How weird. I turn 29 this month and I just moved back into to my parents' house.
Okay, okay, I, my wife and two children are house-sitting for them. Still, I guess there's hope for a new career in the unauthorised software industry.
MNG, the multiple-image extension to PNG, does this quite well.
MNG isn't really designed for multiple images with different resolutions (e.g. mipmap textures) or different colour depths (e.g. colour-corrected images with separate alpha). I think you can hack the former by choosing a frame size which is as big as the largest image, but it's not a good fit.
Tiled images:
Do you mean the tile-graphics images used by 2D video game consoles, or do you refer to a sense of "tiled images"
that doesn't match the sense that comes to mind when I see the term "tiled images"?
I mean dicing the image into rectangles and specifying each separately to allow caching of large images. The rectangles may have differing sizes, colour depths and so on.
By "device resolution" do you mean "number of pixels per physical inch"?
That's one aspect, but bear in mind that the "pixel" is necessarily a concept which makes sense to all output devices. In many printers, for example, the distance between two "dots" is different than the physical size of a "dot", which requires two different measures of device resolving power, each of which could reasonably be called "dots per inch".
White and black points.
Are white and black points relevant to selling images online and having them display in the browser, as the other user mentioned?
That depends on what kind of images you're selling. Calibrated film "negatives" need them, for example, and I can imagine those being sold.
PNG supports CMYK data, and its (admittedly limited) color correction system specifies XYZ coordinates.
I stand corrected.
What is "floating point image data"? PNG supports up to 16 bits per channel.
I mean IEEE floating point channels. They're not quite so critical for high dynamic range data if you have 16 bits per channel and white and black points (that way you can assign, say, 0 to reference black, 1024 to reference white and use the area 1025-65536 as your headroom) or some other appropriate non-integer type (e.g. rational pixels, Radiance-style "real pixels" and so on). PNG also doesn't directly support signed pixels, but I guess you could just interpret a 16 bit quantity as a signed number.
One other thing I forgot to mention is storing an arbitrary number of channels per image, which is important for some applications.
If anything, I'd say that an absolute blank on the disk is closer to Cage's original intent than a recording of a live
performance. It forces the listener to strain his ears trying to figure out what's going on, resulting in him listening to
ambient sounds.
Interesting thought. I think it depends on how you listen to the CD. If you're concentrating on something else (say, you put the headphones on while you're coding), you don't strain your ears. At a live concert, you have a more captive and focussed audience.
Of course it would be a strange person indeed who used Cage as background music. But then, I occasionally put on THRaKaTTaK while coding, so who am I to judge?
John Cage's piece, 4'33", was actually very clever and quite a novel idea for its time.
One of the themes of his work is to let sounds be themselves. To that end, he composed a piece which involved a pianist holding his hands over a piano keyboard for 4 minutes and 33 seconds. The music was not silence, but rather the sound of the audience slowly realising to what was going on.
As such, this piece can never really be recorded (unless you actually record an audience listening to it, and even then, it's not the same thing; once the sound is recorded, it is no longer the same kind of performance), and claiming that a recording of silence is even close to being the same thing as 4'33" is ludicrous.
Mike Batt's problem is crediting Cage on the album. Yes, he did it for a laugh, but by doing so, did he inadvertantly claim legal liability?
Personally, I think John Cage would have gotten a real kick out of the whole proceedings. It would have appealed to his sense of whimsy.
Multiple images in the one file. (For example, multiresolution images.)
Tiled images.
Device resolution (not just pixel resolution).
White and black points.
Colour spaces other than RGB (e.g. XYZ, La*b*, CMYK).
Floating point image data.
Basically, PNG is unsuitable for pretty much everything except losslessly transferring images compactly over a network (which it does extremely well; arguably much better than TIFF).
Renderings per frame may
drop to minutes instead of hours, but it probably won't be interactive.
You have no idea how wrong that statement is. The first half, that is. The second half is perfectly accurate.
