Image compositing? I can think of plenty of applications on Windows that do that...in fact, here's one that only does image compositing just for you...and it's free... http://www.topshareware.com/Image-Inc.-download-44 355.htm
Er, yeah, it's kind of like that, except it does movies, CG effects, and is actually used by professionals in the field it was written for. You didn't read GP's link, did you?
Anyway, I use Mac regularly, and I think that Quicksilver and Adium are the killer apps. iChat AV is pretty nice too if you want to video chat (granted, it sucks for AIM but that's what Adium is for), and the UNIX subsystem is nice for running POSIX-based OSS (I'm a CS student, so I do a lot of work with OSS). I also think that the interface is, on average, more polished than Linux or Windows. Finder looks terrible with brushed metal, but hey, why use Finder when you can use Quicksilver.
I'm not sure it's worth it, though. Those apps are really nice, and Apple's support is really good. On the other hand, their hardware is mediocre (one button, heavy for its size, runs hot and whiny, mediocre battery life) and slightly more expensive than the competition (at least in what I was looking for). For my next laptop, I might just get a Dell/Toshiba/HP/Sony/Lenovo/whatever and slap Linux on it.
Since this is localhost-only, and (unlike Bernstein's and Boneh's attacks) can't be extended remote, it's not going to impact SSL or (single-user) SSH.
What's more, they didn't break OpenSSL even on the same machine. To quote from the paper:
We used the RSA implementation from OpenSSL version 0.9.7e as a template and made some modifications to convert this implementation into the simple one as stated above. To be more precise, we changed the window size from 5 to 1, removed the CRT mode, and added the dummy reduction step. (emphasis mine)
If the window size is 5 by default, most of the bits don't affect branches at all, because the branch is replaced by a lookup table. If it's a sliding window, something like 1 bit in 5 results in a key-dependent branch, and if it's a fixed window, none of them do. Oh, and as a side effect, the signing/decryption will be significantly (~30%) faster. Of course, you can use cache-timing attacks to snoop on the lookup table.
But now to quote from the OpenSSL changelog:
Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
...
*) Make a new fixed-window mod_exp implementation the default for
RSA, DSA, and DH private-key operations so that the sequence of
squares and multiplies and the memory access pattern are
independent of the particular secret key. This will mitigate
cache-timing and potential related attacks.
BN_mod_exp_mont_consttime() is the new exponentiation implementation,
and this is automatically used by BN_mod_exp_mont() if the new flag
BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
will use this BN flag for private exponents unless the flag
RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
(emphasis mine)
Now, I haven't gone over the latest OpenSSL code with a fine-toothed comb, but I have gone over it with a coarse-toothed one (as part of an analysis of a related attack). The BN_mod_exp_mont_consttime() in 0.9.8d appears (both from code structure and comments) to go to some lengths to avoid key-dependent branches, and I expect that the same is true in the older maintained branch, 0.9.7l. If this is done correctly, it should totally prevent this attack.
In other words, this attack doesn't work on any real implementation, because existing optimizations hide most of the bits. As to the remaining bits, OpenSSL fixed this over a year ago as part of a patch to a different localhost-only attack (the "Hyperthreading considered harmful" attack). To make their attack, these guys used a year-old version of the code, and even then they had to rewrite it to be more vulnerable.
Of course, their paper has some value: it shows once again that if you use a naive implementation of RSA, you can be hacked. It improves a previous attack, and demonstrates that one obvious defense doesn't prevent the attack. But the real world, at least the UNIX side, has moved on from there.
I have to mostly agree with this: Bussard's talking tactics were pretty sleazy. His distinctions between "physics" and "engineering" problems were largely vacuous, and he glossed over a lot of stuff. He sounded a lot like a crank in several places, not least when he threatened (repeatedly) to give the tech to China. Also, his spiel on what this machine would do if it worked is unnecessary: we all know that a high-efficiency fusion machine would change the world, but we need to be convinced that he can build one.
He also suggested that the panel to decide whether this is workable should consist mainly of his 70+year-old friends, which is pretty shady.
