Unless you're talking about some harmonic the rest of us can't hear or predict, the whine from a TV should be at 15,734 Hz, which is its horizontal scanning frequency (different, but not much, for PAL instead of NTSC). That's just if I'm right about it just being something loose that vibrates in the magnetic feild, but I don't know what else it could be.
Yes - if you write software and publish it online, you could have unknowingly infringed on someone's patent, and thus become legally vulnerable. This endangers your right to excersize free speech online. (See sig.)
Honest question - in what specific instances has that issue frustrated them?
I can think of basically two - typing a filename into a load or save dialog box, and during tab completion on the command line. In both cases, the entire directory has already been scanned, so it should be easy to do the case thing - if the typed file doesn't exist but exactly one which differs only in case does, open that instead, when saving, warn if the cases don't match (with an option to always assume they should), and when tab completing, change the case to be right.
There's also what started this, include files, but I think by that point, you're not really just a user (besides, the rest of the language is case sensitive.) And for the commands themselves, it's bad for the shell to do that, as not all arguments are interpreted as filenames. Though I wouldn't be opposed to a flag to fopen(3) that would tell glibc to, if it feels like it (older versions would ignore it, as they do "b", do a case insensitive match, returning an error if there were no normal matches but multiple insensitive matches - then user-visible stuff that doesn't need performance can just use that. (And all this can get integrated with the localization support that already exists at the libc level to work correctly in other languages.)
Suppose I murder grandparent's family. I write a story about exactly how I did it, including everything except the minor detail that that I did it, not you. Then, I steal grandparent's credit cards, drivers license, SSN, and Best Buy frequent buyer's card. I include all of that info in my story, then pretend to be a journalist and publish it. Police come knocking, and I say I got it from you.
Point is, such a story isn't good evidence, because the journalist can name whomever he wants, regardless of who the real author is. Now, back to reality, if I did that, I'd be very suspect myself ("No, officer, I didn't write that, liquidsin did!") and I'd probably get caught. Now, in reality, no innocent journalist would publish the story, and if they did, they'd have an enormous personal incentive to tell truthfully who their source is.
So in practice, it's best when (a) nothing prevents a journalist from naming his sources, but (b) nothing forces him to.
This is not a case of telling a journalist *ABOUT* an illegal act, this is a case where telling the journalists *IS* the illegal act, and the journalist was party to this illegal act.
In order to give my counterargument weight, I'll follow you in ignoring the distinction between criminal and civil law...
If a journalist receiving protected information from someone with an NDA is committing an illegal act, then does it not violate the Fifth Amendment to force the journalist to identify his source (thus confirming his source was someone under an NDA)?
The whole thing is inconsistent, as it needlessly introduces a very complex many-to-one mapping between user-visible filenames and the filenames that are actually used for lookup. And thankfully, unix kernel developers have a preference for elegance and simplicity that you just seem to lack, if only because they're the ones who would have to maintain it.
The kernel is supposed to be simple and fast, and so there is a trend to move as much as possible out of it, not into it (eg. udev vs. devfs). Userspace offers a lot more flexibility for things of that nature - it can be upgraded seperately, it can more easily be customized between distros and by individuals, already has mature localization technologies, and so on. And of course a more simple and spartan kernel is easier to strip down for embedded devices.
Now, sometimes you can make the case that things should be in the kernel for performance - as Windows does for a lot of its GUI code. But here, it is unquestionably faster to just not translate the names and do an ordinary case-insensitive match in the kernel, and implement translation as a library used in the few applications that actually need it.
Software is one of the happy few fields where the easy thing is often the right thing, if only because it's fun to work on something elegant, and hard and thankless work to contribute to a quagmire of ugly hackish code. Which is probably why Linux doesn't do it - it's ugly, and Linux has the luxury of not being forced into stuff like that by backwards compatibility.
I've said about all I care to say, but if you care, you can easily find more about this on the web (try the LKML archives). It's basically a bunch of Windows users whining about how they don't care if it's a big ugly hack, and the people who know what they're talking about and would have to fix it when it breaks saying no thanks. (That sounds horribly elitist, and it is, but that's the breaks.)
You start off sounding just like the proverbial PHB - "I don't understand it, so it must be easy."
Of course you should care about implementation difficulties - they translate directly into bugs for you. And you should care about elegance - the lack of it translates into bloated software for you.
Windows actually implements case insensitivity by translating everything into one case-sensitive bitstream for saving on the disk (along with an 8.3 mangled name - ugh), and back while reading it. This is a minor performance hit. And they didn't implement it for the users, they implemented it for DOS campatibility, DOS having been written when everyone important used ASCII and case insensitivity was one extra line of code, foreigners be damned!
