now Fedora wants to change the filesystem (to look like Mac OS ???)...
I sure as heck hope the do! Let's see... Windows, Mac iOS, and Android all have the concept of an application directory where all the files for an app go. Installing an app just means copy the directory to/Applications or some such thing, and maybe doing a few system related things to inform the OS of your icon, uninstall program, etc. Android runs apps in restricted mode so they can only see their own directory, unless users give the app explicit permission to do more. As a result, any old fool can publish an app in the Android Market with 1/10th the experience and skill required to be in Debian. Do you think it is a good or bad thing that less than 10% of FOSS developers have the skills needed to contribute to Debian?
I have an idea... maybe we should instead dump all of an applications files into shared directories, where they can clobber each other... Unix was first, Gnu/Linux is brilliant, but we retain baggage that's just embarrassing. Merging/lib and/usr/lib? What a complete waste of time.
While also IANANE, I read the same thing, which is very cool, and potentially useful for small remote sites needing power. However to have a major impact on our energy future, it needs to be mainstream attached to the grid. Until we have so many reactors that they produce more than base load combined, it just doesn't make economic sense to run then at less than 100%. Nuclear power has the cheapest fuel cost, cheaper than coal. Load following means you turn off (or down) a bunch of reactors when not needed. Naturally, you turn off the ones that burn the most expensive fuel.
Combine this with the engineering challenges of building a plant that is expected to cycle a few times a year versus every day. The high temperatures in molten salt reactors makes the material challenges formidable. The MSRe had problems with tiny cracks forming in the pipes. The stress to the system from cycling is far higher in these reactors than even the already insanely high stress in a normal "low temperature" reactor. That said, It is extremely cool that the MSRe was turned off every weekend so people could go home.
Geeks interested in safe practical thorium power really need to read the history of molten salt reactors here. I hope India and China have the sense to invest in this path. The LFTR is the long term theoretical evolution of the molten salt reactor path. My only problem with the whole LFTR hype is it's pushing for massive research instead of building reactors we know how to build now. We should get back in the game now, first building a new MSR taking into account what we learned in the 60's and new advances since then, and then build a few commercial plants.
To be specific about some of the hype I don't like, check out the claimed advantages of LFTRs. Some of the advantages that LFTR theoretically inherit from MSR I wont dispute, including inherent safety, small size, and low operational cost, as MSR research proved that already in the 60's. However, I take issue with "load following" which means ramping the reactor up and down to follow the load. That's what all our other generators are good for, but to get your investment out of a nuclear reactor, you want to take advantage of it's low fuel cost and run it at 100% capacity almost all the time. This also greatly simplifies the engineering involved, and given the economics, there's simply no way our early LFTRs will be designed for load following. Then they claim minimal end-of-life expense. Cleaning up the MSR plant turned out to be massively more expensive than anyone would have guessed, though with knowledge gained from that experience, we should be able to do a better job next time. Then, they assume that the first LFTRs will use a new turbine design, rather than standard steam turbines. That might be where we eventually get, but build the first plants using cheaply available and well understood technology! This sort of hype looks more like fishing for DARPA grants than solving the energy crisis.
Well, obviously you disagree that mathematical algorithms should not be patented, though I'll point out that your example of Red Hat is a company that works as much or more than any other to end software patents. I patent software too, also for defensive reasons. It's a huge waste of time and money, but the lawyers making money of software patents and big companies that shut down competition with them through anti-competitive practices have far more influence in government that open source developers and small companies.
This is sound advice. The executor of your will is probably a good person to trust with the passwords location. This is probably what you should do.
However, if you want to think high tech, there are tons of solutions. For example, have a little server where you keep your passwords automatically check that you have been reading email in your g-mail account in the last few days. If there's been too long a period of inactivity, automatically send an email to your son with a list of all your passwords and maybe a few last words of wisdom. You could have that program do all sorts of fun stuff, like send any last minute insults you didn't have the guts to say while alive. Heh heh...
The other problem is that patents really fail at their primary purpose: encouraging disclosure.
Exactly! They do the opposite. They force companies to keep their software secret, rather than disclosing source, to reduce the chance of being sued for software patent infringement.
It's not just mathematical formulas that in theory are not patentable, but mathematical algorithms. If the US PTO stuck to it's mandate, rather than caving to pressure, it would never have allowed most software patents. Now "slide to unlock" and such things can be considered a design patent, though they should also be disallowed. As they say, if the big auto companies each patented their interface to their cars, you'd have to re-learn how to drive when you bought a new brand. We'd have steering wheels in Ford, and joy sticks in GM. The same applies to desktop environments, tablets, phones, and web pages.