However, you're casting it in slightly too negative a light. It's reasonable to believe that a lot of the trouble toward the end was due to the lack of funding, and that funding is actually unattainable for bureaucratic reasons. Furthermore, the first step of his research would be to rebuild the original machine with better coils, and take very, very careful measurements of the thing. This was indended to take a year and cost a few million dollars, which sounds entirely reasonable. His point about building small models is that his equations show that only a large machine can be at all efficient, so building a quarter-size or half-size machine would prove nothing about the engineering side. Once he's proved the physics, he wants to move directly to a full-scale demo.
There is even some small amount of merit in his distinction between physics and engineering problems. You obviously need a huge power supply to run this thing. We know how to build such power supplies, but they cost money and he doesn't have money, so he's running it from a capacitor bank for half a millisecond at a time. We know how to build fast-response gas injectors, but he doesn't have those either, because they also cost money, so he's using slow ones. We know how to build megavolt standoffs, but they cost money... On the other hand, his spiel on how easy helium extraction is may be entirely bullshit. He claims they have a paper on it, so they've thought about it and it's probably not complete bullshit, but it's not a standard affair in the engineering community. I also don't believe him that arcing is an engineering problem.
He's also clearly not lying, at least not about whether the machine is possible, because we all know he'll get caught in the first stage and that's not the kind of legacy he's trying to leave. However, it's quite possible that he's hallucinating the data, or reading too much into it, or something, and he's clearly got a serious case of tunnel vision.
Personally, I'd like to see notebook makers focus on getting the kind of battery life you see in the Nintendo DS (but using a larger battery, of course). Even if speeds are knocked in half, that's okay.
Lenovo's "ultraportable" X60s series gets 8 hours of battery life if you get it with the bigger battery option. It's using a Core Duo, dunno if it'll get Core 2 soon or not (it's more a business notebook, so not much holiday upgrade pressure).
If Lenovo is too expensive for you, or if the X60s is too barebones (it doesn't have an optical drive, for instance), there are some Taiwanese notebooks with VIA C7-M chips. The C7-M is anemic (except for crypto, where it has an accelerator), but the 1.2GHz version draws 0.1W at idle, scaling linearly up to 7W at full load. The scaling supposedly handles bursty loads better than Intel processors, too.
In any case, remember that the CPU is only a ~30% slice of the notebook power consumption pie. The biggest slice is the display with its backlight. Wireless, HD, chipset, RAM and video card (if you have one) are significant as well.
Either way, he signed the fucking thing, even though he has actively did his special magical presidential veto more times than all other presidents in history.
Actually, Bush has only vetoed one bill, on stem cell research.
I hope so. I hope you're the only one so hypersensative to the word that like==blond. Save your pedantry for where it's really warranted, like misuse of apostrophies!
Oh no! We won't be able to replace $TIME_HONORED_CLASSIC with $MODERN_TRIPE?
Are you suggesting that the author of every great work died before 1936?
What about Joyce? CS Lewis? HG Wells? JRR Tolkien? Cather? Adams? Orwell? TS Eliot? Woolf?
Some of their works are already classics, but there are surely dozens of other works, by these and other authors, that won't become classics because it is difficult to distribute them.
I would not expect a lot of minorities in any of the ivy league schools. They are by and large affirmitive action for rich people, movie stars, models and such. The main criterea for entrance is fame and fortune...
That may have been true 20 years ago, but it isn't true anymore.
and the main crieterea for graduation is "went to some ofo my classes and didn't stay high the entire time"...
Bold words from someone who can't spell "criterion". The pressure at MIT at least is pretty insane.
If we'd put the music of the Voyager Golden Record on a USB key in iTunes Fairplay ALC, I have little doubt that aliens with the capacity to recover a space probe from interstellar space could have decoded it (they have to be at least as smart as DVD Jon). Hell, it might have made it interesting for them.
Uh uh. We could probably retrieve an interstellar probe if it wandered through our solar system with its radio on. But without a program available to play a Fairplay-encoded AAC, it would surely look like noise to us, and probably to the aliens as well.
Even the great DVD Jon can't crack Fairplay without the license file.
Everything runs fine with jpg, java, javascript, and flash.
Java and Javascript are not image formats. Flash is much broader, is a non-accessible resource hog, and is most commonly used for irritating ads (not unlike animated GIFs, I suppose).
