That was sort of how NCR's SCSI CAM implementation worked. The BIOS on the SCSI card contained two drivers for the card (two because it had to work with DOS, Win 3.x, Win95, WinNT, SCO Unix, OS/2, and Netware, so we had a 16-bit driver and a 32-bit driver).
Designing the interface to the driver in the BIOS was a bit of a challenge, because it had to work with everything from the 5380 (a very low level SCSI chip--basically all it implemented was those parts of SCSI that were too fast to do in software efficiently, like REQ/ACK handshakes during data transfer) all the way to the 537xx, which was basically a SCSI coprocessor, capable of doing everything, including management of a queue of requests.
To use the BIOS driver, you had to load it into memory, and relocate it (a relocation table was included in the BIOS), and give it a table of callback functions that it could use to do things like enable/disable interrupts, allocate and free memory, map virtual to physical addresses, and stuff like that. So, for each supported OS, we had to write a driver that could do that. This part tended to be fairly simple.
This worked very well. You could take a Unix system with a 5380, and decide you needed better performance. Shutdown. Replace 5380 card with 53720 card. Reboot. You now have better SCSI performance.
A way I found helpful to think about this while desiging it was to think about this question: "If you could design the ultimate intelligent SCSI card, what would you want it to do? How would you interface between it and the OS?". We basically then designed for that interface, and that's what the driver in the BIOS implemented. The NCR cards were essentially intelligent SCSI cards that just happened to not have a CPU onboard...so they had to make use of the system CPU.
I went to the American Airline site, and the story is simply wrong. You can use the site just fine, without ever seeing that agreement. You only see the agreement when you try to sign up for their frequent flyer program.
There is nothing at all newsworthy here, for nerds, or anyone else. Come on, editors...don't accept junk like this.
I remember years ago there was a big flap in the computer press about how the leading candidates for HDTV standards that the FCC was considering would not work well with computers.
The reason for this was that the broadcasters and the TV manufacturers and pretty much everyone else remotely interested in HDTV standards had tons of lobbyists working full time to push their interests, except for companies in the computer industry. A couple computer companies had a couple of part-time lobbyists working on this, or something pathetic like that.
Microsoft is not doing something bad here. The ones doing something bad are all the other companies that should be on that list but aren't.
Counterexample. We have a counter that starts at 0. Each player on his turn adds an integer from 1 to 9 to the counter. The winner is the player whose turn puts the counter over 99.
That game is deterministic, sequential, no draws are possible, and with perfect play the second player wins.
Why do you assume that "Black can force a win" is not possible? I have not heard of anyone proving that chess with perfect play is not a loss for white.
Go pieces, once placed on the board, cannot move anymore. Chess pieces can still move from one place to the other. This means that as more and more Go pieces are placed on the board, there are less and less positions the computer has to consider
Uhm...so after 100 moves per side, Go is down to ~160 possible moves per ply...still way higher than Chess.
I'm amazed at the number of people making fun of people who want better sounding musical reproduction. Yes, most of us can't hear the difference between CD and these other formats. Hell, a lot of probably can't tell the difference between MP3 and CD.
However, there ARE people who can hear the difference. Why make fun of them wanting something that sounds better?
All you people making fun of them...what kind of computer to you have? After all, most people would not notice the difference between a ~1 GHz Celeron with a GeForce2 MX, and the latest Athlon or P4 and a GeForce 4 Ti4600. Anyone who has spent more than $1000 on their computer setup needs to shut up about the audiophiles.
Actually, DSD encoding, which SACD uses, was developed for the purpose of being a good enough format to archive from analog masters, so it is not surprising that Sony will be releasing a lot of old, great recordings.
Uhm...WTF? You are very confused about what a OTP is. You can't plug the suspected key phrase in somewhere and see if it makes sense, because it makes sense *everywhere*, as does every other possible phrase.
Going to Pentium (x86) would be a step backward, into a braindead and inefficient architecture, and probably cause a riot among developers
First of all, developers don't give a damn what the architecture of the CPU is. We use compilers nowadays. The people who write compilers might care, but the rest of just care how it performs, not how ugly or beautiful it is to assembly programmers.
Second, define "inefficient". Pentium is as fast or faster than Motorola's PowerPC, Sparc, Alpha, MIPS, and everything else except IBM's high-end version of PowerPC, according to SPEC, and pretty much every other benchmark.
Third, what the heck do drivers have to do with it? Going to x86 for the CPU doesn't mean they have to go to commodity PC hardware.
Fourth, regardless of my third point, they've ALREADY went to commodity PC hardware. Every peeked inside a Mac? It's basically a PC with a PowerPC motherboard in a nice case. The drivers don't care what the CPU is--see my first point.
