MS is trying to reduce the number of MP3 players that expose an easy to use file system interface rather than whatever DRM-of-the-day system is popular. MP3 players are becoming an _exceedingly_ low margin, high volume business, and MS is trying to make it more expensive to give people what they want ("just plug your device in and it works").
If it wasn't a big deal, there wouldn't be a quiet but forceful attempt to invalidate the patents going on.
Second Rennassance (which was utterly brilliant, incidentally) was fascinating. But it wasn't The Matrix; it was an entirely different movie, with different heros, different villains (we're the villains), and a vastly different universe.
They couldn't tell that story. They had the wrong cast. Animatrix was a brilliant mechanism for getting the story out there, so brilliant it probably should have preceeded one of the two movies proper. But this was before Kill Bill had made it OK to mix real life and animation (and it still probably wouldn't have worked -- we expect weird things from Quentin, like sword holsters on airplanes).
Revolutions didn't work without being attached to Reloaded. Both movies bled to death on their separation from one another. Saying either was the incarnation of a pre-planned, completed trilogy is just ludicrous.
I actually don't doubt that a trilogy was planned; quite a bit of backstory was hinted at with regards to a prequel (and a little for a sequel). The problem is, a prequel couldn't have Keanu, and Keanu sold the show.
So there might have been a trilogy, but I assure you: This wasn't it.
Nearly all sequels are written in response to the success of the first. Some are pre-written in more depth than others, but anything produced after a reaction to an original is altered by that reaction. Welcome to the creative process.
The first movie ends with Neo saying he was going to show people the world they lived in. The last movie ends with Neo dead and the status quo perfectly maintained. I posit that the former is an interesting future to imagine, and the latter is just stupid. "So, everyone remains enslaved, Zion lives happily ever after, and nobody inside figures out the nature of their virtual universe?"
Right.
My point with the spoon is -- where's the payoff? Why is it neat that there's a spoon, except to show off Jar Jar Matrix?
There was no payoff to Revolutions, and that's why an otherwise independantly OK movie is so very reviled.
Precisely. Harry Potter 2 was not as well received as the first, with the director pulled off the third movie even before the second came out. The funds for the third were guaranteed, but something big happened to alter where those funds went. The same actors might not be around for Harry Potter 7 (probably 8 or 9, given the growing girth of the books), but the series will not be ground into the tarmac.
By comparison, The Matrix trilogy is _dead_, in a way that eclipses the failures of Godfather 3, or to a lesser extent, Terminator 3. We've been seeing some billion dollar bets lately. This was a lost one.
As of late, there's been a resurgence in so-called "franchise" movies, where the funding for and expectation of a sequel is a foregone conclusion during the production process. Recently, we've seen three variants of this:
A) Lord Of The Rings, which filmed all three episodes in one monster shoot, then spent a year between each tweaking for maximum quality.
B) The Matrix Trilogy, which filmed the second and third episode in a less-monstrous shoot, and originally planned to unveil the conclusion a mere three months after the return. Tweaking was not originally planned for.
C) Harry Potter, which does not appear to begin production of the next chapter until the previous movie has finished its theatrical run.
Given these three case studies, it's worth noting that two of them (LOTR and HP) have their plotlines and characters fully fleshed out from day one, far in advance of movie production. Meanwhile, The Matrix sequels were written in response to the success of the original, meaning the third one got a screenplay before the second saw any public scrutiny.
I think this was the problem.
Unlike LOTR and HP, which had a healthy community of readers who could be tapped to determine which parts were most interesting and which parts could be sacrificed to the cutting room floor, the Wachowski's flew blind when concluding their series. They tried to show everything they could do, rather than explore the dimensions people were most interested in. When they realized their conclusion answered none of the new questions people couldn't help but ask -- they had no opportunity to recover their loss, save to push a worldwide release.
It's sad, too. Matrix Revolutions should have been a revolution inside the Matrix; the humans taking over their own virtual world, perhaps saving their own, perhaps abandoning it to the machines. Fundamentally, it should have been about the many within, not the grungy escapees. And so many interesting opportunities were abandoned...the spoon from the Matrix showing up in Zion, for instance. E
I don't know what happened. But I do know -- the serial format has brought some astonishing successes, and alot of money -- but when it fails, it seems to fail big.