That leaves JPEG, which is actually an image format, but a totally different one. GIF was designed, for logos: it is lossless, has a very limited color palette, and allows for some amount of transparency. JPEG was designed for photos: it's lossy, has a broad color palette, no transparency, and it looks terrible on things with crisp lines, like text or diagrams.
The real competitor to GIF is PNG, which is still lossless, but has better transparency and more colors. Unfortunately, it also has poorly-specified gamma correction, which makes it painful to use in web design.
You're right: string theory is not a scientific theory. It's a mathematical theory. That is, it's a collection of mathematical axioms and related proofs (and an extremely unpleasant one, according to a physicist friend of mine).
String theory does provide a model of physics. That is to say, if you set the parameters right, you get something that looks kind of like quantum field theory (which, by the way, is also a mathematical theory in addition to a scientific one). Unfortunately, the math is too hard to deteremine how they differ, and even once a determination is made, string theory has a lot of parameters which will have to be set before real predections are possible. Note that quantum field theories are testable, but only barely. For instance, Howard Georgi's "representations of SU(5)" theory was disproved by experiments in proton decay.
Finally, once string theory does make real predictions, they will be hard to test. In particular, they are likely to require enormous amounts of energy, and accelerator experiments can take years to run and analyze. So it will be a long time yet before string theory becomes scientific.
Go to any elite engineering school and take a survey of the top 10% of the students there. I would be shocked if at least 50% of those students are not chinese. I don't mean chinese americans, I mean chinese from china.
I call bullshit (on your statistic, not your general point).
I'm in the Stanford CS PhD program, and there are surprisingly few Chinese. Sure, there are plenty of Asians, but a reasonable fraction of them are Korean or Indian, and there are plenty of Caucasians (disproportionately many Greeks), Hispanics and Jews. I've only noticed a few Chinese from China, but I'm oblivious that way. I have no idea who the top students are, but if you count the whole PhD program, it's certainly not true.
At Harvard it wasn't true either (of course, Harvard is not an engineering school). If anything, there were more Jews than Chinese at the top, and almost all of the Chinese were 2nd or 3rd generation; this was true both undergrad and CS grad (though not as much of math grads).
At MIT, from what I know, it wasn't true either: most of the elite students I knew there were either Jewish or Caucasian. I didn't know that many, though.
At my summer jobs, there have been a lot of smart Chinese people, but only a few of them were from China.
At the math olympiad programs the years I went, half the team was Chinese, but only one or so each year from China. Of the students there in general, maybe a third were Chinese. (China wins the olympiad almost every year, though.)
This is not to say that Chinese people are stupid. Just like you, I know plenty of really smart Chinese people, although most of them are 2nd or 3rd generation.
On Windows it is legal to compare two strings like this:
if(str1 == str2)
Actually... no, it isn't. I was doing Java development this summer (on Java 1.5), and I did that a few times by accident, and it doesn't work. It asks whether str1 and str2 are the same String object, rather than asking whether the strings contained in them are equal.
You're trolling, but I'm bored, so I'll respond anyway.
OO code was developed by people who think in calculus and set theory.
Not so. The people who think in calculus and set theory developed Haskell. And if you were arguing against Haskell, you might have had a point.
Perl is good, Ruby is good, these are things that alleviate the problem a lot.
Perl was designed for writing short, simple programs which do lots of string manipulation. For that it's often the best thing out there. For anything large and complicated, it often becomes difficult to maintain.
Ruby is as object-oriented as they come, and very clean (but slow) as a result. In fact, it's directly based on the grand-daddy of object-oriented programming, Smalltalk.
Um, let me make a simpler argument, the same one I'm responding to, and it's a damned good one. Look at the VERBIAGE on those "modern" programs. What the HELL are you people thinking? Just typing all that slows down development enormously.
That would, in fact, be Java. Also Objective-C. Occasionally C++, but not so much. Pretty much everything else out there has shorter lines.
I guess the next thing to do (if I'm so clever) would be to develop a simple, usable language for non-CS students. I kinda think someone did it already, many times. But no students will ever see it.
Hah, like AppleScript? Really, I think a better thing to do would be to teach a Smalltalk-derived OO language like Python or Ruby.