Other than stupid interface patents, the real menace with software patents is mathematical algorithm patents. No one can write a 50 thousand line algorithm and tell their boss that they are confident of not violating any patents. So long as you keep your software closed source, it's probably no big deal because no one will ever discover that you violate their patent. As a result, companies are encouraged to keep their code secret, which is why we have this stupid situation with NVDA and AMD closed source device drivers, and why Skype on Linux never gets entirely fixed. It's why kids of my generation (before software patents) learned anything they wanted to about how computers worked, while today's generation of Windows users live with the machine being a black box, and have to switch to an entirely geek culture (linux) to learn how computers actually work.
I disagree to some extent. The US system is being more strict about software patents now days, which I feel is a good thing, as the rest of the world just laughs at our software patents, and coders here in the US have to dumb down their code. IBM and many other companies have made binding commitments not to sue open source projects, and have even donated patents to an open source defense pool. Closed source projects aren't effected much by software patents because no one will ever prove you violate them, unless it's some stupid GUI related thing like menus or one click buy buttons, or if two companies have already gone to war and convince judges to allow their guys to peek at each other's code. So, in effect, all we really need is to stop allowing stupid GUI patents like multi-touch gestures on a small screen, and discourage law suits against open source developers, and maybe stop allowing companies to force other companies to reveal source code. It's hard, but there's movement in that direction.
All of my software patents used to come with full source code. They stopped asking for that several years ago. I don't think having source or not makes any difference. If idea is expressed clearly in the body of the patent, a coder in the field not only should be capable of writing it, but in my experience, they typically prefer to write it themselves. It's not easy to get other people to use your software, even when you give them source. I've got maybe 20 open source projects out there, and maybe three or four get any significant use at all, and I doubt anyone looks at my code. The ideas are the important part. Mathematical algorithms simply should not qualify as patentable. It's bad for innovation, as he original patent system creators understood.
In general, the fact that no else seems to be using your idea is enough to show that it's either A) not obvious, or B) not useful. If the patent office allows you to patent a not useful idea, no one is harmed, so it's simplest for them to just assume that it is A, and not argue the inventors over how wonderful or not their ideas are.
The main problem with this is when people file sub-marine patents as patent trolls. For example, I'm going to guess that tablets and phones will continue to integrate more sensors. So, I could file a patent on one that has a 3D motion sensor for controlling software by waving your hands in front of one or two cameras in the device. I could patent a phone with a heart rate monitor, or a one button 911 feature that puts an ongoing crime's location and video instantly on the web and alerts the police. I could file a ton of dumb ideas, and delay having them published (I just finished a patent that took 10 years - not on purpose). Once Samsung or Apple is pushing it as their next hot feature, I get my patent published and sue them for insane amounts of money. I don't know the answer for patent trolls, but we need to deal with the problem. The stupid big-business answer that seems to be going into effect is to limit damages, rather than troll patents. This in effect allows them to sue you for everything your worth (a millon is enough for most of us), while we can only sue them for chump change they don't care about.
Masks are copyrighted. The designs they implement can generally be patented, but those designs are described in patents as certain interconnections of devices typically. Violations of circuit patent is generally easily verified. You just send the competitor's device to have the mask layers extracted from an actual die and run circuit extraction on the result.
VHDL and Verilog are normally copyrighted. When they implement patented algorithms, it's effectively a software patent, which should also be banned as patenting mathematical algorithms. However, most patents are things like "Connecting a first fifo to a blah blah blah", where rather than patenting an algorithm, they patent a certain connection of circuit blocks. I would argue that is not a mathematical algorithm.
The original authors of our patent system were wise to ban mathematical algorithm patents. Such patents present a real threat to the free flow of ideas in universities and in open source software. They hold back progress. While it is easy to find patent violations in open source software, it's nearly impossible in proprietary binary-only software. As a result, we get vague threats like "Linux violates 100 of our patents. Pay us royalties *or else*." When you ask which patents you violate, they say it's a secret. The system is so screwed up, it's unbelievable it has supporters at all. I personally have several software patents. For the first few years, I refused to file them, and only filed hardware patents, but then a competitor patented an algorithm we'd been using for years, so I gave up and patented everything that was allowed by law. The result was a huge waste of time and money at both our companies, with zero benefit. Only the lawyers win.
Academic studies always show that we prefer female voices, but in reality, those who count on computer generated voices all day prefer male voices. Note all the readers mentioning James Earl Jones. The reason studies like this get it wrong is simple. They get random groups of people who never use computer voices to take part in their experiments, and such people initially prefer female voices. If the experiments were to run long enough for participants to become expert listeners, they would find they trend towards male voices. I do a lot of looking at speech signals, and my unsupported theory as to why we switch to male voices is that male voices cover a broader portion of the sweet spot in our hearing, where we perceive sound most sensitively. This makes male voices easier to listen to if you have to listen for long periods of time. They are also easier to understand in noisy environments, thus thus the classic low male ham radio voice.