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
The nCipher accelerates modular exponentiation -- stuff like 2^3000 mod 15001(or the remainder of 2 to the three thousandth power divided by 15001). This is inherently an integer op, and is only fast because you can break it down like one.
The computational pathways of GPU's are not appropriate for such calculations.
We've talked a decent amount about doing crypto on GPU's. The fundamental issue is that such processors are massively optimized for operating on floating point numbers, and almost all crypto is integer based -- lots of bitshifts, MODs, and XOR's, only the latter of which this gear handles correctly. Even if the problem with getting data back off the card was solved, the card itself couldn't do the job.
Indeed, I only know of one crypto hack that uses floats -- being from DJB, it's predictably brilliant. Basically, it's easy to compute the floating point error from a given operation, but computationally hard to find an operation that yields a given error. So you can effectively sign (or at least MAC) arbitrary content. Nice!
Alan's pulled off some gnarly stunts with the 2.2ac series; some definite hackworhty brilliance. Remember, hacker isn't a prejorative, any more than soldier is. I do understand the distinction you're trying to make, but it's pretty likely that Alan would call himself a kernel hacker.
I should point out they weren't talking about lines of code as much as public respect. Who's the lead coders of Samba? I know Tridge and Jeremy, but most don't. How bout Apache? KDE? Even if you can name them, they don't compare to Alan and Linus for not only coding so much themselves, but so successfully managing those hundreds of other kernel developers.
Yes, you need devtools to build a kernel. That doesn't make the kernel unimpressive.
It's very interesting what would have happened had Linux not shown up. I expect BSD would have forked into what Linux became. Had AT&T's lawsuit not slowed things down for years, Linux would not have happened, as BSD would have been far more mature at the appropriate time. How strange is that?
There are millions of users out there still running Windows 98. There may very well be more users of 98 than XP. Pulling 98 off MSDN means it will become significantly harder for developers to test their code on the platform that still large numbers of users have.
This is utterly disastrous. Companies that say there are risks of working with Linux should see what it means to work with Sun.
Can't you use high-power lasers to detect remote vibrations? Fire the beams at the edge of craters to maximize luminance delta, place them on a super-stable air platform (or just drop it in space), and "listen".
We do the same w/ IR lasers and glass windows, so it's not unimaginable (glass window = drum to our speech).
I'd like to take this moment to publically thank, congratulate, and otherwise embarass my illustrious former roommate, David Weekly. Alot of people talk about problems, like how alot of people talk about "Gee, how tragic is Zero Tolerance for all these kids, oh look, another one just got expelled for learning the word 'Knife'!"
Most people don't do anything. David Weekly did.
He stepped up, fought back against Diebold, and brought justice -- not just for himself (he's the founder of the California Community Colocation Project, so the ISP takedown notices directly affected him), not just for the four college kids attacked by Diebold, but for all of us here and for everybody with a stake in the perceived integrity of the American vote.
That's some damn fine work, David. Thanks! And thanks to everyone at EFF and OPG who fought this battle with him too!
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
Ah. But languages are living structures for communication....perhaps the meaning of the very word shifted to accomodate the horrors wrought by the Balrog.
If --gate can be suffixed onto any particularly scandsalous politiical event, and Dan Savage can exact ervenge on a certain senator of his choosing, I don't see it being difficult to imagine the word Moria evolving into Black Pit, after the fact.
Anonymous cash-based transactions are basically only acceptable for relatively small-value products and services.
Noticably, you can't get a hotel room at any major chain without a major credit card (Debit doesn't count).
Cash is basically the micropayment solution at the moment, because it's faster and cheaper then what the card vendors can provide. The larger the transactions get, though, the less acceptable the medium becomes.
You must declare any large amount of currency you're flying with, and all flights pose a confiscation risk for everything but your passport.