Niagara is actually very nice if you have a massively parallel (~ 32 threads always on the CPU) task with no floating point operations. So it shines for enormous webservers, or servers combining every tier on one machine. Unfortunately, there's only one FPU on the whole chip, so floating point pretty much kills the performance.
Also, for anything running single-threaded, the performance is terrible. The processor runs at 1-1.2 GHz and isn't superscalar.
The combination of these factors means that not too many people will be buying SunFire T2000 servers at $13k apiece.
Enjoy paying a hundred bucks a year for features like this as an upgrade?
I'm a student, so it's only $80. And no, as I said I'm happy to use VirtueDesktops. But I enjoy paying $80 a year for upgrades like iChat AV (with H.264), Core Image/Vector/Data/Audio, Expose (which is actually quite nice), launchd, fast user switching (IIRC, there's still not a good interface for that in Ubuntu?), and so on. As it happens, I've only done this once (10.2 -> 10.3; was using Ubuntu when 10.4 came out), but I'll do it again in a heartbeat if 10.5 has virtualization or ZFS support. Without those, we'll see. DTrace is awesome, and the new iChat stuff looks pretty cool, and Time machine should be useful (if corny), but I'm not sure if that will be enough.
Leopard is going to have virtual desktops. In the mean time, there's Desktop Manager or VirtueDesktops for free, and CodeTek Virtual Desktops for not-so-free. The first two work pretty well; I've never used CodeTek before.
As an EE and and audio engineer and a musician, I have a simple reply to this: There are no "insignificant" bits. Only bits that are lost due to one factor or another. Lossy compression isn't something to figure out. It is something to abandon.
The point still stands. Some bits are more significant than others, and as an engineer you must know this. Start with a very high-bandwidth encoding (96k x 6 x 24, or 6-channel high-frequency sigma-delta, or whatever). It's not practical to haul it around, so you have to compress it to something reasonable, say 500kb/s. Since all compression down to a given bitrate "loses" the same amount of information (assuming that it is optimal, and so cannot be compressed further). It's equally lossy.
What's to say that the optimal encoding (in whatever psychoacoustic sense) is a function of the 44k x 2 x 8 CD-quality version?
Slashdot Mirror
... maybe they should call it 3DNow or something?
Image compositing? I can think of plenty of applications on Windows that do that...in fact, here's one that only does image compositing just for you...and it's free...4 355.htm
http://www.topshareware.com/Image-Inc.-download-4
Er, yeah, it's kind of like that, except it does movies, CG effects, and is actually used by professionals in the field it was written for. You didn't read GP's link, did you?
Anyway, I use Mac regularly, and I think that Quicksilver and Adium are the killer apps. iChat AV is pretty nice too if you want to video chat (granted, it sucks for AIM but that's what Adium is for), and the UNIX subsystem is nice for running POSIX-based OSS (I'm a CS student, so I do a lot of work with OSS). I also think that the interface is, on average, more polished than Linux or Windows. Finder looks terrible with brushed metal, but hey, why use Finder when you can use Quicksilver.
I'm not sure it's worth it, though. Those apps are really nice, and Apple's support is really good. On the other hand, their hardware is mediocre (one button, heavy for its size, runs hot and whiny, mediocre battery life) and slightly more expensive than the competition (at least in what I was looking for). For my next laptop, I might just get a Dell/Toshiba/HP/Sony/Lenovo/whatever and slap Linux on it.
Since this is localhost-only, and (unlike Bernstein's and Boneh's attacks) can't be extended remote, it's not going to impact SSL or (single-user) SSH.
What's more, they didn't break OpenSSL even on the same machine. To quote from the paper:
If the window size is 5 by default, most of the bits don't affect branches at all, because the branch is replaced by a lookup table. If it's a sliding window, something like 1 bit in 5 results in a key-dependent branch, and if it's a fixed window, none of them do. Oh, and as a side effect, the signing/decryption will be significantly (~30%) faster. Of course, you can use cache-timing attacks to snoop on the lookup table.