I understand case insensitivity can make things easier to use - but the proper place to implement this is outside of the kernel, in some library that could be used by tab-completion and GUI code. You as a user shouldn't care how this works, you should just file feature requests for accomodating users who are too lazy to press shift once in a while in user-visible software. And when you write software, you aren't a user, so just suck it in and keep track of the case of your filenames - otherwise the interpreter, compiler, or worse, kernel, has to use another MB or so of code, or else scan the entire directory when an open() would suffice, to compensate.
And finally, Windows doing something DOES NOT IMPLY that it should be done.
From Red Hat's email: "Moreover, our client does not allow others to provide links to our client's web site without permission."
God damn. Overzealous enforcement of trademark rights is one thing, but that's something else entirely! Nobody (should) have the right to control who links to their site on the Internet! If you don't want a page linked to, then don't put it up, or at least block outside referrers. It's not like the page being linked to is facilitating copyright infringement - there really is no legal reason why they can't link to anyone they damn well please, and nothing short of an act of Congress can change that.
If we weren't in the real world, but rather in some incredible physical law simulating program, what differences would you notice? What about the simulated world would make you think what you just said was false, while you believe it in this world? I say nothing.
I guess my point is, creating a virtual universe computer program and simulating life in it is nothing less than creating a universe and a physics. The only difference between the virtual universe and ours is that the virtual one is much smaller and its physics are much simpler.
Here things get really interesting - if an inhabitant of our virtual universe posited that his universe was designed by an intelligent being to foster life, he'd be right. It seems to me that the entire evolution vs creationism debate is really just a question of how much the creator, if he exists, micromanages. For example, intelligent design proponents in the Tierra universe would be right, as the author actually did write the first program. But in other artifical universes, the intent of the creator is evident only in the physics, and the specifics of life are entirely the result of a random initial state and the process of evolution.
Naturally we won't ever be able to prove or disprove the existence of a creator, much less his intents. We certainly have no good evidence that he did any Tierra-style micromanagement to get life started (or to interfere with the lives of a select few ancient people.) But all that said, a non-micromanaging creator experimenting with virtual universes (to borrow a sibling's phrase, "Jimmy playing on his quantum computer") is certainly an interesting, though wholly unscientific, model.
Well, I doubt we'll ever have complete knowledge of the present state for even a vanishingly small area. (Complete knowledge of the Earth, for example, would only allow you to make predictions for a second or two, due to the speed of light.) And of course we'd also need near-infinite (or infinite, if you want to do the whole universe) computing resources and an exactly correct understanding of physics.
As to free will - how does anybody know they have free will? Tempting as it is, a belief in free will isn't enough to make a scientific statement about the determinism of the universe. However, it's quite possible science will never know, because to prove the universe is indeterminate, you'd need to go outside the universe and use your perfect program and infinite computer to predict the universe (if your prediction was wrong, then the universe is indeterminate). And even with this method, you could never prove the universe is completely determinable. (Actually, that's a bit like proving within a system that that system is self-consistent.)
That said, our current understanding of physics agrees with you that at least some small-scall things are indeterminate. I guess my point is, while that's probably true, people thinking they have free will isn't very good supporting evidence. Probably better would be to ask grandparent if he could point out where the physicists are wrong and how they can predict the things they say are unpredictable.
Hmm, this seems to imply that people don't want to pay twice what something's worth just for the privelage of being annoyed with increasingly pervasive ads.
This is a real problem with the idea of treating not even just trademarks, but "advertising work" as protected property. We need some ads, and we certainly have them. We don't really want more ads, so there's no reason our government should protect advertising investment. (I'm not against trademarks, I just think they should only be used to let brands build a reputation for quality and keep them from lying to consumers about which company produced what.)
Consider also that a free market is most efficient when the consumers are totally informed, and that in general, we should consider this a good thing. Thus, we demand that ads not be outright lies, we demand that they don't misrepresent the origin of a product, etc. Along these lines, isn't it best if a consumer is informed about all the competing suppliers for a product he wants?
Don't worry, nobody's trying to move all blame from SCO to the lawyers - but just because it's not only the lawyers's fault doesn't mean it's not also their fault. If there's a smackdown, SCO management and its lawyers should get it equally - they both freely made a choice between money and not evil, they both sold out, and dammit, that's not OK!