To understand what people like when they have to listen a lot to computer generated voices, just ask the blind. I was the tech lead for Vinux 3.0, which is Linux for the Vision Impaired. That doesn't make me an expert, but here are my observations. The most popular voices for blind programmers are male, probably eloquence first (it can play very fast), followed by espeak (because it's free and everywhere), followed by various low-speed commercial male voices. The most popular Mac voice for the blind is Adam, a mechanical guy with a decent voice that can play at decent speed. The female voices are often discussed, usually with adjectives like sexy, emotional, sultry, and so on, but in the end the blind go back to their male TTS engines to get work done.
I did a lot of testing to try and speed up voices to the speeds the blind like to listen. The result is the sonic library, which powers speech speed-up in various programs like the Astro Nova player at up to 6X playback speed. At least one blind lister can listen with high comprehension to a male voice (eloquence) at about 1,500 words per minute, or 7X the default speed of this high speed voice. At this speed, the original vowels are typically compressed to one or two pitch periods. It's incredible that a blind listener can still perceive these as whole phonemes. To achieve higher speed, I've told him he needs to consider listening to a female voice, where I could get perhaps twice as many pitch periods into the same 10-ish milliseconds where he currently perceives one phoneme. The problem is that at higher fundamental pitch, this voice will register on a smaller portion of his hearing bandwidth, making it harder to get as much information out of it high speed. We've not yet had any luck with high speed female voices.
As a person losing central vision, I experienced all this myself. When I first started using computer generated voices, I tried to find a female voice I could live with. I tried a couple of smooth female Cepstral voices, but before long they sounded grating and frustratingly slow. My blind friend told me to avoid the "natural" voices and go with something that I could listen to at high speed, but I just couldn't stand the mechanical voice he was pushing - eloquence. Well, he was right. I eventually migrated to eloquence just like him and many blind people. James Earl Jones has the perfect voice for computers. It's low enough to take advantage of every bit of bandwidth we easily perceive. It's broad spectrum, taking advantage of the high frequencies as well, and very consistent, making his voice addictive. Your ear likes consistency.
Hey, great post. I guess this thread is pretty dead, but I love learning from the people who were actually there. I didn't know they actually switched OS-es... amazing because I was using their hardware at the time. I'm not familiar with HP's open-source effort... I guess it was so late and so lame it had little impact. Sun at least went out with the world's highest respect. In comparison, HP is very sad indeed.
I couldn't agree more. I mentioned Sun because everyone talks about them as a disappointing failure. However, I believe that a company like Sun that did the best that any company ever did at innovating in their market deserves tremendous respect. It's not Sun's fault that workstations are no longer needed. Rather than bashing them for not re-inventing themselves as cell-phone manufacturers or some such nonsense, we should realize that they practically invented the modern workstation, and stood by them to the end, much to their customer's delight. In addition, in their later years, Sun became the single most charitable company to open source efforts. You just have to love these guys, even if they all had to find new jobs.
Jobs, Gates, Torvalds... I've spent a couple of decades slowly coming around to the idea that it's individual leaders who make the real difference, but everyone else who does the real work. Down below a guy is talking about all the brilliant engineers at Apple working 80+ hours per week, and how that's what really counts, and this CEO worship is stupid. Do you think Sun didn't have those guys? Did they fail because they were stupid and lazy?
The world is far better off for having let Jobs lead. It's not that guys like Jobs are impossible to find, but getting them into places of power where they have permission to change the world... that is impossible to find. Sony had a guy like Jobs, and when he died, it left the music player market wide open for Jobs. The cell phone industry hadn't had a brilliant innovator with permission to make a difference since the Razor. Jobs fixed it. Bill Gates hasn't been a factor at Microsoft for a decade, and Jobs has walked all over them. The world is far better off due to what Jobs accomplished.
That said, Jobs was an ass hole, who both saved us from the stupidity of failing innovation in several markets, and shacked us with the most oppressive crap ever invented. If you ever wondered why e-books now cost the same everywhere you normally by them, that's 100% pure evil genius, courtesy of Jobs, called the "agency model". Sony is banned from iPads, and GPL 3 software can't even run on any iOS device legally. Jobs arrogance had no bounds. He wanted no less than 30% of every dollar you spend, even if it was on bread. No one knows of any significant charity supported by this man.
I hope Apple can find a new middle ground going forward that is somewhat innovative and somewhat less evil. However, the great man has passed through the veil, and while I'm not bashing Apple, there's no way their board could possibly let another Jobs take control. My guess is stock price is headed down... however, my stock advice is generally wrong.