Like I said -- vaguely illegal. Makes the whole "new 20" thing kind of funny. I mean, I understand the reasons behind all the rules -- to say Russia has problems with nonpayment of taxes, not to mention "underground economies", is an understatement. And, yeah, war on drugs. Just worth noting the collateral damage.
Because anonymous financial transactions are a difficult and vaguely illegal proposition, while anonymous votes are a mission-critical top priority line item.
So you recompile once into the new instruction set, and now you're retrieving transmeta ops from main memory instead of x86 ops.
Except for really tight inner loops, you're always flying off to system RAM for one thing or another. While there's a static penalty because of code morphing, I'd wager it's a "lost in the backwash" effect -- oh, so a given stream of ops took a few extra million cycles to start cranking. BFD; we've got half a billion of 'em per second. The real question is why we don't have a full billion -- or more.
Code morphing might make Transmeta a little slower, but fundamentally the chip doesn't spend half its cycles translating code -- which, of course, is what an interpretative emulator would do. I think it's Amdahl's law that limits how much a given subprocess can be accelerated to the benefit of the master process -- code morphing could take zero time, and the chip still wouldn't be fast, because the damn chip can't be using the morpher all that much.
Intel and AMD are fast because they threw huge piles of cash at improving the silicon and making it fundamentally smarter. Transmeta's got some neat tricks going on -- but whatever Code Morphing translates to, I see no reason to believe it's fundamentally as powerful as what Intel and AMD can bring to the table.
That's what annoyed me about this article -- they go for the complex answer (it's slow because it morphs code) instead of the simple one (it's slow because it's a less advanced processor design, with fewer computational resources available being covered for by saying it drains less power). Given that this was an article about advanced processor designs, you'd think this might come up.
Since the author of this article is lurking here, I thought I'd ask:
You make a rather big deal about Transmeta needing to run all x86 code through a "code morpher" (dynamic recompiler, actually), and come up with a decently large set of conclusions based on it.
What's the big deal? No processor executes raw x86 anymore. Everything translates into an internal microcode that bears little resemblance to the original asm. Of course, normal chips have hardware accelerated microcode translaters, whereas Transmeta must recode in software -- but Transmeta's entire architecture was designed from day one to do that, and concievably they have more context available to do recoding by involving main memory in the process.
And what is it with you neglecting the equivalence of main memory? Yes, transistors are necessary to store the translated program. They're also necessary to store the original one -- the Mozilla client I'm presently tapping away inside sure as hell doesn't fit in L1 on my C3! Outside of a small static penalty on load, and a smaller dynamic penalty from ongoing profiling, you can't blame performance on the fact that software needs to be in RAM. Software always needs to be in RAM.
Don't get me wrong -- Transmeta's a performance dog, and everyone's known that since day one. But I think it's reasonable to say the cause is mostly one of attention -- every man hour they threw into allowing the system to emulate x86 took away from adding pipelines, increasing clock rates, tweaking caches, etc. In other words, yes it's a feat that they got the code to work, but you don't need to blame the feat for the quality of work -- they simply did alot of work nobody else had to waste time on, and fell behind because of it.
Much easier explanation. Might even be true.
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
Slashdot Mirror
MS is trying to reduce the number of MP3 players that expose an easy to use file system interface rather than whatever DRM-of-the-day system is popular. MP3 players are becoming an _exceedingly_ low margin, high volume business, and MS is trying to make it more expensive to give people what they want ("just plug your device in and it works").
If it wasn't a big deal, there wouldn't be a quiet but forceful attempt to invalidate the patents going on.
Your paragraphs contradict one another.
Second Rennassance (which was utterly brilliant, incidentally) was fascinating. But it wasn't The Matrix; it was an entirely different movie, with different heros, different villains (we're the villains), and a vastly different universe.
They couldn't tell that story. They had the wrong cast. Animatrix was a brilliant mechanism for getting the story out there, so brilliant it probably should have preceeded one of the two movies proper. But this was before Kill Bill had made it OK to mix real life and animation (and it still probably wouldn't have worked -- we expect weird things from Quentin, like sword holsters on airplanes).
