Re:What happens with deceased people's code?
on
Drafting GPL3
·
· Score: 1
What happens with deceased people's code?
AFAIK the rights get inherited just like anything else. And in principle, after some time the copyright expires and the code gets public domain (in reality, Dinsney & Co. will probably make sure that it doesn't). IANAL however.
Will I inherit granny's 200,000 lines of code?
Depends. What did granny write in her last will?
What do I do then?
Well, whatever you do with your own code. Except if you have (inherited) any contract concerning that code, of course.
Most emulators do not do a lot of FP so the Cells will probably not be used much for the lower end emulators.
Are you sure the Cells are only useful for FP? Since one of the main features of the Cells is that they can be pipelined, I could imagine an emulator using them in a similar way as real processors use their pipeline.
As a Linux user, I have never encountered something that I wish my shell (bash) could do that it couldn't.
Really? I can think of tons of things I'd like bash to do that it can't. For example, it could (reliably) give me the lotto numbers of next week. It could make the weather better. Or it could automatically refill my refrigerator. Not that I'd expect MSH to have those features:-)
Since when does freedom include the freedom to decide about your own punishment? "Oh, you want me to go to jail? Ok, let's say I agree to not leave this planet."
BSOD = Ball Shouldn't Occasionally Drop. I've never seen this meaning before, so it's new. Of course it has nothing to do with Windows (you didn't think it had, did you?);-)
but 3 is still a nice number I just hope that I can see it one day.
Here's a simple version of it: ||| The starred version of this is: *** In English texts you'll see it as: three In German texts you find it as: drei In French texts, look out for: trois In Italian texts you'll find: tre As roman numeral, it's: III. This is the decimal form: 3 In base 2, it's 11 However, computers will usually use 00000011 Or rather 0000000000000011 Actually, most systems today will use 00000000000000000000000000000011 And some even 00000000000000000000000000000000000000000000000000 00000000000011 Unless it's text, then they'll probably use 00110011 Which often is expressed as '\063' Or as '\x33'
In the x86, there is no such explicit alignment, so the CPU must at least CHECK for such an alignment issue on every instruction. That test is very likely to be an extra micro-opt.
Couldn't this distinction be done directly in hardware? Checking if the last 2 bits are zero is just a gate, and the output of that gate e.g. might just physically change the "micro-opt pointer" (I assume there's such a thing in the hardware?) so the unaligned access microcode would be skipped.
Maybe there's a fundamental flaw in my thinking, but to me it looks like being possible.
Now lets consider normal operation of the x86. It is not to load a byte into memory, but to perform a CISC "op [addr x2 + offset] [addr x2 + offset] [addr x2 + offset]".
Which byte-sized x86 operation takes three addresses?
Ok, it's a while since I last looked at the x86 instruction set, so I may miss some newer developments, but the only ones which come into my mind that take even two of them (implicitly) are the "string instructions" (movsb etc.). Most other instructions (like mov, add, sub) have only one memory access (you can do add eax, [ebx], but you cannot do e.g. add [ecx], [ebx]. BTW, all those operations are two-operand anyway. Did Intel add three-operand instructions later?
A large section of the community will be entirely closed-minded; if it's made of artificial parts, then it's not alive.
Well, a simulation of a brain on a normal computer will most certainly not be alive. But it might be conscious. On the other hand, a bacterium is certainly alive, but surely not conscious.
Being alive and being conscious are two entirely separate issues.
"If the human mind was simple enough to understand, we'd be too simple to understand it." -- Emerson Pugh
I think there's a very basic misconception at work: The idea that for X understanding Y, X must be more complex than Y. I think the idea is that you have to map Y in its full complexity into X.
But I'd argue, if you have to do that, it's clearly a sign of non-understanding. I'll explain this with two examples:
1. Imagine we'd have a machine which can automatically scan any physical object and build a computer program which exactly simulates the scanned object (and a computer powerful enough to do this simulation for any object this machine scans). Now imagine we would let this machine scan a brain, and put the program into the computer. Now, would we understand the brain? Or would the computer understand the brain? Well, I'd say neither.
2. A modern computer is a very complex machine. I doubt anyone can get the full complexity into his brain (how many million transistors are now typically on a CPU alone?), not to speak about all the connections between them (and the other pieces). Nevertheless, we do understand our computers. And that not just despite the fact that it's too complex to fit into our brain, but if we could just emulate the complete computer in our brain, I'd claim that we wouldn't understand it.
