That's an understatement if I've ever heard one. The thing is fricking GINORMOUS! 9MB for "Hello, World"? 900MB for a UML application?!??!?!
Re:What about ad-hoc cash transfers?
on
Cashless Society
·
· Score: 1
This can't really be ubiquitous until everyone has a reader in their home (hooked to their PC, or maybe just a cheap standalone box) that can do transfers and check value and stuff. Also everything that has a reader to make purchases should be able to show your card's value for free. Also the $100 limit needs to be higher. Also it should be totally free (and Free) and provided by the government as a way of replacing high-maintenance paper and metal money.
Personally I think it would be really nice to carry around a "cash card" instead of actual bills/coins if all the above conditions were met. Smaller, easier to handle, and less smelly. If e-ink-type-stuff were added to the card to show its value at all times, then it would be a no-brainer. Of course, people older than about 35 would hate the new system "just because", so the transition couldn't be complete until they all are too old to have any influence.
That is the vision of GStreamer. Progress seems fairly slow, but steady so far. I think GStreamer is even distributed with Gnome now. Hopefully GStreamer will soon mature to the point where gst-player can replace mplayer/xine as a general media player.
Freetype's font auto-hinter does this. So if you're using anti-aliased fonts on XFree86, you are seeing this idea in action (unless you're using the truetype font hints in violation of Apple's patents, but they do much the same thing anyway). I don't know if any SVG implementation does this yet.
They've probably never heard of mplayer. After all, it was only fairly recently that Sorensen became available through mplayer. It does indeed play the trailer (unfortunately with some sound lag on my machine though). I don't think Apple wants to publicly endorse mplayer, since it is kind of a hack that required reverse-engineering the interface quicktime uses to call quicktime codecs on Windows. Most companies are strangely reluctant to endorse hacks, even hacks that work well.
Once again, for a Linux player that can handle this Quicktime Sorensen file, look no further than the superlative mplayer. It has had the capability to play quicktime movie trailers for a while now, but I figure some people still haven't heard. Just make sure you have version 0.90 rc1 or better, because that's when they reverse-engineered the windows quicktime codecs.
Today has been plenty fast for me. I haven't noticed any problems yet (well, other than not being able to register Visual Studio.NET, is there a crack out there?). The internet really is quite resilient.
You shouldn't spend less time auditing. You should not sacrifice security
But you're not sacrificing security. If you spend less time auditing code that has much fewer security holes, you get the same level of security in the end. You can never guarantee perfect security, only extremely good security. If the language gives you a head start, it takes less time to achieve the same level of security.
Pairing alloc/free's in properly designed code takes all of about a tenth of a second.
Oh, right. And next you'll tell me that loop constructs are for lazy programmers because constructing them with gotos only takes "about a tenth of a second". Give me a break. For one thing, it takes longer than that in any reasonably complex project. For another thing, humans will occasionally make a mistake. Poof, security hole. Compilers don't make those kind of mistakes. Why not let the compiler do it, both for coding speed and security reasons?
The numerical range of values that can be stored in an int or float isn't really the issue. You can simply scale the values to get any range you desire, and the mapping of actual intensities on the range is arbitrary anyway. The post above explains that the reason they use floats is because the amount of precision changes logarithmically over the range of values and the percieved brightness of light is logarithmic with the intensity, which makes more sense.
How is having a 16-bit floating point value per channel different from having an 16-bit integer per channel? You can't represent more data with a 16-bit floating point number than you can with a 16-bit integer. 16 bits of information is 16 bits of information no matter how you slice it.
I also don't see the difference you are talking about with the "intensity" values as opposed to color. If the 8-bit 'r' value in a 24-bit pixel isn't measuring the intensity of red light at that pixel, then what is it measuring?
I guess my basic question is: Why is this so special? Why couldn't someone, for example, extend PNG to have 16 bits per channel (if it doesn't already) and forget about OpenEXR?
Oh come on. Of course any language can be insecure. It is just a matter of how easy it is to make a mistake that compromises security. It is much, much easier to make such a mistake in C (buffer overflow, double free, wild pointer, null pointer, etc) than in Java. Therefore Java code is less likely to have security holes. It is that simple. Using C for all your secure code just because it is possible to write insecure code in Java is stupid.
"The language is secure so I don't need to audit"
Who said that? Did I say that? No, I did not say that. I said you spend less time auditing. Please don't put words in my mouth. Since there are (for example) no double free errors in Java, you don't have to check for them. Therefore you spend less time auditing. You still have to audit for errors that still can occur in Java. Duh.
I'm done talking to you.
Classic "uh-oh, I'm losing the argument" cop-out. Oh well.