Sun's plan worked brilliantly. They dominated workstation markets pretty much everywhere, and their open source policy was central to that. The main reason for Sun's early success (remember how they killed Apollo?) was they just ran the same software that all the universities were using. I remember how my own code just ran on the things... it was awesome. Their open-source policy, which was long before the phrase "open source" was coined, enabled them to trash the competition. Then they got big, and as the number one company in their space, they got cold feet about open source. IMO, the single biggest mistake Sun ever made was to take Berkeley Unix private, and relabel it Sun-OS (and later, Solaris). It was unbelievably super-dumb. Had they kept it open, there never would have been any compelling reason for Linux, much less BSD. My guess is that they realized their mistake, and tried to the end to make up for it by becoming radical supporters of open source. It was too little, too late. However, it wasn't that their open source strategy failed. It was their choice of back-stabbing the community that killed them. Well, that and Moore's Law, and the lack of an evil marketing genius like Jobs.
Anyway, Meg isn't the marketing genius HP needs. She pretty much is a nail in the coffin.
I just don't get how people view Sun. They were the greatest workstation provider ever. They were number one, and never lost that spot. They were the opposite of stupid... they were the geniuses who pioneered high end personal computing. What killed Sun was Moore's Law. It wasn't their fault that workstations weren't wanted anymore. In fact, Intel is facing a similar problem right now. Who want's to pay hundreds of dollars for a CPU when my ARM based phone has more power than I need? Sure, if Sun had an evil marketing genius like Jobs, they could have morphed into Google or some such nonsense. Or maybe it could have been JBL or Zenith or Atari or Sony or HP or Motorola, or any other great company who's time had ended, though I hate to compare dumb companies like these who simply failed to execute to Sun, who innovated to the end. I just can't blame a company for being number one in their market while making their customers happy. The market went away, and Sun followed. Why do we feel they should have morphed into some other company with nothing in common with building workstations? Do we really think Sun should be selling on-line advertising or ebooks?
Anyway, Oracle is on my sh*t-list. They "dumped most of the loser projects", including virtually all support for accessibility software for the blind. This is a hot-button issue for me, and as far as I'm concerned, the decision makers at Oracle deserve to burn in Hell.
I suppose this isn't very helpful advice, but I recommend switching to free software for all the tools where you don't have to have the best possible solution. There's almost no Microsoft software in use where I work, and while I am a fan of Microsoft, I like the freedom free (as in speech) software brings us.
I'm somewhat tipsy (I'm hanging out at my in-laws, and need the drink), but I'll dive into the endless language debate. I LOVE Python, and have only good things to say about it. But, I wont use it for CPU intensive tasks, as it's not meant for it. For CPU intensive algorithms, it's pretty much C, C++, or Java (sorry Lua). C++ by default (without writing your own memory allocators) is about the same speed as Java, in my experience. Of course, in C, you have to do it all. If you do that well, you'll kick ass.
You sound pretty knowledgeable about Java and the Dalvik JVM, and I enjoyed reading your post. I'll just point out that the speed and memory issues of Java from Oracle vs Dalvik are so in the noise compared to real optimization, it's pointless to bicker about it. Both systems suffer from insane inefficiency due to memory layout, resulting in horrible cache performance. If you want to talk about pure CPU efficiency for applications that are not memory hogs, I can tell you from benchmarks I've performed that Oracle's JVM is very close to hand coded C performance. I found that my Sonic speech speed-up library performed only 4% slower in Oracle's Java and OpenJDK. There is simply no way that the Dalvik JVM does much better. So, if you really think performance is a compelling reason to choose Dalvik, I implore you to offer some compelling evidence that there is any performance to be had. Both Dalvik and Oracle's JVM suck, but both are quite comparable in their lack of performance.
The fact that this is the first comment on this article is pretty ironic, given that it's these kind of attitudes that keep women away.
Good point. I do the sort of algorithms where you really need a solid personality disorder to do well. If you like working mostly by yourself for years on end to achieve algorithms that deliver better performing ICs, you're a geek like me and those I've hired for two decades. I know some brilliant women who are programmers, but not even one woman has ever applied for one of my jobs. There are differences between men and women, but no other test seems to separate the sexes so strongly as what I call hard core algorithms. But, our field is dominated by men who are really boys who never grow up.
I don't have the scientific credentials to question either. However, the article throws around statements like "warming caused by carbon dioxide (not much)" without any material backing it up. I find the average article about alien landings in National Enquirer more compelling. How does this guy rate a/. post?
One has to consider, then, why "blacks" aren't as 'smart' as whites or Asians, now, then
Go read Guns, Germs, and Steel. Ideas like northern Europeans conquered the globe because we're smarter, or we're smarter because it's harder to live in cold climates are rubbish. This book is pretty long, and has big words. Don't hurt your brain.
If you look at northern Europeans (like me), it's easy to see we are decended from Neanderthals. Where do you think the white skin, blond hair, blue eyes, and huge noses came from? Duh. The stupid white Neanderthals mated with super-smart blacks from Africa, creating the modern race of whites.