The point is, the understanding comes from abstraction. We know that several transistors and other components together form certain complex structures like logical gates, flip-flops or dynamic memory cells. Those are easy enough to understand in terms of transistors. And then, we go on from there, with greatly reduced complexity: We don't look at individual transistors any more, but we look at gates, flip-flops, memory cells etc. Those we group into more complex objects like adders, registers, etc. Again, we reduced complexity, since we basically only have to look at this level and can mostly forget about the gates. And this way we can go on until we have the coarse-grained description of the computer as a CPU, Memory, Disks etc. with a couple of busses connecting them. This splitting in hierarchies (and separate entities inside each hierarchic layer), or in short abstraction, is the core of understanding.
Since evolution favours more economic designs, it's IMHO quite likely that the human brain also allows grouping into such hierarchic levels (and actually we've already found some of those hierarchic levels: the neurons, the columns, and certain specialized areas in the brain). I'd expect a few more of those hierarchic levels, but the point is, if you understand the behavious of one hierarchic level, then you can describe the next level using the previous one as building blocks.
The factorial of 42 is a two-digit number? What base are you using? Base 37483411234209726053065806? (because in this base, it would be two-digits: The more significant digit would have the value 37483411234209726053065805, and the less significant digit would have the value 33187259034871818286636170)
In x86 land, you have byte-addressable instructions, but it's highly probable that these will be micro-opt'd into lesser efficient accesses.
Why? Given that the exact action is hidden, I don't see why a byte read couldn't be microcoded as a word read followed by a byte extraction, if that is really faster. No one actually sees if the cache is accessed byte-wise (and the main memory should only be accessed cache-line-wise anyway), nor if a hidden extra register is used internally.
Of course you could also have a full-screen browser showing slashdot on your other screen. The increase of productivity would be negative, but probably substantial.
So now the only real problem is keeping the human population down to a size the planet has enough raw meterials for.
Well, I'm sure someone will take care about that problem. After all, you always could rely on someone starting a war (especially when the population is high), and I'm sure they'll also find a way to produce bombs etc. with those robots. Or they just re-program the robots to use human source material for some products... actually, zombie robots would probably be far more dangerous than the typical zombie computer.
Slashdot Mirror
AFAIK the rights get inherited just like anything else. And in principle, after some time the copyright expires and the code gets public domain (in reality, Dinsney & Co. will probably make sure that it doesn't). IANAL however.
Depends. What did granny write in her last will?
Well, whatever you do with your own code. Except if you have (inherited) any contract concerning that code, of course.
No. The law addresses that.
Are you sure the Cells are only useful for FP? Since one of the main features of the Cells is that they can be pipelined, I could imagine an emulator using them in a similar way as real processors use their pipeline.
But Bochs is. And running ReactOS on top of it would give a completely free Windows replacement. If ReactOS were already ready, of course
Really? I can think of tons of things I'd like bash to do that it can't. For example, it could (reliably) give me the lotto numbers of next week. It could make the weather better. Or it could automatically refill my refrigerator.
Not that I'd expect MSH to have those features
Wouldn't that have to be renamed to "Lines"?
Publish the password on the web. Then whenever you've forgotten it, you just go to Google and search for "Password of professorhojo".
Make sure that you publish it on a high-ranking site. For example, put it on slashdot.
Since when does freedom include the freedom to decide about your own punishment? "Oh, you want me to go to jail? Ok, let's say I agree to not leave this planet."
Rejected: The obligatory phrase "using a computer" was not found.
But that wouldn't have been a whole new meaning!
BSOD = Ball Shouldn't Occasionally Drop. I've never seen this meaning before, so it's new. Of course it has nothing to do with Windows (you didn't think it had, did you?) ;-)
Well, Windows has no root, so they at best could take over the "Administrator of all Evil".
It seems not even to work if the MSDN side is in another window, but that window currently shows another tab.
Here's a simple version of it: |||
The starred version of this is: ***
In English texts you'll see it as: three
In German texts you find it as: drei
In French texts, look out for: trois
In Italian texts you'll find: tre
As roman numeral, it's: III.