Java isn't any less "free" than C even though it is a much more secure language. It is just as turing-complete as C is. If you use a language with security features built in you spend less time auditing and more time being productive. As we all know, programmer time is more expensive than CPU time, so a little less efficiency is excusable. Even so, the technology behind new languages is maturing to the point where it's hard to tell the difference anymore.
C was never designed to be secure. Blame C programmers for insecure code, not the language.
There is enough blame for everyone. When a C programmer makes a mistake, he deserves some of the blame. But if the language is inherently insecure, it is also partly the fault of the language. The way you're talking, we should just all use C for our secure code and punish programmers when they make mistakes because it's their fault. Instead you can realize that part of the fault is with the language and replace the language. Even though C wasn't designed to be secure you can still blame it for security related problems.
In the end, the semantics of the word "blame" aren't important. The important thing is that people realize that C programmers make security mistakes because C is an insecure language, and don't use C in security-critical areas, or review security-related C code extremely thoroughly before release.
Well I guess that makes all VB coders geniuses then, since they never get buffer overflows. They must just be the greatest bunch of programmers around.
Seriously, that's a stupid attitude to have. When you're talking about security, you have to assume the worst from the start. That means assuming coders will make mistakes, no matter how good they are. Even the best coders do make mistakes. In a language designed to be secure, the impact of coding mistakes is minimized.
It's very simple. Congress makes the laws. They can do whatever the heck they want, provided they don't violate the constitution or laws that they previously enacted. The Supreme Court doesn't have the authority to say "This law is stupid and bad, so we strike it down." The only thing they can rule on is whether it violates the Constitution or some other law. They ruled that it didn't. They are not endorsing the extension of copyright. Here is a quote from a justice explaining this: "We are not at liberty to second-guess Congressional determinations and policy judgments of this order, however debatable or arguably unwise they may be."
Lessig's arguments, when I read them, sounded like they were reaching pretty far. He was basically trying to give the Court a way it could rule that this monumentally stupid law was, in fact, unconstitutional. Unfortunately, he didn't quite succeed.
Two-thirds of any unclaimed settlements, up to $1.1 billion, will be given to California's neediest public schools to be used for computer equipment and related services in a program to be administered by the state's Department of Education. Microsoft would keep the final one-third of the unclaimed portion.
You have to look for a report on this by a California newspaper. It will have much more detail about the exact terms, since CA residents are the ones who are getting the vouchers. From the San Jose Mercury News:
Under the terms announced today, Microsoft will issue vouchers, ranging in value from $5 to $29 per licensed product, to consumers and businesses. The vouchers can be used to buy computers and software from any maker, including Microsoft rivals.
For claims totaling $100 or less, no documentation of purchases is needed, they said.
Plaintiffs' lawyers estimate consumers will recoup about one-third of what they spent for Microsoft products, including the Windows operating systems and popular Office and Word programs, between Feb. 18, 1995, and Dec. 15, 2001.
For example, consumers will receive a voucher for $29 for each copy of Microsoft Office purchased, $26 for each copy of Excel, $16 for each copy of the Windows operating system and $5 for each copy of the word-processing program Word, according to plaintiffs' lawyers.
The only requirement for a solution to be "enough", IMHO, is that it prevents Microsoft from continuing to practice anti-competitive monopolistic behavior. As we have seen, they won't stop on their own. The fact that they exist pretty much makes it inevitable that they will behave in an anti-competitive manner. The only real solution to the problem that I see is a split, but that has problems of its own.
This settlement is a joke. 1/40th of Microsoft's cash reserves? It may be a lot of money, but they've got money out the wazoo. The lawyers get rich, the consumer gets a $20 voucher, whoopee. Not to mention that most of the voucher money will probably go unclaimed, and the claimed part will largely be spent in a manner that only strengthens Windows's hold on the market (Office, Windows software, Windows itself, Windows PCs, Windows-specific hardware). And there is no mechanism to ensure they won't turn around and do the same anti-competitive things tomorrow. The settlement is the equivelant of a verbal warning. "Bad Microsoft! Don't do that again!"
The problem with monopolies is that they are huge and slow-moving. Yeah, a telco monopoly might be efficient for now, but what happens when the technology improves and changes? The lumbering monopoly will be left in the dust by a healthy network of competitive companies struggling to adapt and survive. In the computer world, things change and improve so fast that IMHO it is more efficient in the long run to have a slightly less efficient but adaptable capitalist solution than a monopolistic one. By the time the monopoly has learned how to be efficient using yesterday's technology, tomorrow's technology will be here and it will be twice as efficient anyway.