Revolutions didn't work without being attached to Reloaded. Both movies bled to death on their separation from one another. Saying either was the incarnation of a pre-planned, completed trilogy is just ludicrous.
--Dan
I actually don't doubt that a trilogy was planned; quite a bit of backstory was hinted at with regards to a prequel (and a little for a sequel). The problem is, a prequel couldn't have Keanu, and Keanu sold the show.
So there might have been a trilogy, but I assure you: This wasn't it.
--Dan
Nearly all sequels are written in response to the success of the first. Some are pre-written in more depth than others, but anything produced after a reaction to an original is altered by that reaction. Welcome to the creative process.
The first movie ends with Neo saying he was going to show people the world they lived in. The last movie ends with Neo dead and the status quo perfectly maintained. I posit that the former is an interesting future to imagine, and the latter is just stupid. "So, everyone remains enslaved, Zion lives happily ever after, and nobody inside figures out the nature of their virtual universe?"
Right.
My point with the spoon is -- where's the payoff? Why is it neat that there's a spoon, except to show off Jar Jar Matrix?
There was no payoff to Revolutions, and that's why an otherwise independantly OK movie is so very reviled.
--Dan
Precisely. Harry Potter 2 was not as well received as the first, with the director pulled off the third movie even before the second came out. The funds for the third were guaranteed, but something big happened to alter where those funds went. The same actors might not be around for Harry Potter 7 (probably 8 or 9, given the growing girth of the books), but the series will not be ground into the tarmac.
By comparison, The Matrix trilogy is _dead_, in a way that eclipses the failures of Godfather 3, or to a lesser extent, Terminator 3. We've been seeing some billion dollar bets lately. This was a lost one.
--Dan
So it's kind of interesting.
As of late, there's been a resurgence in so-called "franchise" movies, where the funding for and expectation of a sequel is a foregone conclusion during the production process. Recently, we've seen three variants of this:
A) Lord Of The Rings, which filmed all three episodes in one monster shoot, then spent a year between each tweaking for maximum quality.
B) The Matrix Trilogy, which filmed the second and third episode in a less-monstrous shoot, and originally planned to unveil the conclusion a mere three months after the return. Tweaking was not originally planned for.
C) Harry Potter, which does not appear to begin production of the next chapter until the previous movie has finished its theatrical run.
Given these three case studies, it's worth noting that two of them (LOTR and HP) have their plotlines and characters fully fleshed out from day one, far in advance of movie production. Meanwhile, The Matrix sequels were written in response to the success of the original, meaning the third one got a screenplay before the second saw any public scrutiny.
I think this was the problem.
Unlike LOTR and HP, which had a healthy community of readers who could be tapped to determine which parts were most interesting and which parts could be sacrificed to the cutting room floor, the Wachowski's flew blind when concluding their series. They tried to show everything they could do, rather than explore the dimensions people were most interested in. When they realized their conclusion answered none of the new questions people couldn't help but ask -- they had no opportunity to recover their loss, save to push a worldwide release.
It's sad, too. Matrix Revolutions should have been a revolution inside the Matrix; the humans taking over their own virtual world, perhaps saving their own, perhaps abandoning it to the machines. Fundamentally, it should have been about the many within, not the grungy escapees. And so many interesting opportunities were abandoned...the spoon from the Matrix showing up in Zion, for instance. E
I don't know what happened. But I do know -- the serial format has brought some astonishing successes, and alot of money -- but when it fails, it seems to fail big.
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
The nCipher accelerates modular exponentiation -- stuff like 2^3000 mod 15001(or the remainder of 2 to the three thousandth power divided by 15001). This is inherently an integer op, and is only fast because you can break it down like one.
The computational pathways of GPU's are not appropriate for such calculations.
--Dan
We've talked a decent amount about doing crypto on GPU's. The fundamental issue is that such processors are massively optimized for operating on floating point numbers, and almost all crypto is integer based -- lots of bitshifts, MODs, and XOR's, only the latter of which this gear handles correctly. Even if the problem with getting data back off the card was solved, the card itself couldn't do the job.