Slashdot Mirror
I sure as heck hope the do! Let's see... Windows, Mac iOS, and Android all have the concept of an application directory where all the files for an app go. Installing an app just means copy the directory to /Applications or some such thing, and maybe doing a few system related things to inform the OS of your icon, uninstall program, etc. Android runs apps in restricted mode so they can only see their own directory, unless users give the app explicit permission to do more. As a result, any old fool can publish an app in the Android Market with 1/10th the experience and skill required to be in Debian. Do you think it is a good or bad thing that less than 10% of FOSS developers have the skills needed to contribute to Debian?
I have an idea... maybe we should instead dump all of an applications files into shared directories, where they can clobber each other... Unix was first, Gnu/Linux is brilliant, but we retain baggage that's just embarrassing. Merging /lib and /usr/lib? What a complete waste of time.
While also IANANE, I read the same thing, which is very cool, and potentially useful for small remote sites needing power. However to have a major impact on our energy future, it needs to be mainstream attached to the grid. Until we have so many reactors that they produce more than base load combined, it just doesn't make economic sense to run then at less than 100%. Nuclear power has the cheapest fuel cost, cheaper than coal. Load following means you turn off (or down) a bunch of reactors when not needed. Naturally, you turn off the ones that burn the most expensive fuel.
Combine this with the engineering challenges of building a plant that is expected to cycle a few times a year versus every day. The high temperatures in molten salt reactors makes the material challenges formidable. The MSRe had problems with tiny cracks forming in the pipes. The stress to the system from cycling is far higher in these reactors than even the already insanely high stress in a normal "low temperature" reactor. That said, It is extremely cool that the MSRe was turned off every weekend so people could go home.
Geeks interested in safe practical thorium power really need to read the history of molten salt reactors here. I hope India and China have the sense to invest in this path. The LFTR is the long term theoretical evolution of the molten salt reactor path. My only problem with the whole LFTR hype is it's pushing for massive research instead of building reactors we know how to build now. We should get back in the game now, first building a new MSR taking into account what we learned in the 60's and new advances since then, and then build a few commercial plants.
To be specific about some of the hype I don't like, check out the claimed advantages of LFTRs. Some of the advantages that LFTR theoretically inherit from MSR I wont dispute, including inherent safety, small size, and low operational cost, as MSR research proved that already in the 60's. However, I take issue with "load following" which means ramping the reactor up and down to follow the load. That's what all our other generators are good for, but to get your investment out of a nuclear reactor, you want to take advantage of it's low fuel cost and run it at 100% capacity almost all the time. This also greatly simplifies the engineering involved, and given the economics, there's simply no way our early LFTRs will be designed for load following. Then they claim minimal end-of-life expense. Cleaning up the MSR plant turned out to be massively more expensive than anyone would have guessed, though with knowledge gained from that experience, we should be able to do a better job next time. Then, they assume that the first LFTRs will use a new turbine design, rather than standard steam turbines. That might be where we eventually get, but build the first plants using cheaply available and well understood technology! This sort of hype looks more like fishing for DARPA grants than solving the energy crisis.
Well, obviously you disagree that mathematical algorithms should not be patented, though I'll point out that your example of Red Hat is a company that works as much or more than any other to end software patents. I patent software too, also for defensive reasons. It's a huge waste of time and money, but the lawyers making money of software patents and big companies that shut down competition with them through anti-competitive practices have far more influence in government that open source developers and small companies.
This is sound advice. The executor of your will is probably a good person to trust with the passwords location. This is probably what you should do.
However, if you want to think high tech, there are tons of solutions. For example, have a little server where you keep your passwords automatically check that you have been reading email in your g-mail account in the last few days. If there's been too long a period of inactivity, automatically send an email to your son with a list of all your passwords and maybe a few last words of wisdom. You could have that program do all sorts of fun stuff, like send any last minute insults you didn't have the guts to say while alive. Heh heh...
Exactly! They do the opposite. They force companies to keep their software secret, rather than disclosing source, to reduce the chance of being sued for software patent infringement.
It's not just mathematical formulas that in theory are not patentable, but mathematical algorithms. If the US PTO stuck to it's mandate, rather than caving to pressure, it would never have allowed most software patents. Now "slide to unlock" and such things can be considered a design patent, though they should also be disallowed. As they say, if the big auto companies each patented their interface to their cars, you'd have to re-learn how to drive when you bought a new brand. We'd have steering wheels in Ford, and joy sticks in GM. The same applies to desktop environments, tablets, phones, and web pages.
