that a Congressman could do this, they have access to a whole lot of classified data â" you don't have to be paranoid to imagine a major "coup."
The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.
It doesn't take a terribly "liberal interpretation" of the 4th amendment to see a "right to privacy" here. I mean, it would be pretty extreme to claim this prevents surveillence in a public place, but I think TIA, Carnivore, etc. constitute unreasonable searches against people's papers and effects and are certainly done without probable cause.
This isn't as broad a "right to privacy" as some might like, but it's not a stretch at all to claim that it rules out trying to spy on as much of the country as you can manage.
Oop. you're right
Um. You know, it really doesn't matter what your keysize is after a certain amount, once the cheapest attack is to sneak into your house and boot in single-user or attach a keylogger. Or do you keep a smart-card under your pillow/in your pocket for that purpose?
Remember that for cheaper than the feds could break your key, they could subpoena you.
Most of my AIM friends fall into fairly large cliques. I mean, what's to stop them from using a PGP-like (non-)structure, where you just get introduced by your friends? Or are the masses too stupid to make this model practical (check fingerprints etc). Maybe those who actually need secure messaging will do this...
Your reasoning is off-base. You're correct that if you choose a "co-Sophie Germain prime" (you want p=2q+1, not q), the discrete logs problem becomes much nastier, because you have a group of order q, instead of smaller groups for each factor to break.
But there are much faster ways to do discrete logs than by brute force. They use quite advanced math, but the current best attacks on factoring use methods which can be applied to discrete logs with very little modification. Google "index calculus."
For pretty much exactly the reason you mentioned, DH is harder to break than RSA at the same keysize, but not nearly as hard to break as RSA at twice the keysize. And remember how quickly RSA-129 (a 426-bit or so key) was factored by distributed.
You could break DH-128 on a PC if you had to, or in a lab more quickly with some clever programming. It might take them a while, but even if it took weeks running as their PC screensaver, they'd have your whole conversation if they needed it. It'd be enough to keep the jackass down the hall from listening in on your cybersex, but not enough to hide your plan/idea/whatever confidential data you might send over IM/? from a hacker.
Easy to deal with. Put in the default comment block for that key, "This signature is for authentication purposes only. It is not legally binding."
In fact, you probably don't have to worry about it anyway. You aren't legally bound to stuff you write in say a postcard, even if you sign it, unless you put in a bunch of legalese.
[Obligatory -- IANAL]
- encryption algorithm: one-time pad should just about do it.
- random key selection process: come on, if you have this quantum crypto thing running, you surely have enough money for a decent RNG. And anyway, you can use a quantum one with all the equipment you'd have.
- "check if tapped" procedure (that quantum stuff): right. According to the guys doing this, 1/4 per bit, which means 1/2 per bit actually usable as key... you could probably mess with it a bit to get even better guarantees by adding a few extra bits.
So... this would be pretty much unbreakable. Not that ssh isn't also pretty much unbreakable (though you never know... those NSA...)
If so, 30 years from now, all communications might be so secure that we wouldn't have to worry about eavesdroppers
Nope. I mean, it wouldn't be so expensive today to encrypt point-to-point links with a stream cipher. But the problem is, it has to go through a router at some point. And you just have to put a bug in the router, have it copying traffic... this stuff is multi-stage, there's no way you could tell if the router were hacked/bugged from the timing.
I think if you're going to fantasize about a future with no eavesdroppers, you may as well fantasize about IPSec.
While POV-Ray doesn't benefit from dual procs (normally, I think there's a patch), faster procs would be a HUGE IMPROVEMENT when it takes some of my scenes 8 hours or more to render at screen resolution.
Those who don't learn by history are doomed to repeat it. ...And those who do learn by it are doomed to see it coming.
kind of reminds me of changing some autocorrect to replace a bunch of intransitive words with "fornicate." Like, for instance, "type".
I can fornicate at 120 wpm...
it's not a RAID if it's striped. that's why you do a real RAID.
Hm. I assume that means their servers are down. I haven't been to that site in months, and then it was working properly (I went to get the floorplan of Random Hall).
I'm a Harvard student anyway.
Not just Bonzai, but all of those ads that try to 'trick' users into clicking through: fake 'X' buttons, bogus radio buttons, etc...
How many people, once fooled into the company's web site, decide "Hey, I'm already here. Why not give them my credit card number for some useless piece of shit I don't need."
Well, the Bonzai people made it look like a button, so that you didn't know you were being tricked. It just looked like clicking through a regular dialog...
Interesting point on the music store... but I disagree with that about Linux. Windows may be more intuitive (though not so much as MacOS), but for programming I prefer Linux to Windows, and find Windows' lack of a decent command line a damning flaw. I mean, sure, X11 sucks, but Windows doesn't have anything which comes even close to bash.
And on Slashdot, I'd say you're outnumbered.
Yeah. Or find out which bathroom to use?