Paying bounties to get third parties to do the work in dealing with a nuisance can be a good idea. It kind of reminds me of the laws that deal with short swing trading. Short swing trading is when you buy or sell a stock that you recently bought or sold. Certain officers of public companies are not allowed to do short swing trading. I forget the exact rule, but basically, you can't change the direction you are going (buying or selling) more than once every several months (I think it is six months).
So, for example, if Bill Gates sells some MS stock today, he can't buy MS stock tomorrow.
The way the SEC enforces this is very clever. The law is that any shareholder of the company can sue to nail a short swing trader. If the suit is successful, the short swing trader has to turn over to the company any profit they made, AND they have to pay the attorney fees of the suing shareholder. The profits are calculated in the least favorable (to the short swing trader) way--find the highest selling price he got in the last six months, and the lowest buying price...match those shares up, and count the difference as profit. So, if you buy at 100, sell at 90, buy at 80, and sell at 70, you have really lost 20, but as far as the short swing laws go, you made 10 (the sell at 90 less the buy at 80), and so you have to pay 10.
The final brilliant piece of the short swing law is that the shareholder who brings suit does NOT have to have been a shareholder at the time of the trading--they only have to be a shareholder at the time of the suit.
Combine that with the winner getting attorney fees, and what happens is that attorneys check the public records, find dumb corporate officers who tried to sneak in some short swing trading, go out and buy a share of the company to get standing to sue, and sue.
This has pretty much completely eliminated illegal short swing trading, with the SEC having to spend no money to track it down and enforce the law.
And the real kicker..."determine if they're 'rooted in the community' -- or have an unusual history that indicates a potential threat." So, if you didn't grow up in the same place your family has lived in for the last six or seven generations you must be a terrorist
Where did anything imply that something like that would be a determining factor? Go back last week on slashdot and read the articles on spam filtering by statistical methods. It sounds like this is really nothing different than that, except applied to terrorist filtering.
Sun's obsession with Microsoft doesn't hurt Microsoft, but it is going to kill Sun. While they are busy wasting all their resources "fighting" Microsoft (a company not even in the same market as they are for the most part), IBM is going to squash them on the high end, while Intel finishes taking the low and medium end hardware from them.
This is the second case to make it to the appeals court level, and so become interesting from a precedential point of view. The first was Procd v. Zeidenberg, in 1996.
What I don't understand is why shows like TNG didn't get help with the technobable. I recall when TNG was on, I used to go back to the student house I had lived in when I was at Caltech, and there was always a huge crowd for each episode. If the producers had come by Caltech, and asked, they would have had at least 50 people volunteering to give them good, free, technobable, that would not have been blatantly bogus, and would have been consistent with all the prior technobable.
Well, for a Hollywood movie, we can simplify this quite a bit. Jack Woodford's summary of the typical plot: "Boy meets girl. Girl gets boy into a pickle. Boy gets pickle into girl".
I've been using OS X at work, first to develop a product, and then for the last few months as a box to SSH to the Linux box I'm working on.
My main machine at home runs Linux, with KDE for the desktop.
I've also got a WinXP machine for when I do Windows programming for work at home, and, or course, for Everquest. I'm reasonably good at using all these from a user's point of view, although I've never done much RTFMing for the GUIs, just experimented.
I was an exclusive Mac user at home from 1985 to 1994, and a Unix user at work from 1981, so am reasonably familiar with them.
Here's what I've found. OS X is beautiful. However, it is full of little annoyances because Apple is stubborn, and won't admit that anyone else ever did anything better. E.g., little things like not allowing windows to resize from any side.
There's no doubt that KDE has a steeper learning curve, and is not as beautiful, but it is not that steeper, and once I've learned something, it generally works better on KDE. Basically, at the cost of being a little clunkier at some things, KDE gets in my way a lot less.
So, among technical users, I certainly have no trouble believing Linux is beating Mac on the desktop. However, among home users, I don't see it. It's just too hard for the average home user to acquire a Linux machine, compared to a Mac.
I recall a work of "art" (it must have been art...it was at an art gallery) called "A Matter of Time". It consisted of an aquarium with a goldfish in it, and a mechanism that, at random intervals, shot a spear into the aquarium from the top, and then pulled it out.
Since there are way more Windows installations than Unix installations, the fact that 24% of the migrations are from Unix and 31% are from Windows means that Linux is hitting Unix WAY harder than it is hitting Windows.