But now to quote from the OpenSSL changelog:
Now, I haven't gone over the latest OpenSSL code with a fine-toothed comb, but I have gone over it with a coarse-toothed one (as part of an analysis of a related attack). The BN_mod_exp_mont_consttime() in 0.9.8d appears (both from code structure and comments) to go to some lengths to avoid key-dependent branches, and I expect that the same is true in the older maintained branch, 0.9.7l. If this is done correctly, it should totally prevent this attack.
In other words, this attack doesn't work on any real implementation, because existing optimizations hide most of the bits. As to the remaining bits, OpenSSL fixed this over a year ago as part of a patch to a different localhost-only attack (the "Hyperthreading considered harmful" attack). To make their attack, these guys used a year-old version of the code, and even then they had to rewrite it to be more vulnerable.
Of course, their paper has some value: it shows once again that if you use a naive implementation of RSA, you can be hacked. It improves a previous attack, and demonstrates that one obvious defense doesn't prevent the attack. But the real world, at least the UNIX side, has moved on from there.
I have to mostly agree with this: Bussard's talking tactics were pretty sleazy. His distinctions between "physics" and "engineering" problems were largely vacuous, and he glossed over a lot of stuff. He sounded a lot like a crank in several places, not least when he threatened (repeatedly) to give the tech to China. Also, his spiel on what this machine would do if it worked is unnecessary: we all know that a high-efficiency fusion machine would change the world, but we need to be convinced that he can build one.
He also suggested that the panel to decide whether this is workable should consist mainly of his 70+year-old friends, which is pretty shady.
However, you're casting it in slightly too negative a light. It's reasonable to believe that a lot of the trouble toward the end was due to the lack of funding, and that funding is actually unattainable for bureaucratic reasons. Furthermore, the first step of his research would be to rebuild the original machine with better coils, and take very, very careful measurements of the thing. This was indended to take a year and cost a few million dollars, which sounds entirely reasonable. His point about building small models is that his equations show that only a large machine can be at all efficient, so building a quarter-size or half-size machine would prove nothing about the engineering side. Once he's proved the physics, he wants to move directly to a full-scale demo.
There is even some small amount of merit in his distinction between physics and engineering problems. You obviously need a huge power supply to run this thing. We know how to build such power supplies, but they cost money and he doesn't have money, so he's running it from a capacitor bank for half a millisecond at a time. We know how to build fast-response gas injectors, but he doesn't have those either, because they also cost money, so he's using slow ones. We know how to build megavolt standoffs, but they cost money... On the other hand, his spiel on how easy helium extraction is may be entirely bullshit. He claims they have a paper on it, so they've thought about it and it's probably not complete bullshit, but it's not a standard affair in the engineering community. I also don't believe him that arcing is an engineering problem.
He's also clearly not lying, at least not about whether the machine is possible, because we all know he'll get caught in the first stage and that's not the kind of legacy he's trying to leave. However, it's quite possible that he's hallucinating the data, or reading too much into it, or something, and he's clearly got a serious case of tunnel vision.
Personally, I'd like to see notebook makers focus on getting the kind of battery life you see in the Nintendo DS (but using a larger battery, of course). Even if speeds are knocked in half, that's okay.
Lenovo's "ultraportable" X60s series gets 8 hours of battery life if you get it with the bigger battery option. It's using a Core Duo, dunno if it'll get Core 2 soon or not (it's more a business notebook, so not much holiday upgrade pressure).
If Lenovo is too expensive for you, or if the X60s is too barebones (it doesn't have an optical drive, for instance), there are some Taiwanese notebooks with VIA C7-M chips. The C7-M is anemic (except for crypto, where it has an accelerator), but the 1.2GHz version draws 0.1W at idle, scaling linearly up to 7W at full load. The scaling supposedly handles bursty loads better than Intel processors, too.
In any case, remember that the CPU is only a ~30% slice of the notebook power consumption pie. The biggest slice is the display with its backlight. Wireless, HD, chipset, RAM and video card (if you have one) are significant as well.
You really should upgrade your system to get rid of those ^H's. ^W is the new tech...
Either way, he signed the fucking thing, even though he has actively did his special magical presidential veto more times than all other presidents in history.
Actually, Bush has only vetoed one bill, on stem cell research.
No, because then it would be an invisible rootkit underlord.
... by a Swarm of Green Robots.