Without support for mainstream media (WiMP, QT, Flash 6, Real), Microsoft Office, and DirectX (negotiable, but witness how many games use the "industry standard" OpenGL), Linux can't even get a seat to watch the game, let alone actually play.
MPlayer is awesome. I guarantee it can play every single video file you have on your computer right now, and every one you're likely to come across surfing the web. When did you last try it? Yes, it supports Sorenson now. And WMV. And Real. And DivX. All out-of-the-box, all in one player, with no DRM, auto-updaters, horrible licenses, or advertisements.
Star/OpenOffice opens Microsoft Office documents perfectly. Other open-source office efforts are leveraging this code to produce their own document filters.
Cutting-edge games are still a problem. But there are lots of people who don't play games on their computers other than Solitaire and Space Cadet Pinball. And computer games are becoming less relevant as consoles become more and more powerful.
I'd say the Linux desktop's time is near. The pieces are falling into place.
The biggest problem with folders is no one wants to be a file clerk and weed, sort, and file their docs.
That's not true. Often, people with a fairly high level of computer experience start compulsively organizing their stuff into nice neat folders when they're bored. It is almost theraputic, a place for everything and everything in its place. The problem is that most people don't really "get" the whole filesystem metaphor and don't ever get the idea that they can change it themselves to organize their files. It is almost scary how few people really know how the filesystem works. If the "Open" dialog doesn't start up in the usual folder, they're lost. Creating new folders and moving files are black computer voodoo magic.
Slashdot Mirror
That's an understatement if I've ever heard one. The thing is fricking GINORMOUS! 9MB for "Hello, World"? 900MB for a UML application?!??!?!
Personally I think it would be really nice to carry around a "cash card" instead of actual bills/coins if all the above conditions were met. Smaller, easier to handle, and less smelly. If e-ink-type-stuff were added to the card to show its value at all times, then it would be a no-brainer. Of course, people older than about 35 would hate the new system "just because", so the transition couldn't be complete until they all are too old to have any influence.
"Chess Masters Hold Off Computer Programmers... For Now"
That is the vision of GStreamer. Progress seems fairly slow, but steady so far. I think GStreamer is even distributed with Gnome now. Hopefully GStreamer will soon mature to the point where gst-player can replace mplayer/xine as a general media player.
Freetype's font auto-hinter does this. So if you're using anti-aliased fonts on XFree86, you are seeing this idea in action (unless you're using the truetype font hints in violation of Apple's patents, but they do much the same thing anyway). I don't know if any SVG implementation does this yet.
They've probably never heard of mplayer. After all, it was only fairly recently that Sorensen became available through mplayer. It does indeed play the trailer (unfortunately with some sound lag on my machine though). I don't think Apple wants to publicly endorse mplayer, since it is kind of a hack that required reverse-engineering the interface quicktime uses to call quicktime codecs on Windows. Most companies are strangely reluctant to endorse hacks, even hacks that work well.
Once again, for a Linux player that can handle this Quicktime Sorensen file, look no further than the superlative mplayer. It has had the capability to play quicktime movie trailers for a while now, but I figure some people still haven't heard. Just make sure you have version 0.90 rc1 or better, because that's when they reverse-engineered the windows quicktime codecs.
Today has been plenty fast for me. I haven't noticed any problems yet (well, other than not being able to register Visual Studio .NET, is there a crack out there?). The internet really is quite resilient.
But you're not sacrificing security. If you spend less time auditing code that has much fewer security holes, you get the same level of security in the end. You can never guarantee perfect security, only extremely good security. If the language gives you a head start, it takes less time to achieve the same level of security.
Pairing alloc/free's in properly designed code takes all of about a tenth of a second.
Oh, right. And next you'll tell me that loop constructs are for lazy programmers because constructing them with gotos only takes "about a tenth of a second". Give me a break. For one thing, it takes longer than that in any reasonably complex project. For another thing, humans will occasionally make a mistake. Poof, security hole. Compilers don't make those kind of mistakes. Why not let the compiler do it, both for coding speed and security reasons?
The numerical range of values that can be stored in an int or float isn't really the issue. You can simply scale the values to get any range you desire, and the mapping of actual intensities on the range is arbitrary anyway. The post above explains that the reason they use floats is because the amount of precision changes logarithmically over the range of values and the percieved brightness of light is logarithmic with the intensity, which makes more sense.
Interesting. Thanks for the informative explanation!
I also don't see the difference you are talking about with the "intensity" values as opposed to color. If the 8-bit 'r' value in a 24-bit pixel isn't measuring the intensity of red light at that pixel, then what is it measuring?