Other than stupid interface patents, the real menace with software patents is mathematical algorithm patents. No one can write a 50 thousand line algorithm and tell their boss that they are confident of not violating any patents. So long as you keep your software closed source, it's probably no big deal because no one will ever discover that you violate their patent. As a result, companies are encouraged to keep their code secret, which is why we have this stupid situation with NVDA and AMD closed source device drivers, and why Skype on Linux never gets entirely fixed. It's why kids of my generation (before software patents) learned anything they wanted to about how computers worked, while today's generation of Windows users live with the machine being a black box, and have to switch to an entirely geek culture (linux) to learn how computers actually work.
I disagree to some extent. The US system is being more strict about software patents now days, which I feel is a good thing, as the rest of the world just laughs at our software patents, and coders here in the US have to dumb down their code. IBM and many other companies have made binding commitments not to sue open source projects, and have even donated patents to an open source defense pool. Closed source projects aren't effected much by software patents because no one will ever prove you violate them, unless it's some stupid GUI related thing like menus or one click buy buttons, or if two companies have already gone to war and convince judges to allow their guys to peek at each other's code. So, in effect, all we really need is to stop allowing stupid GUI patents like multi-touch gestures on a small screen, and discourage law suits against open source developers, and maybe stop allowing companies to force other companies to reveal source code. It's hard, but there's movement in that direction.
All of my software patents used to come with full source code. They stopped asking for that several years ago. I don't think having source or not makes any difference. If idea is expressed clearly in the body of the patent, a coder in the field not only should be capable of writing it, but in my experience, they typically prefer to write it themselves. It's not easy to get other people to use your software, even when you give them source. I've got maybe 20 open source projects out there, and maybe three or four get any significant use at all, and I doubt anyone looks at my code. The ideas are the important part. Mathematical algorithms simply should not qualify as patentable. It's bad for innovation, as he original patent system creators understood.
In general, the fact that no else seems to be using your idea is enough to show that it's either A) not obvious, or B) not useful. If the patent office allows you to patent a not useful idea, no one is harmed, so it's simplest for them to just assume that it is A, and not argue the inventors over how wonderful or not their ideas are.
The main problem with this is when people file sub-marine patents as patent trolls. For example, I'm going to guess that tablets and phones will continue to integrate more sensors. So, I could file a patent on one that has a 3D motion sensor for controlling software by waving your hands in front of one or two cameras in the device. I could patent a phone with a heart rate monitor, or a one button 911 feature that puts an ongoing crime's location and video instantly on the web and alerts the police. I could file a ton of dumb ideas, and delay having them published (I just finished a patent that took 10 years - not on purpose). Once Samsung or Apple is pushing it as their next hot feature, I get my patent published and sue them for insane amounts of money. I don't know the answer for patent trolls, but we need to deal with the problem. The stupid big-business answer that seems to be going into effect is to limit damages, rather than troll patents. This in effect allows them to sue you for everything your worth (a millon is enough for most of us), while we can only sue them for chump change they don't care about.
Masks are copyrighted. The designs they implement can generally be patented, but those designs are described in patents as certain interconnections of devices typically. Violations of circuit patent is generally easily verified. You just send the competitor's device to have the mask layers extracted from an actual die and run circuit extraction on the result.
VHDL and Verilog are normally copyrighted. When they implement patented algorithms, it's effectively a software patent, which should also be banned as patenting mathematical algorithms. However, most patents are things like "Connecting a first fifo to a blah blah blah", where rather than patenting an algorithm, they patent a certain connection of circuit blocks. I would argue that is not a mathematical algorithm.
The original authors of our patent system were wise to ban mathematical algorithm patents. Such patents present a real threat to the free flow of ideas in universities and in open source software. They hold back progress. While it is easy to find patent violations in open source software, it's nearly impossible in proprietary binary-only software. As a result, we get vague threats like "Linux violates 100 of our patents. Pay us royalties *or else*." When you ask which patents you violate, they say it's a secret. The system is so screwed up, it's unbelievable it has supporters at all. I personally have several software patents. For the first few years, I refused to file them, and only filed hardware patents, but then a competitor patented an algorithm we'd been using for years, so I gave up and patented everything that was allowed by law. The result was a huge waste of time and money at both our companies, with zero benefit. Only the lawyers win.
Academic studies always show that we prefer female voices, but in reality, those who count on computer generated voices all day prefer male voices. Note all the readers mentioning James Earl Jones. The reason studies like this get it wrong is simple. They get random groups of people who never use computer voices to take part in their experiments, and such people initially prefer female voices. If the experiments were to run long enough for participants to become expert listeners, they would find they trend towards male voices. I do a lot of looking at speech signals, and my unsupported theory as to why we switch to male voices is that male voices cover a broader portion of the sweet spot in our hearing, where we perceive sound most sensitively. This makes male voices easier to listen to if you have to listen for long periods of time. They are also easier to understand in noisy environments, thus thus the classic low male ham radio voice.