Open sores peach wreck ignition is final ready.
I hope so. I hope you're the only one so hypersensative to the word that like==blond. Save your pedantry for where it's really warranted, like misuse of apostrophies!
You misspelled "apostrophes."
Oh no! We won't be able to replace $TIME_HONORED_CLASSIC with $MODERN_TRIPE?
Are you suggesting that the author of every great work died before 1936?
What about Joyce? CS Lewis? HG Wells? JRR Tolkien? Cather? Adams? Orwell? TS Eliot? Woolf?
Some of their works are already classics, but there are surely dozens of other works, by these and other authors, that won't become classics because it is difficult to distribute them.
I would not expect a lot of minorities in any of the ivy league schools. They are by and large affirmitive action for rich people, movie stars, models and such. The main criterea for entrance is fame and fortune...
That may have been true 20 years ago, but it isn't true anymore.
and the main crieterea for graduation is "went to some ofo my classes and didn't stay high the entire time"...
Bold words from someone who can't spell "criterion". The pressure at MIT at least is pretty insane.
If we'd put the music of the Voyager Golden Record on a USB key in iTunes Fairplay ALC, I have little doubt that aliens with the capacity to recover a space probe from interstellar space could have decoded it (they have to be at least as smart as DVD Jon). Hell, it might have made it interesting for them.
Uh uh. We could probably retrieve an interstellar probe if it wandered through our solar system with its radio on. But without a program available to play a Fairplay-encoded AAC, it would surely look like noise to us, and probably to the aliens as well.
Even the great DVD Jon can't crack Fairplay without the license file.
Parent is troll, but I'll bite.
Everything runs fine with jpg, java, javascript, and flash.
Java and Javascript are not image formats. Flash is much broader, is a non-accessible resource hog, and is most commonly used for irritating ads (not unlike animated GIFs, I suppose).
That leaves JPEG, which is actually an image format, but a totally different one. GIF was designed, for logos: it is lossless, has a very limited color palette, and allows for some amount of transparency. JPEG was designed for photos: it's lossy, has a broad color palette, no transparency, and it looks terrible on things with crisp lines, like text or diagrams.
The real competitor to GIF is PNG, which is still lossless, but has better transparency and more colors. Unfortunately, it also has poorly-specified gamma correction, which makes it painful to use in web design.
You're right: string theory is not a scientific theory. It's a mathematical theory. That is, it's a collection of mathematical axioms and related proofs (and an extremely unpleasant one, according to a physicist friend of mine).
String theory does provide a model of physics. That is to say, if you set the parameters right, you get something that looks kind of like quantum field theory (which, by the way, is also a mathematical theory in addition to a scientific one). Unfortunately, the math is too hard to deteremine how they differ, and even once a determination is made, string theory has a lot of parameters which will have to be set before real predections are possible. Note that quantum field theories are testable, but only barely. For instance, Howard Georgi's "representations of SU(5)" theory was disproved by experiments in proton decay.
Finally, once string theory does make real predictions, they will be hard to test. In particular, they are likely to require enormous amounts of energy, and accelerator experiments can take years to run and analyze. So it will be a long time yet before string theory becomes scientific.
Go to any elite engineering school and take a survey of the top 10% of the students there. I would be shocked if at least 50% of those students are not chinese. I don't mean chinese americans, I mean chinese from china.
I call bullshit (on your statistic, not your general point).
I'm in the Stanford CS PhD program, and there are surprisingly few Chinese. Sure, there are plenty of Asians, but a reasonable fraction of them are Korean or Indian, and there are plenty of Caucasians (disproportionately many Greeks), Hispanics and Jews. I've only noticed a few Chinese from China, but I'm oblivious that way. I have no idea who the top students are, but if you count the whole PhD program, it's certainly not true.
At Harvard it wasn't true either (of course, Harvard is not an engineering school). If anything, there were more Jews than Chinese at the top, and almost all of the Chinese were 2nd or 3rd generation; this was true both undergrad and CS grad (though not as much of math grads).
At MIT, from what I know, it wasn't true either: most of the elite students I knew there were either Jewish or Caucasian. I didn't know that many, though.
At my summer jobs, there have been a lot of smart Chinese people, but only a few of them were from China.