Don't know what you're talking about, but I remember when graphics hardware used to suck, and the most common way to make something selected was to overlay it with a halftone of blue. So what you would do is, figure out where that halftone would go, and in the pixels that remain exposed, mix in your porn image, at say about 25% opacity. Now, on the pixels that are obscured by the halftone, mix in the inverse of your porn image at the same opacity. When the halftone is gone, it would be hard to notice the change - the most you would notice is a subtle checkerboard effect where the porn was contrasting with the flowers. But when the halftone obscured the negative that previously was balancing the positive porn image in adjacent pixels, you would see the porn in much higher contrast.
Read his book. I don't think it's at all unfair to say Linus is against software patents (though in favor of copyrights, mind you) - the only thing confusing about this article is why it's considered news. He, like most sane programmers, have been against them for years. (Maybe the new thing here is that he's bringing some of his political clout to bear on the issue, which is rare.)
Well lobbying is bad, but it's necessary because Microsoft, SCO, and Sun also lobby. (Groupthink is doubleplusgood!) But seriously - there should be limits on how much non-single-human entities are allowed to influence government.
You seem to imply that GCC's C++ exceptions don't actually work, that we have to resort to setjmp()/longjmp(), that templates don't work, that GCC's STL strings aren't copy-on-write, etc. All of these implications are, to put it bluntly, false. (If you didn't mean this, no offense, but you did imply it.)
And yes, C# is (a) pretty cool, and (b) different from C++. That's why we have Mono:-). As for debugging, I don't do that much (usually stack traces are enough), and my "IDE" is kwrite and a command line, but KDevelop, Eclipse, and many others do indeed have integrated debugging - if it craps out, file a bug report, don't just bitch on Slashdot.
Ok - for some reason my mental image was of those butane lighter with a long handle things. But still - did they have a fireplace? It's just not very strong evidence.
You've made a glaring oversight - Microsoft's development tools, for the most part (I know about the free command line compiler), are expensive! Linux as a platform has many more free development tools than Windows has, and almost infinitely more than Microsoft offers. This may not seem important in a corporate setting, but if you just want to tinker, don't have a school with "the first hit's free" contracts, and don't want to infringe copyrights, Linux is a much better platform. (I speak from personal experience here; this is why I first switched.)
Also, a big, integrated IDE is really a matter of preference, and nothing more. You gain some initial ease-of-use, but you loose a great deal of flexibility. Windows developers will probably disagree because this is what they're used to, but why would I want my code editor to also be a compiler, build system, debugger, and GUI designer? Is it really that hard to press Alt-Tab? I'm not saying it's not cool to have your editor highlight the lines compile errors are on, but it's easy to do that, not only with one big IDE made by one company with one development paradigm, but also with a plugin to parse the output of "third-party," modular, independant tools. It's the Unix way, and it's nothing if not flexible.
One minor nitpick I'd make their - Qt doesn't necessarily use X - it usually does, but it can also run on the framebuffer, or embedded devices, or Windows. (Of course, so can SDL).
I've actually been planning to give Freevo a try - primarily because mythbackend has a habit of dying on me, and a few other quirks.
Uh, maybe he bought firestarters, because he needed them? Like, to light birthday candles or something? Geez, two people, probably from the same town, using the same brand of fire-starter means absolutely nothing. Nothing at all.
This is somewhat offtopic, but I don't believe nerve cells are necessary for conciousness. (I'm refering here to weak AI, though, and I agree that nerve cells are necessary for conciousness in current humans.)
Slashdot Mirror
Unless you're talking about some harmonic the rest of us can't hear or predict, the whine from a TV should be at 15,734 Hz, which is its horizontal scanning frequency (different, but not much, for PAL instead of NTSC). That's just if I'm right about it just being something loose that vibrates in the magnetic feild, but I don't know what else it could be.
Yes - if you write software and publish it online, you could have unknowingly infringed on someone's patent, and thus become legally vulnerable. This endangers your right to excersize free speech online. (See sig.)
I can think of basically two - typing a filename into a load or save dialog box, and during tab completion on the command line. In both cases, the entire directory has already been scanned, so it should be easy to do the case thing - if the typed file doesn't exist but exactly one which differs only in case does, open that instead, when saving, warn if the cases don't match (with an option to always assume they should), and when tab completing, change the case to be right.
There's also what started this, include files, but I think by that point, you're not really just a user (besides, the rest of the language is case sensitive.) And for the commands themselves, it's bad for the shell to do that, as not all arguments are interpreted as filenames. Though I wouldn't be opposed to a flag to fopen(3) that would tell glibc to, if it feels like it (older versions would ignore it, as they do "b", do a case insensitive match, returning an error if there were no normal matches but multiple insensitive matches - then user-visible stuff that doesn't need performance can just use that. (And all this can get integrated with the localization support that already exists at the libc level to work correctly in other languages.)