People in the visual effects/animation business have something they call "Blinn's Law", which is the flip-side of Moore's Law. It states: Renders will always take the same amount of time. It's true. On average, computing frames for Monsters, Inc took about the same time as they did for Luxo, Jr. The reason why this is the case is that audience's expectations increase at about the same rate as the power of hardware. Yes, eventually we may well have prman in your graphics card, but by then, the CG films of the turn of the century will look quaint at best.
You'd be surprised though, by how tolerant of other views some religious people can be, and likewise, how intolerant some atheists can be.
Doesn't surprise me in the least. People are people. Presence or absence of religion doesn't change that.
My inner armchair psychologist thinks that many "theists" are exclusive (i.e. exclude other religions and non-religions by definition) and many "atheists" are "anti" (i.e. define their position by being "not theists", thus setting up an irrevocably contrary position from the start). This makes the whole field of religion in particular a breeding ground for intolerance no matter what side you're considering.
Disclaimer: By "theist" and "atheist" here, I mean something specific, namely, someone who bases their identity, or perhaps the wrong part of their identity, around the concept, rather than someone who just happens to believe or not believe in one or more deities. Used in this sense, Madeleine Murray O'Hare was an atheist, but Alan Turing was someone who just happened not to believe in any deities. I hope that readers will forgive the abuse of terminology, but I hope you take my point.
The emphasis in all this is on the word "some". As in all controversies, it's the most extreme elements who are the loudest. Some people are very tolerant. Some people are very intolerant. Most don't care, or at the very least, it never becomes an issue.
I've personally found that geeks are some of the most tolerant people you'll find. I think it's something to do with having spent an entire childhood getting hell for being different. It makes you more likely to tolerate difference in others, so long as you are different for the right reasons. Sheep are not tolerable, but non-sheep can be found in all creeds and we respect that. That way we can get on with the serious flaming over issues that are truly important, like which software licence is best, or which operating system kernel will take over the world.
The law in question only covers sites hosted by ISPs in Australia. Google isn't affected by this law, just like if the Australian content provider shifted their material to a server located in Sweden.
One of the reasons why the EFA would like a list of the banned sites issued is so they can construct a list of where the sites have moved to, thus showing that the only real effect of the censorship law is to move business overseas.
Some have mentioned the dragon book (Aho, Ullman et all) and the tiger book (Appel's "Modern Compiler Implementation"). Both excellent references.
If you're at all mathematically minded, you should look into a good book on formal languages, such as Harrison's "Introduction to Formal Language Theory", or R. Nigel Horspool's "LR Parsing". (Caution: Don't buy these. Look for them in your local university library.)
Once you have the basics, you should look into the work of Mark Hopkins. The guy is a genius, and understands more of the theory of parsing than anyone I've ever seen, and has invented some techniques which far surpass the "state of the art" of lex and yacc.
To answer your question: You have always had fair use rights. Plus, thanks to the DMCA[1], you now have proxying and caching rights. You have never had republishing rights except where explicitly granted. Indeed, on most corporate web sites that I've seen, republishing is explicitly disallowed in the relevant corporate disclaimers.
IANAL either, but lawyers and courts are usually not impressed by the "slippery slope" arguments that we geeks (I do include myself) usually come up with in these situations.
[1] You heard that right, by the way. If it weren't for the constitution-violating Chapter 12, the DMCA would actually be a pretty good law. It has some lovely shiny new rights for Americans to enjoy. This is one of them.
If you're straight, it's actually not as hard as you might think to find a girl geek. Seriously.
The secret is to think more broadly. Your girl geek may not be a coder, an open source zealot or the like, but remember that geeks come in all shapes and sizes. Mine is a graphic designer, for example. Despite being more at home in Photoshop, she can also do sysadmin duties like editing DNS zone files.
She asked for broadband for her birthday. *giddy sigh*
Even if that were true, you're missing the point. We still have FreeDOS.
NT isn't actually too badly designed as OS designs go. Most of the problems with developing under NT actually have little to do with the kernel proper (e.g. OLE2, DLL hell etc).