It lets everyone else run the same code. The idea's been around for decades, but Java is the first thing to do it and do it well
What about Emacs? (No, I'm not joking...I use vim, but Emacs does run everywhere, and programs written in its version of Lisp run everwhere Emacs runs).
This reminds me of the days when Apple mostly used SCSI peripherals. Back in those days, if you had a Mac, and bought a generic unformatted hard drive, and hooked it up to your Mac, it would not work. Apple's SCSI software refused to recognize drives that were not bought from Apple (it could tell, because Apple's drives had the IDENTIFY information changed to say APPLE instead of QUANTUM or MAXTOR or SEAGATE). So, you had to go buy third party SCSI driver software to use a generic drive with your Mac.
In the vary early days of SCSI, this made sense. The command set was not standardized, and drivers needed to be written for specific models of drives. However, it didn't take long for the drive manufactures to defined and implement the Common Command Set, and generic drivers became easy to write.
Compare to both Linux and Windows. Both of these provided generic drivers for various classes of SCSI device, and so you did NOT need any special software to use a bare-bones drive.
It was fairly easy in a debugger to get around Apple's brand check. They had left debugging symbols in their HD setup program, so you could just set a breakpoint on "strcmp" and watch for one where it was comparing "APPLE" to "QUANTUM" (or whatever your drive was) and change the return value to say it matched.:-)
Slashdot Mirror
Designing the interface to the driver in the BIOS was a bit of a challenge, because it had to work with everything from the 5380 (a very low level SCSI chip--basically all it implemented was those parts of SCSI that were too fast to do in software efficiently, like REQ/ACK handshakes during data transfer) all the way to the 537xx, which was basically a SCSI coprocessor, capable of doing everything, including management of a queue of requests.
To use the BIOS driver, you had to load it into memory, and relocate it (a relocation table was included in the BIOS), and give it a table of callback functions that it could use to do things like enable/disable interrupts, allocate and free memory, map virtual to physical addresses, and stuff like that. So, for each supported OS, we had to write a driver that could do that. This part tended to be fairly simple.
This worked very well. You could take a Unix system with a 5380, and decide you needed better performance. Shutdown. Replace 5380 card with 53720 card. Reboot. You now have better SCSI performance.
A way I found helpful to think about this while desiging it was to think about this question: "If you could design the ultimate intelligent SCSI card, what would you want it to do? How would you interface between it and the OS?". We basically then designed for that interface, and that's what the driver in the BIOS implemented. The NCR cards were essentially intelligent SCSI cards that just happened to not have a CPU onboard...so they had to make use of the system CPU.
There is nothing at all newsworthy here, for nerds, or anyone else. Come on, editors...don't accept junk like this.
The reason for this was that the broadcasters and the TV manufacturers and pretty much everyone else remotely interested in HDTV standards had tons of lobbyists working full time to push their interests, except for companies in the computer industry. A couple computer companies had a couple of part-time lobbyists working on this, or something pathetic like that.
Microsoft is not doing something bad here. The ones doing something bad are all the other companies that should be on that list but aren't.
Counterexample. We have a counter that starts at 0. Each player on his turn adds an integer from 1 to 9 to the counter. The winner is the player whose turn puts the counter over 99.
That game is deterministic, sequential, no draws are possible, and with perfect play the second player wins.
Why do you assume that "Black can force a win" is not possible? I have not heard of anyone proving that chess with perfect play is not a loss for white.
Uhm...so after 100 moves per side, Go is down to ~160 possible moves per ply...still way higher than Chess.
Correction: it is believed that perfect play leads to a draw. It has not been proven.
However, there ARE people who can hear the difference. Why make fun of them wanting something that sounds better?
All you people making fun of them...what kind of computer to you have? After all, most people would not notice the difference between a ~1 GHz Celeron with a GeForce2 MX, and the latest Athlon or P4 and a GeForce 4 Ti4600. Anyone who has spent more than $1000 on their computer setup needs to shut up about the audiophiles.
Actually, DSD encoding, which SACD uses, was developed for the purpose of being a good enough format to archive from analog masters, so it is not surprising that Sony will be releasing a lot of old, great recordings.
How is making a clone of the Archos Jukebox but using a smaller form-factor disk (which wasn't available when the Archos was designed) innovating?
Uhm...WTF? You are very confused about what a OTP is. You can't plug the suspected key phrase in somewhere and see if it makes sense, because it makes sense *everywhere*, as does every other possible phrase.
Double clicking goes back to at least Apple. It was in the original Mac in 1984.
First of all, developers don't give a damn what the architecture of the CPU is. We use compilers nowadays. The people who write compilers might care, but the rest of just care how it performs, not how ugly or beautiful it is to assembly programmers.