Point is, such a story isn't good evidence, because the journalist can name whomever he wants, regardless of who the real author is. Now, back to reality, if I did that, I'd be very suspect myself ("No, officer, I didn't write that, liquidsin did!") and I'd probably get caught. Now, in reality, no innocent journalist would publish the story, and if they did, they'd have an enormous personal incentive to tell truthfully who their source is.
So in practice, it's best when (a) nothing prevents a journalist from naming his sources, but (b) nothing forces him to.
In order to give my counterargument weight, I'll follow you in ignoring the distinction between criminal and civil law...
If a journalist receiving protected information from someone with an NDA is committing an illegal act, then does it not violate the Fifth Amendment to force the journalist to identify his source (thus confirming his source was someone under an NDA)?
The kernel is supposed to be simple and fast, and so there is a trend to move as much as possible out of it, not into it (eg. udev vs. devfs). Userspace offers a lot more flexibility for things of that nature - it can be upgraded seperately, it can more easily be customized between distros and by individuals, already has mature localization technologies, and so on. And of course a more simple and spartan kernel is easier to strip down for embedded devices.
Now, sometimes you can make the case that things should be in the kernel for performance - as Windows does for a lot of its GUI code. But here, it is unquestionably faster to just not translate the names and do an ordinary case-insensitive match in the kernel, and implement translation as a library used in the few applications that actually need it.
Software is one of the happy few fields where the easy thing is often the right thing, if only because it's fun to work on something elegant, and hard and thankless work to contribute to a quagmire of ugly hackish code. Which is probably why Linux doesn't do it - it's ugly, and Linux has the luxury of not being forced into stuff like that by backwards compatibility.
I've said about all I care to say, but if you care, you can easily find more about this on the web (try the LKML archives). It's basically a bunch of Windows users whining about how they don't care if it's a big ugly hack, and the people who know what they're talking about and would have to fix it when it breaks saying no thanks. (That sounds horribly elitist, and it is, but that's the breaks.)
Of course you should care about implementation difficulties - they translate directly into bugs for you. And you should care about elegance - the lack of it translates into bloated software for you.
Windows actually implements case insensitivity by translating everything into one case-sensitive bitstream for saving on the disk (along with an 8.3 mangled name - ugh), and back while reading it. This is a minor performance hit. And they didn't implement it for the users, they implemented it for DOS campatibility, DOS having been written when everyone important used ASCII and case insensitivity was one extra line of code, foreigners be damned!
I understand case insensitivity can make things easier to use - but the proper place to implement this is outside of the kernel, in some library that could be used by tab-completion and GUI code. You as a user shouldn't care how this works, you should just file feature requests for accomodating users who are too lazy to press shift once in a while in user-visible software. And when you write software, you aren't a user, so just suck it in and keep track of the case of your filenames - otherwise the interpreter, compiler, or worse, kernel, has to use another MB or so of code, or else scan the entire directory when an open() would suffice, to compensate.
And finally, Windows doing something DOES NOT IMPLY that it should be done.
God damn. Overzealous enforcement of trademark rights is one thing, but that's something else entirely! Nobody (should) have the right to control who links to their site on the Internet! If you don't want a page linked to, then don't put it up, or at least block outside referrers. It's not like the page being linked to is facilitating copyright infringement - there really is no legal reason why they can't link to anyone they damn well please, and nothing short of an act of Congress can change that.
I guess my point is, creating a virtual universe computer program and simulating life in it is nothing less than creating a universe and a physics. The only difference between the virtual universe and ours is that the virtual one is much smaller and its physics are much simpler.
Here things get really interesting - if an inhabitant of our virtual universe posited that his universe was designed by an intelligent being to foster life, he'd be right. It seems to me that the entire evolution vs creationism debate is really just a question of how much the creator, if he exists, micromanages. For example, intelligent design proponents in the Tierra universe would be right, as the author actually did write the first program. But in other artifical universes, the intent of the creator is evident only in the physics, and the specifics of life are entirely the result of a random initial state and the process of evolution.
Naturally we won't ever be able to prove or disprove the existence of a creator, much less his intents. We certainly have no good evidence that he did any Tierra-style micromanagement to get life started (or to interfere with the lives of a select few ancient people.) But all that said, a non-micromanaging creator experimenting with virtual universes (to borrow a sibling's phrase, "Jimmy playing on his quantum computer") is certainly an interesting, though wholly unscientific, model.
Subjectively, same as we measure each other's conciousness.