I think the original poster confused two different techniques to accomplish the same ends.
In the following discussion, let M be the number of people who get partial keys and K be the number of partial keys required to reconstruct the secret key.
The first technique (I forget who came up with it) was to consider a key as a point in K-dimensional space. You randomly generate M hyperplanes of dimension (K-1), each of which contain the secret key, and give one to each person. When M partial keys are presented, you have enough information to find the secret key by solving the linear equations.
The problem with this technique is that if you have some partial keys but not enough to generate the secret key, you still have some information which could be used to speed up a brute-force attack. Shamir's technique is superior in this respect, since having some partial keys gives you no information about the secret key, all other things being equal (i.e. assuming the fake coefficients were chosen truly randomly). As you said, "it could be anywhere".
This actually makes Shamir's secret sharing algorithm one of the very few provably secure cryptographic algorithms in existence, which is quite a remarkable thing when you think about it. Of course that doesn't automatically make your protocol or your secure, but you can't have everything.
Design patterns are nothing more and nothing less than a certain kind of engineering experience. If you're only writing trivial or small programs, you don't need them. If you're developing flexible medium- to large-sized applications or systems, the rules are different.
There are a number of differences between programming and software engineering. This has been one of them.
There is one exception, and that's commerce. If you do business with someone in another jurisdiction, you are subject to the commerce laws in that jurisdiction.
Of course, this is taken to stupid extremes, like the California court which was convinced that Matthew Pavlovich could be tried there in the DeCSS case because a lot of movies are made there and a lot of computers are made there, and so anything which involves computers and movies are fair game.
It's not like C++ can't interface with C libraries. ..
It's often pretty painful, though. Off the top of my head:
Complex C APIs, especially GUI systems (e.g. Win32, Motif) often try to implement an object model themselves, which usually doesn't map well onto C++'s object model.
APIs with callbacks are almost never exception safe.
Varargs at the C level usually implies const incorrectness at the C++ level.
Don't get me started on concurrency, though at least modern C APIs tend to get this right.
Memory management/ownership issues occasionally crop up, though this is a bit more rare in a well-written C API.
That doesn't work with, say, movies. At least the producer, the director, the screenwriter, the editor, the composer, the production designer, the costume designer and possibly even the actor might all have some claim on the copyright of a certain shot in a film. Securing rights to use that scene, and ensuring appropriate copyright attributions, would be nothing less than a nightmare.
I think the original comment ignores the fact that nowadays, creativity is almost always a team effort.
Slashdot Mirror
How weird. I turn 29 this month and I just moved back into to my parents' house.
Okay, okay, I, my wife and two children are house-sitting for them. Still, I guess there's hope for a new career in the unauthorised software industry.
pH33r mY sK1lLz!!!!!!1!!!!!
Personally, I've never understood why exiting a program is a "File" operation, either.
Disclaimer: I hear what you're saying. Raytracing with POV-Ray is a fun passtime. However...
Only in bad raytracers. Good rendered images (raytraced or otherwise) don't look like raytraced images unless the user specifically wants them to.
No arguing over taste, of course, but I find raytraced images that look raytraced distracting as hell.
On multiple images in one file:
MNG isn't really designed for multiple images with different resolutions (e.g. mipmap textures) or different colour depths (e.g. colour-corrected images with separate alpha). I think you can hack the former by choosing a frame size which is as big as the largest image, but it's not a good fit.
Tiled images:
I mean dicing the image into rectangles and specifying each separately to allow caching of large images. The rectangles may have differing sizes, colour depths and so on.
That's one aspect, but bear in mind that the "pixel" is necessarily a concept which makes sense to all output devices. In many printers, for example, the distance between two "dots" is different than the physical size of a "dot", which requires two different measures of device resolving power, each of which could reasonably be called "dots per inch".
White and black points.
That depends on what kind of images you're selling. Calibrated film "negatives" need them, for example, and I can imagine those being sold.