To understand what people like when they have to listen a lot to computer generated voices, just ask the blind. I was the tech lead for Vinux 3.0, which is Linux for the Vision Impaired. That doesn't make me an expert, but here are my observations. The most popular voices for blind programmers are male, probably eloquence first (it can play very fast), followed by espeak (because it's free and everywhere), followed by various low-speed commercial male voices. The most popular Mac voice for the blind is Adam, a mechanical guy with a decent voice that can play at decent speed. The female voices are often discussed, usually with adjectives like sexy, emotional, sultry, and so on, but in the end the blind go back to their male TTS engines to get work done.
I did a lot of testing to try and speed up voices to the speeds the blind like to listen. The result is the sonic library, which powers speech speed-up in various programs like the Astro Nova player at up to 6X playback speed. At least one blind lister can listen with high comprehension to a male voice (eloquence) at about 1,500 words per minute, or 7X the default speed of this high speed voice. At this speed, the original vowels are typically compressed to one or two pitch periods. It's incredible that a blind listener can still perceive these as whole phonemes. To achieve higher speed, I've told him he needs to consider listening to a female voice, where I could get perhaps twice as many pitch periods into the same 10-ish milliseconds where he currently perceives one phoneme. The problem is that at higher fundamental pitch, this voice will register on a smaller portion of his hearing bandwidth, making it harder to get as much information out of it high speed. We've not yet had any luck with high speed female voices.
As a person losing central vision, I experienced all this myself. When I first started using computer generated voices, I tried to find a female voice I could live with. I tried a couple of smooth female Cepstral voices, but before long they sounded grating and frustratingly slow. My blind friend told me to avoid the "natural" voices and go with something that I could listen to at high speed, but I just couldn't stand the mechanical voice he was pushing - eloquence. Well, he was right. I eventually migrated to eloquence just like him and many blind people. James Earl Jones has the perfect voice for computers. It's low enough to take advantage of every bit of bandwidth we easily perceive. It's broad spectrum, taking advantage of the high frequencies as well, and very consistent, making his voice addictive. Your ear likes consistency.
Hey, great post. I guess this thread is pretty dead, but I love learning from the people who were actually there. I didn't know they actually switched OS-es... amazing because I was using their hardware at the time. I'm not familiar with HP's open-source effort... I guess it was so late and so lame it had little impact. Sun at least went out with the world's highest respect. In comparison, HP is very sad indeed.
I couldn't agree more. I mentioned Sun because everyone talks about them as a disappointing failure. However, I believe that a company like Sun that did the best that any company ever did at innovating in their market deserves tremendous respect. It's not Sun's fault that workstations are no longer needed. Rather than bashing them for not re-inventing themselves as cell-phone manufacturers or some such nonsense, we should realize that they practically invented the modern workstation, and stood by them to the end, much to their customer's delight. In addition, in their later years, Sun became the single most charitable company to open source efforts. You just have to love these guys, even if they all had to find new jobs.
Jobs, Gates, Torvalds... I've spent a couple of decades slowly coming around to the idea that it's individual leaders who make the real difference, but everyone else who does the real work. Down below a guy is talking about all the brilliant engineers at Apple working 80+ hours per week, and how that's what really counts, and this CEO worship is stupid. Do you think Sun didn't have those guys? Did they fail because they were stupid and lazy?
The world is far better off for having let Jobs lead. It's not that guys like Jobs are impossible to find, but getting them into places of power where they have permission to change the world... that is impossible to find. Sony had a guy like Jobs, and when he died, it left the music player market wide open for Jobs. The cell phone industry hadn't had a brilliant innovator with permission to make a difference since the Razor. Jobs fixed it. Bill Gates hasn't been a factor at Microsoft for a decade, and Jobs has walked all over them. The world is far better off due to what Jobs accomplished.
That said, Jobs was an ass hole, who both saved us from the stupidity of failing innovation in several markets, and shacked us with the most oppressive crap ever invented. If you ever wondered why e-books now cost the same everywhere you normally by them, that's 100% pure evil genius, courtesy of Jobs, called the "agency model". Sony is banned from iPads, and GPL 3 software can't even run on any iOS device legally. Jobs arrogance had no bounds. He wanted no less than 30% of every dollar you spend, even if it was on bread. No one knows of any significant charity supported by this man.
I hope Apple can find a new middle ground going forward that is somewhat innovative and somewhat less evil. However, the great man has passed through the veil, and while I'm not bashing Apple, there's no way their board could possibly let another Jobs take control. My guess is stock price is headed down... however, my stock advice is generally wrong.
Loved the post. Seriously... you didn't even get one mod point for that? Slashdotters...