Indeed, I only know of one crypto hack that uses floats -- being from DJB, it's predictably brilliant. Basically, it's easy to compute the floating point error from a given operation, but computationally hard to find an operation that yields a given error. So you can effectively sign (or at least MAC) arbitrary content. Nice!
--Dan
Tom--
Alan's pulled off some gnarly stunts with the 2.2ac series; some definite hackworhty brilliance. Remember, hacker isn't a prejorative, any more than soldier is. I do understand the distinction you're trying to make, but it's pretty likely that Alan would call himself a kernel hacker.
I should point out they weren't talking about lines of code as much as public respect. Who's the lead coders of Samba? I know Tridge and Jeremy, but most don't. How bout Apache? KDE? Even if you can name them, they don't compare to Alan and Linus for not only coding so much themselves, but so successfully managing those hundreds of other kernel developers.
Yes, you need devtools to build a kernel. That doesn't make the kernel unimpressive.
It's very interesting what would have happened had Linux not shown up. I expect BSD would have forked into what Linux became. Had AT&T's lawsuit not slowed things down for years, Linux would not have happened, as BSD would have been far more mature at the appropriate time. How strange is that?
--Dan
I had a Touchdrive. Nice idea, implemented ok.
Terrible accuracy. Lost it, didn't even care to look for it.
Fun to show off, though.
--Dan
Kermit and Kermit 95 have been invaluable tools to improve our computing efficiency
holy crap someone used these two words in the same sentence world coming to end STOP
--dan
There are millions of users out there still running Windows 98. There may very well be more users of 98 than XP. Pulling 98 off MSDN means it will become significantly harder for developers to test their code on the platform that still large numbers of users have.
This is utterly disastrous. Companies that say there are risks of working with Linux should see what it means to work with Sun.
--Dan
Sockets are overrated.
Can't you use high-power lasers to detect remote vibrations? Fire the beams at the edge of craters to maximize luminance delta, place them on a super-stable air platform (or just drop it in space), and "listen".
We do the same w/ IR lasers and glass windows, so it's not unimaginable (glass window = drum to our speech).
--Dan
First, you can't use Smalley fingers, because they're too sticky: Whatever you're manipulating will attach to the finger due to proximity effects.
Then, you can't use positional assembly, because objects don't stick to eachother just because they get close to one another.
So, uh, which is it? Do or do not compatible molecules bind or adhere when brought in close proximity?
I'm not a chemical engineer (obviously), but even I picked up on this interesting contradiction.
--Dan
No offense, but what idiot thought to use 2.4ghz inside the body?
Yes, of course it barely transmits, 2.4ghz is the frequency used by microwaves to heat food, because water absorbs it so well.
Hint: We're mostly water too.
Now if we could just dessicate people utterly, those transmitters would work just fine...wouldn't be much of a life monitor, though...
--Dan
I'd like to take this moment to publically thank, congratulate, and otherwise embarass my illustrious former roommate, David Weekly. Alot of people talk about problems, like how alot of people talk about "Gee, how tragic is Zero Tolerance for all these kids, oh look, another one just got expelled for learning the word 'Knife'!"
Most people don't do anything. David Weekly did.
He stepped up, fought back against Diebold, and brought justice -- not just for himself (he's the founder of the California Community Colocation Project, so the ISP takedown notices directly affected him), not just for the four college kids attacked by Diebold, but for all of us here and for everybody with a stake in the perceived integrity of the American vote.
That's some damn fine work, David. Thanks! And thanks to everyone at EFF and OPG who fought this battle with him too!
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
Maybe there was always a black pit there.
In other words, perhaps the pit preceeds Balrog. Perhaps the dwarves became so rich because of resources that could be mined from the pit.
--Dan
Ah. But languages are living structures for communication....perhaps the meaning of the very word shifted to accomodate the horrors wrought by the Balrog.
If --gate can be suffixed onto any particularly scandsalous politiical event, and Dan Savage can exact ervenge on a certain senator of his choosing, I don't see it being difficult to imagine the word Moria evolving into Black Pit, after the fact.