This is the decimal form: 3
In base 2, it's 11
However, computers will usually use 00000011
Or rather 0000000000000011
Actually, most systems today will use 00000000000000000000000000000011
And some even 0000000000000000000000000000000000000000000000000
Unless it's text, then they'll probably use 00110011
Which often is expressed as '\063'
Or as '\x33'
Seen enough of it?
Couldn't this distinction be done directly in hardware? Checking if the last 2 bits are zero is just a gate, and the output of that gate e.g. might just physically change the "micro-opt pointer" (I assume there's such a thing in the hardware?) so the unaligned access microcode would be skipped.
Maybe there's a fundamental flaw in my thinking, but to me it looks like being possible.
Which byte-sized x86 operation takes three addresses?
Ok, it's a while since I last looked at the x86 instruction set, so I may miss some newer developments, but the only ones which come into my mind that take even two of them (implicitly) are the "string instructions" (movsb etc.). Most other instructions (like mov, add, sub) have only one memory access (you can do add eax, [ebx], but you cannot do e.g. add [ecx], [ebx]. BTW, all those operations are two-operand anyway. Did Intel add three-operand instructions later?
Over my head? No, over your head!
Depends. If there's a strong enough lobby of humans who can and are willing to enforce the machine's wishes, that's enough.
Well, a simulation of a brain on a normal computer will most certainly not be alive. But it might be conscious. On the other hand, a bacterium is certainly alive, but surely not conscious.
Being alive and being conscious are two entirely separate issues.
I think there's a very basic misconception at work: The idea that for X understanding Y, X must be more complex than Y. I think the idea is that you have to map Y in its full complexity into X.
But I'd argue, if you have to do that, it's clearly a sign of non-understanding. I'll explain this with two examples:
1. Imagine we'd have a machine which can automatically scan any physical object and build a computer program which exactly simulates the scanned object (and a computer powerful enough to do this simulation for any object this machine scans). Now imagine we would let this machine scan a brain, and put the program into the computer. Now, would we understand the brain? Or would the computer understand the brain? Well, I'd say neither.
2. A modern computer is a very complex machine. I doubt anyone can get the full complexity into his brain (how many million transistors are now typically on a CPU alone?), not to speak about all the connections between them (and the other pieces). Nevertheless, we do understand our computers. And that not just despite the fact that it's too complex to fit into our brain, but if we could just emulate the complete computer in our brain, I'd claim that we wouldn't understand it.
The point is, the understanding comes from abstraction. We know that several transistors and other components together form certain complex structures like logical gates, flip-flops or dynamic memory cells. Those are easy enough to understand in terms of transistors. And then, we go on from there, with greatly reduced complexity: We don't look at individual transistors any more, but we look at gates, flip-flops, memory cells etc. Those we group into more complex objects like adders, registers, etc. Again, we reduced complexity, since we basically only have to look at this level and can mostly forget about the gates. And this way we can go on until we have the coarse-grained description of the computer as a CPU, Memory, Disks etc. with a couple of busses connecting them. This splitting in hierarchies (and separate entities inside each hierarchic layer), or in short abstraction, is the core of understanding.
Since evolution favours more economic designs, it's IMHO quite likely that the human brain also allows grouping into such hierarchic levels (and actually we've already found some of those hierarchic levels: the neurons, the columns, and certain specialized areas in the brain). I'd expect a few more of those hierarchic levels, but the point is, if you understand the behavious of one hierarchic level, then you can describe the next level using the previous one as building blocks.
The factorial of 42 is a two-digit number?
What base are you using? Base 37483411234209726053065806?
(because in this base, it would be two-digits:
The more significant digit would have the value 37483411234209726053065805, and the less significant digit would have the value 33187259034871818286636170)
... that's a lot of work!
...Wulf cluster designed by the simulated brain.
...essorship
Sorry.
LFS even has been covered by slashdot before.
Why? Given that the exact action is hidden, I don't see why a byte read couldn't be microcoded as a word read followed by a byte extraction, if that is really faster. No one actually sees if the cache is accessed byte-wise (and the main memory should only be accessed cache-line-wise anyway), nor if a hidden extra register is used internally.
Of course you could also have a full-screen browser showing slashdot on your other screen. The increase of productivity would be negative, but probably substantial.
Well, I'm sure someone will take care about that problem. After all, you always could rely on someone starting a war (especially when the population is high), and I'm sure they'll also find a way to produce bombs etc. with those robots. Or they just re-program the robots to use human source material for some products