Second, define "inefficient". Pentium is as fast or faster than Motorola's PowerPC, Sparc, Alpha, MIPS, and everything else except IBM's high-end version of PowerPC, according to SPEC, and pretty much every other benchmark.
Third, what the heck do drivers have to do with it? Going to x86 for the CPU doesn't mean they have to go to commodity PC hardware.
Fourth, regardless of my third point, they've ALREADY went to commodity PC hardware. Every peeked inside a Mac? It's basically a PC with a PowerPC motherboard in a nice case. The drivers don't care what the CPU is--see my first point.
So, for example, if Bill Gates sells some MS stock today, he can't buy MS stock tomorrow.
The way the SEC enforces this is very clever. The law is that any shareholder of the company can sue to nail a short swing trader. If the suit is successful, the short swing trader has to turn over to the company any profit they made, AND they have to pay the attorney fees of the suing shareholder. The profits are calculated in the least favorable (to the short swing trader) way--find the highest selling price he got in the last six months, and the lowest buying price...match those shares up, and count the difference as profit. So, if you buy at 100, sell at 90, buy at 80, and sell at 70, you have really lost 20, but as far as the short swing laws go, you made 10 (the sell at 90 less the buy at 80), and so you have to pay 10.
The final brilliant piece of the short swing law is that the shareholder who brings suit does NOT have to have been a shareholder at the time of the trading--they only have to be a shareholder at the time of the suit.
Combine that with the winner getting attorney fees, and what happens is that attorneys check the public records, find dumb corporate officers who tried to sneak in some short swing trading, go out and buy a share of the company to get standing to sue, and sue.
This has pretty much completely eliminated illegal short swing trading, with the SEC having to spend no money to track it down and enforce the law.
Yes you can. Think about it.
Where did anything imply that something like that would be a determining factor? Go back last week on slashdot and read the articles on spam filtering by statistical methods. It sounds like this is really nothing different than that, except applied to terrorist filtering.
Sun's obsession with Microsoft doesn't hurt Microsoft, but it is going to kill Sun. While they are busy wasting all their resources "fighting" Microsoft (a company not even in the same market as they are for the most part), IBM is going to squash them on the high end, while Intel finishes taking the low and medium end hardware from them.
This is the second case to make it to the appeals court level, and so become interesting from a precedential point of view. The first was Procd v. Zeidenberg , in 1996.
Well, for a Hollywood movie, we can simplify this quite a bit. Jack Woodford's summary of the typical plot: "Boy meets girl. Girl gets boy into a pickle. Boy gets pickle into girl".
My main machine at home runs Linux, with KDE for the desktop.
I've also got a WinXP machine for when I do Windows programming for work at home, and, or course, for Everquest. I'm reasonably good at using all these from a user's point of view, although I've never done much RTFMing for the GUIs, just experimented.
I was an exclusive Mac user at home from 1985 to 1994, and a Unix user at work from 1981, so am reasonably familiar with them.
Here's what I've found. OS X is beautiful. However, it is full of little annoyances because Apple is stubborn, and won't admit that anyone else ever did anything better. E.g., little things like not allowing windows to resize from any side.
There's no doubt that KDE has a steeper learning curve, and is not as beautiful, but it is not that steeper, and once I've learned something, it generally works better on KDE. Basically, at the cost of being a little clunkier at some things, KDE gets in my way a lot less.
So, among technical users, I certainly have no trouble believing Linux is beating Mac on the desktop. However, among home users, I don't see it. It's just too hard for the average home user to acquire a Linux machine, compared to a Mac.
Every so often, they had to get a new fish.
Since there are way more Windows installations than Unix installations, the fact that 24% of the migrations are from Unix and 31% are from Windows means that Linux is hitting Unix WAY harder than it is hitting Windows.
What about Emacs? (No, I'm not joking...I use vim, but Emacs does run everywhere, and programs written in its version of Lisp run everwhere Emacs runs).
In the vary early days of SCSI, this made sense. The command set was not standardized, and drivers needed to be written for specific models of drives. However, it didn't take long for the drive manufactures to defined and implement the Common Command Set, and generic drivers became easy to write.
Compare to both Linux and Windows. Both of these provided generic drivers for various classes of SCSI device, and so you did NOT need any special software to use a bare-bones drive.
It was fairly easy in a debugger to get around Apple's brand check. They had left debugging symbols in their HD setup program, so you could just set a breakpoint on "strcmp" and watch for one where it was comparing "APPLE" to "QUANTUM" (or whatever your drive was) and change the return value to say it matched. :-)