I stand corrected.
I mean IEEE floating point channels. They're not quite so critical for high dynamic range data if you have 16 bits per channel and white and black points (that way you can assign, say, 0 to reference black, 1024 to reference white and use the area 1025-65536 as your headroom) or some other appropriate non-integer type (e.g. rational pixels, Radiance-style "real pixels" and so on). PNG also doesn't directly support signed pixels, but I guess you could just interpret a 16 bit quantity as a signed number.
One other thing I forgot to mention is storing an arbitrary number of channels per image, which is important for some applications.
There's a fine line between "troll", "hacker", "deluded self-important kook" and "avant garde genius".
Take your pick. Whichever one you choose, "extremely clever" still seems entirely appropriate.
Interesting thought. I think it depends on how you listen to the CD. If you're concentrating on something else (say, you put the headphones on while you're coding), you don't strain your ears. At a live concert, you have a more captive and focussed audience.
Of course it would be a strange person indeed who used Cage as background music. But then, I occasionally put on THRaKaTTaK while coding, so who am I to judge?
John Cage's piece, 4'33", was actually very clever and quite a novel idea for its time.
One of the themes of his work is to let sounds be themselves. To that end, he composed a piece which involved a pianist holding his hands over a piano keyboard for 4 minutes and 33 seconds. The music was not silence, but rather the sound of the audience slowly realising to what was going on.
As such, this piece can never really be recorded (unless you actually record an audience listening to it, and even then, it's not the same thing; once the sound is recorded, it is no longer the same kind of performance), and claiming that a recording of silence is even close to being the same thing as 4'33" is ludicrous.
Mike Batt's problem is crediting Cage on the album. Yes, he did it for a laugh, but by doing so, did he inadvertantly claim legal liability?
Personally, I think John Cage would have gotten a real kick out of the whole proceedings. It would have appealed to his sense of whimsy.
Let's see...
Basically, PNG is unsuitable for pretty much everything except losslessly transferring images compactly over a network (which it does extremely well; arguably much better than TIFF).
You have no idea how wrong that statement is. The first half, that is. The second half is perfectly accurate.
People in the visual effects/animation business have something they call "Blinn's Law", which is the flip-side of Moore's Law. It states: Renders will always take the same amount of time. It's true. On average, computing frames for Monsters, Inc took about the same time as they did for Luxo, Jr. The reason why this is the case is that audience's expectations increase at about the same rate as the power of hardware. Yes, eventually we may well have prman in your graphics card, but by then, the CG films of the turn of the century will look quaint at best.
Hell, Toy Story already looks quaint.
Doesn't surprise me in the least. People are people. Presence or absence of religion doesn't change that.
My inner armchair psychologist thinks that many "theists" are exclusive (i.e. exclude other religions and non-religions by definition) and many "atheists" are "anti" (i.e. define their position by being "not theists", thus setting up an irrevocably contrary position from the start). This makes the whole field of religion in particular a breeding ground for intolerance no matter what side you're considering.
Disclaimer: By "theist" and "atheist" here, I mean something specific, namely, someone who bases their identity, or perhaps the wrong part of their identity, around the concept, rather than someone who just happens to believe or not believe in one or more deities. Used in this sense, Madeleine Murray O'Hare was an atheist, but Alan Turing was someone who just happened not to believe in any deities. I hope that readers will forgive the abuse of terminology, but I hope you take my point.
The emphasis in all this is on the word "some". As in all controversies, it's the most extreme elements who are the loudest. Some people are very tolerant. Some people are very intolerant. Most don't care, or at the very least, it never becomes an issue.
I've personally found that geeks are some of the most tolerant people you'll find. I think it's something to do with having spent an entire childhood getting hell for being different. It makes you more likely to tolerate difference in others, so long as you are different for the right reasons. Sheep are not tolerable, but non-sheep can be found in all creeds and we respect that. That way we can get on with the serious flaming over issues that are truly important, like which software licence is best, or which operating system kernel will take over the world.