--Dan
Anonymous cash-based transactions are basically only acceptable for relatively small-value products and services.
Noticably, you can't get a hotel room at any major chain without a major credit card (Debit doesn't count).
Cash is basically the micropayment solution at the moment, because it's faster and cheaper then what the card vendors can provide. The larger the transactions get, though, the less acceptable the medium becomes.
You must declare any large amount of currency you're flying with, and all flights pose a confiscation risk for everything but your passport.
Like I said -- vaguely illegal. Makes the whole "new 20" thing kind of funny. I mean, I understand the reasons behind all the rules -- to say Russia has problems with nonpayment of taxes, not to mention "underground economies", is an understatement. And, yeah, war on drugs. Just worth noting the collateral damage.
--Dan
Because anonymous financial transactions are a difficult and vaguely illegal proposition, while anonymous votes are a mission-critical top priority line item.
--Dan
Everyone thinks of this.
Everyone thinks they're the first to think of this.
Everyone's got the right idea -- except Diebold.
--Dan
So you recompile once into the new instruction set, and now you're retrieving transmeta ops from main memory instead of x86 ops.
Except for really tight inner loops, you're always flying off to system RAM for one thing or another. While there's a static penalty because of code morphing, I'd wager it's a "lost in the backwash" effect -- oh, so a given stream of ops took a few extra million cycles to start cranking. BFD; we've got half a billion of 'em per second. The real question is why we don't have a full billion -- or more.
Code morphing might make Transmeta a little slower, but fundamentally the chip doesn't spend half its cycles translating code -- which, of course, is what an interpretative emulator would do. I think it's Amdahl's law that limits how much a given subprocess can be accelerated to the benefit of the master process -- code morphing could take zero time, and the chip still wouldn't be fast, because the damn chip can't be using the morpher all that much.
Intel and AMD are fast because they threw huge piles of cash at improving the silicon and making it fundamentally smarter. Transmeta's got some neat tricks going on -- but whatever Code Morphing translates to, I see no reason to believe it's fundamentally as powerful as what Intel and AMD can bring to the table.
That's what annoyed me about this article -- they go for the complex answer (it's slow because it morphs code) instead of the simple one (it's slow because it's a less advanced processor design, with fewer computational resources available being covered for by saying it drains less power). Given that this was an article about advanced processor designs, you'd think this might come up.
--Dan
Since the author of this article is lurking here, I thought I'd ask:
You make a rather big deal about Transmeta needing to run all x86 code through a "code morpher" (dynamic recompiler, actually), and come up with a decently large set of conclusions based on it.
What's the big deal? No processor executes raw x86 anymore. Everything translates into an internal microcode that bears little resemblance to the original asm. Of course, normal chips have hardware accelerated microcode translaters, whereas Transmeta must recode in software -- but Transmeta's entire architecture was designed from day one to do that, and concievably they have more context available to do recoding by involving main memory in the process.
And what is it with you neglecting the equivalence of main memory? Yes, transistors are necessary to store the translated program. They're also necessary to store the original one -- the Mozilla client I'm presently tapping away inside sure as hell doesn't fit in L1 on my C3! Outside of a small static penalty on load, and a smaller dynamic penalty from ongoing profiling, you can't blame performance on the fact that software needs to be in RAM. Software always needs to be in RAM.
Don't get me wrong -- Transmeta's a performance dog, and everyone's known that since day one. But I think it's reasonable to say the cause is mostly one of attention -- every man hour they threw into allowing the system to emulate x86 took away from adding pipelines, increasing clock rates, tweaking caches, etc. In other words, yes it's a feat that they got the code to work, but you don't need to blame the feat for the quality of work -- they simply did alot of work nobody else had to waste time on, and fell behind because of it.
Much easier explanation. Might even be true.
Yours Truly,
Dan Kaminsky
DoxPara Research
http://www.doxpara.com
The change rate of the actual information does not scale with the increased resolution.
Think how much more a 32x32 window changes than a 320x320 one.
--Dan