As to free will - how does anybody know they have free will? Tempting as it is, a belief in free will isn't enough to make a scientific statement about the determinism of the universe. However, it's quite possible science will never know, because to prove the universe is indeterminate, you'd need to go outside the universe and use your perfect program and infinite computer to predict the universe (if your prediction was wrong, then the universe is indeterminate). And even with this method, you could never prove the universe is completely determinable. (Actually, that's a bit like proving within a system that that system is self-consistent.)
That said, our current understanding of physics agrees with you that at least some small-scall things are indeterminate. I guess my point is, while that's probably true, people thinking they have free will isn't very good supporting evidence. Probably better would be to ask grandparent if he could point out where the physicists are wrong and how they can predict the things they say are unpredictable.
This is a real problem with the idea of treating not even just trademarks, but "advertising work" as protected property. We need some ads, and we certainly have them. We don't really want more ads, so there's no reason our government should protect advertising investment. (I'm not against trademarks, I just think they should only be used to let brands build a reputation for quality and keep them from lying to consumers about which company produced what.)
Consider also that a free market is most efficient when the consumers are totally informed, and that in general, we should consider this a good thing. Thus, we demand that ads not be outright lies, we demand that they don't misrepresent the origin of a product, etc. Along these lines, isn't it best if a consumer is informed about all the competing suppliers for a product he wants?
Don't worry, nobody's trying to move all blame from SCO to the lawyers - but just because it's not only the lawyers's fault doesn't mean it's not also their fault. If there's a smackdown, SCO management and its lawyers should get it equally - they both freely made a choice between money and not evil, they both sold out, and dammit, that's not OK!
Actually, to me it would make more sense for a psychoacoustic model to be copyrighted, not patented.
Don't know what you're talking about, but I remember when graphics hardware used to suck, and the most common way to make something selected was to overlay it with a halftone of blue. So what you would do is, figure out where that halftone would go, and in the pixels that remain exposed, mix in your porn image, at say about 25% opacity. Now, on the pixels that are obscured by the halftone, mix in the inverse of your porn image at the same opacity. When the halftone is gone, it would be hard to notice the change - the most you would notice is a subtle checkerboard effect where the porn was contrasting with the flowers. But when the halftone obscured the negative that previously was balancing the positive porn image in adjacent pixels, you would see the porn in much higher contrast.
How about discrediting NASA's early work because 22,000 mph in a circle doesn't mean shit when you're trying to land a man on the Moon?
Read his book. I don't think it's at all unfair to say Linus is against software patents (though in favor of copyrights, mind you) - the only thing confusing about this article is why it's considered news. He, like most sane programmers, have been against them for years. (Maybe the new thing here is that he's bringing some of his political clout to bear on the issue, which is rare.)
Well lobbying is bad, but it's necessary because Microsoft, SCO, and Sun also lobby. (Groupthink is doubleplusgood!) But seriously - there should be limits on how much non-single-human entities are allowed to influence government.
And yes, C# is (a) pretty cool, and (b) different from C++. That's why we have Mono :-). As for debugging, I don't do that much (usually stack traces are enough), and my "IDE" is kwrite and a command line, but KDevelop, Eclipse, and many others do indeed have integrated debugging - if it craps out, file a bug report, don't just bitch on Slashdot.
Ok - for some reason my mental image was of those butane lighter with a long handle things. But still - did they have a fireplace? It's just not very strong evidence.
Also, a big, integrated IDE is really a matter of preference, and nothing more. You gain some initial ease-of-use, but you loose a great deal of flexibility. Windows developers will probably disagree because this is what they're used to, but why would I want my code editor to also be a compiler, build system, debugger, and GUI designer? Is it really that hard to press Alt-Tab? I'm not saying it's not cool to have your editor highlight the lines compile errors are on, but it's easy to do that, not only with one big IDE made by one company with one development paradigm, but also with a plugin to parse the output of "third-party," modular, independant tools. It's the Unix way, and it's nothing if not flexible.
I've actually been planning to give Freevo a try - primarily because mythbackend has a habit of dying on me, and a few other quirks.
Uh, maybe he bought firestarters, because he needed them? Like, to light birthday candles or something? Geez, two people, probably from the same town, using the same brand of fire-starter means absolutely nothing. Nothing at all.
There's a subtle but important difference in implications between "the 500 patents it owned" and "500 of the many thousands of patents it owns."
This is somewhat offtopic, but I don't believe nerve cells are necessary for conciousness. (I'm refering here to weak AI, though, and I agree that nerve cells are necessary for conciousness in current humans.)