I guess my basic question is: Why is this so special? Why couldn't someone, for example, extend PNG to have 16 bits per channel (if it doesn't already) and forget about OpenEXR?
"The language is secure so I don't need to audit"
Who said that? Did I say that? No, I did not say that. I said you spend less time auditing. Please don't put words in my mouth. Since there are (for example) no double free errors in Java, you don't have to check for them. Therefore you spend less time auditing. You still have to audit for errors that still can occur in Java. Duh.
I'm done talking to you.
Classic "uh-oh, I'm losing the argument" cop-out. Oh well.
Java isn't any less "free" than C even though it is a much more secure language. It is just as turing-complete as C is. If you use a language with security features built in you spend less time auditing and more time being productive. As we all know, programmer time is more expensive than CPU time, so a little less efficiency is excusable. Even so, the technology behind new languages is maturing to the point where it's hard to tell the difference anymore.
There is enough blame for everyone. When a C programmer makes a mistake, he deserves some of the blame. But if the language is inherently insecure, it is also partly the fault of the language. The way you're talking, we should just all use C for our secure code and punish programmers when they make mistakes because it's their fault. Instead you can realize that part of the fault is with the language and replace the language. Even though C wasn't designed to be secure you can still blame it for security related problems.
In the end, the semantics of the word "blame" aren't important. The important thing is that people realize that C programmers make security mistakes because C is an insecure language, and don't use C in security-critical areas, or review security-related C code extremely thoroughly before release.
Well I guess that makes all VB coders geniuses then, since they never get buffer overflows. They must just be the greatest bunch of programmers around.
Seriously, that's a stupid attitude to have. When you're talking about security, you have to assume the worst from the start. That means assuming coders will make mistakes, no matter how good they are. Even the best coders do make mistakes. In a language designed to be secure, the impact of coding mistakes is minimized.
You're in luck, it does indeed have AAlib output.
Lessig's arguments, when I read them, sounded like they were reaching pretty far. He was basically trying to give the Court a way it could rule that this monumentally stupid law was, in fact, unconstitutional. Unfortunately, he didn't quite succeed.
And a link to the article
The only requirement for a solution to be "enough", IMHO, is that it prevents Microsoft from continuing to practice anti-competitive monopolistic behavior. As we have seen, they won't stop on their own. The fact that they exist pretty much makes it inevitable that they will behave in an anti-competitive manner. The only real solution to the problem that I see is a split, but that has problems of its own.
This settlement is a joke. 1/40th of Microsoft's cash reserves? It may be a lot of money, but they've got money out the wazoo. The lawyers get rich, the consumer gets a $20 voucher, whoopee. Not to mention that most of the voucher money will probably go unclaimed, and the claimed part will largely be spent in a manner that only strengthens Windows's hold on the market (Office, Windows software, Windows itself, Windows PCs, Windows-specific hardware). And there is no mechanism to ensure they won't turn around and do the same anti-competitive things tomorrow. The settlement is the equivelant of a verbal warning. "Bad Microsoft! Don't do that again!"
The problem with monopolies is that they are huge and slow-moving. Yeah, a telco monopoly might be efficient for now, but what happens when the technology improves and changes? The lumbering monopoly will be left in the dust by a healthy network of competitive companies struggling to adapt and survive. In the computer world, things change and improve so fast that IMHO it is more efficient in the long run to have a slightly less efficient but adaptable capitalist solution than a monopolistic one. By the time the monopoly has learned how to be efficient using yesterday's technology, tomorrow's technology will be here and it will be twice as efficient anyway.
MPlayer is awesome. I guarantee it can play every single video file you have on your computer right now, and every one you're likely to come across surfing the web. When did you last try it? Yes, it supports Sorenson now. And WMV. And Real. And DivX. All out-of-the-box, all in one player, with no DRM, auto-updaters, horrible licenses, or advertisements.
Star/OpenOffice opens Microsoft Office documents perfectly. Other open-source office efforts are leveraging this code to produce their own document filters.
Cutting-edge games are still a problem. But there are lots of people who don't play games on their computers other than Solitaire and Space Cadet Pinball. And computer games are becoming less relevant as consoles become more and more powerful.
I'd say the Linux desktop's time is near. The pieces are falling into place.
That's not true. Often, people with a fairly high level of computer experience start compulsively organizing their stuff into nice neat folders when they're bored. It is almost theraputic, a place for everything and everything in its place. The problem is that most people don't really "get" the whole filesystem metaphor and don't ever get the idea that they can change it themselves to organize their files. It is almost scary how few people really know how the filesystem works. If the "Open" dialog doesn't start up in the usual folder, they're lost. Creating new folders and moving files are black computer voodoo magic.