At the math olympiad programs the years I went, half the team was Chinese, but only one or so each year from China. Of the students there in general, maybe a third were Chinese. (China wins the olympiad almost every year, though.)
This is not to say that Chinese people are stupid. Just like you, I know plenty of really smart Chinese people, although most of them are 2nd or 3rd generation.
Actually... no, it isn't. I was doing Java development this summer (on Java 1.5), and I did that a few times by accident, and it doesn't work. It asks whether str1 and str2 are the same String object, rather than asking whether the strings contained in them are equal.
You're trolling, but I'm bored, so I'll respond anyway.
OO code was developed by people who think in calculus and set theory.
Not so. The people who think in calculus and set theory developed Haskell. And if you were arguing against Haskell, you might have had a point.
Perl is good, Ruby is good, these are things that alleviate the problem a lot.
Perl was designed for writing short, simple programs which do lots of string manipulation. For that it's often the best thing out there. For anything large and complicated, it often becomes difficult to maintain.
Ruby is as object-oriented as they come, and very clean (but slow) as a result. In fact, it's directly based on the grand-daddy of object-oriented programming, Smalltalk.
Um, let me make a simpler argument, the same one I'm responding to, and it's a damned good one. Look at the VERBIAGE on those "modern" programs. What the HELL are you people thinking? Just typing all that slows down development enormously.
That would, in fact, be Java. Also Objective-C. Occasionally C++, but not so much. Pretty much everything else out there has shorter lines.
I guess the next thing to do (if I'm so clever) would be to develop a simple, usable language for non-CS students. I kinda think someone did it already, many times. But no students will ever see it.
Hah, like AppleScript? Really, I think a better thing to do would be to teach a Smalltalk-derived OO language like Python or Ruby.
Niagara is actually very nice if you have a massively parallel (~ 32 threads always on the CPU) task with no floating point operations. So it shines for enormous webservers, or servers combining every tier on one machine. Unfortunately, there's only one FPU on the whole chip, so floating point pretty much kills the performance.
Also, for anything running single-threaded, the performance is terrible. The processor runs at 1-1.2 GHz and isn't superscalar.
The combination of these factors means that not too many people will be buying SunFire T2000 servers at $13k apiece.
Enjoy paying a hundred bucks a year for features like this as an upgrade?
I'm a student, so it's only $80. And no, as I said I'm happy to use VirtueDesktops. But I enjoy paying $80 a year for upgrades like iChat AV (with H.264), Core Image/Vector/Data/Audio, Expose (which is actually quite nice), launchd, fast user switching (IIRC, there's still not a good interface for that in Ubuntu?), and so on. As it happens, I've only done this once (10.2 -> 10.3; was using Ubuntu when 10.4 came out), but I'll do it again in a heartbeat if 10.5 has virtualization or ZFS support. Without those, we'll see. DTrace is awesome, and the new iChat stuff looks pretty cool, and Time machine should be useful (if corny), but I'm not sure if that will be enough.
Now the RIAA are claiming that Paul can copy an MP3 within 45 minutes
Come on, even on a Macintosh it only takes 20 minutes. Of course, while you're copying it, Netscape does not work...
Leopard is going to have virtual desktops. In the mean time, there's Desktop Manager or VirtueDesktops for free, and CodeTek Virtual Desktops for not-so-free. The first two work pretty well; I've never used CodeTek before.
Guess I'll just have to get by with lilo and busybox.
As an EE and and audio engineer and a musician, I have a simple reply to this: There are no "insignificant" bits. Only bits that are lost due to one factor or another. Lossy compression isn't something to figure out. It is something to abandon.
The point still stands. Some bits are more significant than others, and as an engineer you must know this. Start with a very high-bandwidth encoding (96k x 6 x 24, or 6-channel high-frequency sigma-delta, or whatever). It's not practical to haul it around, so you have to compress it to something reasonable, say 500kb/s. Since all compression down to a given bitrate "loses" the same amount of information (assuming that it is optimal, and so cannot be compressed further). It's equally lossy.
What's to say that the optimal encoding (in whatever psychoacoustic sense) is a function of the 44k x 2 x 8 CD-quality version?