Sun's plan worked brilliantly. They dominated workstation markets pretty much everywhere, and their open source policy was central to that. The main reason for Sun's early success (remember how they killed Apollo?) was they just ran the same software that all the universities were using. I remember how my own code just ran on the things... it was awesome. Their open-source policy, which was long before the phrase "open source" was coined, enabled them to trash the competition. Then they got big, and as the number one company in their space, they got cold feet about open source. IMO, the single biggest mistake Sun ever made was to take Berkeley Unix private, and relabel it Sun-OS (and later, Solaris). It was unbelievably super-dumb. Had they kept it open, there never would have been any compelling reason for Linux, much less BSD. My guess is that they realized their mistake, and tried to the end to make up for it by becoming radical supporters of open source. It was too little, too late. However, it wasn't that their open source strategy failed. It was their choice of back-stabbing the community that killed them. Well, that and Moore's Law, and the lack of an evil marketing genius like Jobs.
Anyway, Meg isn't the marketing genius HP needs. She pretty much is a nail in the coffin.
I just don't get how people view Sun. They were the greatest workstation provider ever. They were number one, and never lost that spot. They were the opposite of stupid... they were the geniuses who pioneered high end personal computing. What killed Sun was Moore's Law. It wasn't their fault that workstations weren't wanted anymore. In fact, Intel is facing a similar problem right now. Who want's to pay hundreds of dollars for a CPU when my ARM based phone has more power than I need? Sure, if Sun had an evil marketing genius like Jobs, they could have morphed into Google or some such nonsense. Or maybe it could have been JBL or Zenith or Atari or Sony or HP or Motorola, or any other great company who's time had ended, though I hate to compare dumb companies like these who simply failed to execute to Sun, who innovated to the end. I just can't blame a company for being number one in their market while making their customers happy. The market went away, and Sun followed. Why do we feel they should have morphed into some other company with nothing in common with building workstations? Do we really think Sun should be selling on-line advertising or ebooks?
Anyway, Oracle is on my sh*t-list. They "dumped most of the loser projects", including virtually all support for accessibility software for the blind. This is a hot-button issue for me, and as far as I'm concerned, the decision makers at Oracle deserve to burn in Hell.
I suppose this isn't very helpful advice, but I recommend switching to free software for all the tools where you don't have to have the best possible solution. There's almost no Microsoft software in use where I work, and while I am a fan of Microsoft, I like the freedom free (as in speech) software brings us.
I'm somewhat tipsy (I'm hanging out at my in-laws, and need the drink), but I'll dive into the endless language debate. I LOVE Python, and have only good things to say about it. But, I wont use it for CPU intensive tasks, as it's not meant for it. For CPU intensive algorithms, it's pretty much C, C++, or Java (sorry Lua). C++ by default (without writing your own memory allocators) is about the same speed as Java, in my experience. Of course, in C, you have to do it all. If you do that well, you'll kick ass.
You sound pretty knowledgeable about Java and the Dalvik JVM, and I enjoyed reading your post. I'll just point out that the speed and memory issues of Java from Oracle vs Dalvik are so in the noise compared to real optimization, it's pointless to bicker about it. Both systems suffer from insane inefficiency due to memory layout, resulting in horrible cache performance. If you want to talk about pure CPU efficiency for applications that are not memory hogs, I can tell you from benchmarks I've performed that Oracle's JVM is very close to hand coded C performance. I found that my Sonic speech speed-up library performed only 4% slower in Oracle's Java and OpenJDK. There is simply no way that the Dalvik JVM does much better. So, if you really think performance is a compelling reason to choose Dalvik, I implore you to offer some compelling evidence that there is any performance to be had. Both Dalvik and Oracle's JVM suck, but both are quite comparable in their lack of performance.
Good point. I do the sort of algorithms where you really need a solid personality disorder to do well. If you like working mostly by yourself for years on end to achieve algorithms that deliver better performing ICs, you're a geek like me and those I've hired for two decades. I know some brilliant women who are programmers, but not even one woman has ever applied for one of my jobs. There are differences between men and women, but no other test seems to separate the sexes so strongly as what I call hard core algorithms. But, our field is dominated by men who are really boys who never grow up.
I don't have the scientific credentials to question either. However, the article throws around statements like "warming caused by carbon dioxide (not much)" without any material backing it up. I find the average article about alien landings in National Enquirer more compelling. How does this guy rate a /. post?
Go read Guns, Germs, and Steel. Ideas like northern Europeans conquered the globe because we're smarter, or we're smarter because it's harder to live in cold climates are rubbish. This book is pretty long, and has big words. Don't hurt your brain.
If you look at northern Europeans (like me), it's easy to see we are decended from Neanderthals. Where do you think the white skin, blond hair, blue eyes, and huge noses came from? Duh. The stupid white Neanderthals mated with super-smart blacks from Africa, creating the modern race of whites.