I'm waiting for the day when someone brings a lawsuit on the grounds that they worship neither the flag nor the republic for which it stands.
As a matter of interest, do non-US-citizens who attend US public schools have to recite the pledge?
The law in question only covers sites hosted by ISPs in Australia. Google isn't affected by this law, just like if the Australian content provider shifted their material to a server located in Sweden.
One of the reasons why the EFA would like a list of the banned sites issued is so they can construct a list of where the sites have moved to, thus showing that the only real effect of the censorship law is to move business overseas.
Some have mentioned the dragon book (Aho, Ullman et all) and the tiger book (Appel's "Modern Compiler Implementation"). Both excellent references.
If you're at all mathematically minded, you should look into a good book on formal languages, such as Harrison's "Introduction to Formal Language Theory", or R. Nigel Horspool's "LR Parsing". (Caution: Don't buy these. Look for them in your local university library.)
Once you have the basics, you should look into the work of Mark Hopkins. The guy is a genius, and understands more of the theory of parsing than anyone I've ever seen, and has invented some techniques which far surpass the "state of the art" of lex and yacc.
To answer your question: You have always had fair use rights. Plus, thanks to the DMCA[1], you now have proxying and caching rights. You have never had republishing rights except where explicitly granted. Indeed, on most corporate web sites that I've seen, republishing is explicitly disallowed in the relevant corporate disclaimers.
IANAL either, but lawyers and courts are usually not impressed by the "slippery slope" arguments that we geeks (I do include myself) usually come up with in these situations.
[1] You heard that right, by the way. If it weren't for the constitution-violating Chapter 12, the DMCA would actually be a pretty good law. It has some lovely shiny new rights for Americans to enjoy. This is one of them.
If you're straight, it's actually not as hard as you might think to find a girl geek. Seriously.
The secret is to think more broadly. Your girl geek may not be a coder, an open source zealot or the like, but remember that geeks come in all shapes and sizes. Mine is a graphic designer, for example. Despite being more at home in Photoshop, she can also do sysadmin duties like editing DNS zone files.
She asked for broadband for her birthday. *giddy sigh*
Even if that were true, you're missing the point. We still have FreeDOS.
NT isn't actually too badly designed as OS designs go. Most of the problems with developing under NT actually have little to do with the kernel proper (e.g. OLE2, DLL hell etc).
I think the original poster confused two different techniques to accomplish the same ends.
In the following discussion, let M be the number of people who get partial keys and K be the number of partial keys required to reconstruct the secret key.
The first technique (I forget who came up with it) was to consider a key as a point in K-dimensional space. You randomly generate M hyperplanes of dimension (K-1), each of which contain the secret key, and give one to each person. When M partial keys are presented, you have enough information to find the secret key by solving the linear equations.
The problem with this technique is that if you have some partial keys but not enough to generate the secret key, you still have some information which could be used to speed up a brute-force attack. Shamir's technique is superior in this respect, since having some partial keys gives you no information about the secret key, all other things being equal (i.e. assuming the fake coefficients were chosen truly randomly). As you said, "it could be anywhere".
This actually makes Shamir's secret sharing algorithm one of the very few provably secure cryptographic algorithms in existence, which is quite a remarkable thing when you think about it. Of course that doesn't automatically make your protocol or your secure, but you can't have everything.
It's the same with most film formats, too. The image is expanded horizontally by an anamorphic lens inside the camera.
The one exception was vistavision, but nobody really cares about that any more.
Design patterns are nothing more and nothing less than a certain kind of engineering experience. If you're only writing trivial or small programs, you don't need them. If you're developing flexible medium- to large-sized applications or systems, the rules are different.
There are a number of differences between programming and software engineering. This has been one of them.
There is one exception, and that's commerce. If you do business with someone in another jurisdiction, you are subject to the commerce laws in that jurisdiction.
Of course, this is taken to stupid extremes, like the California court which was convinced that Matthew Pavlovich could be tried there in the DeCSS case because a lot of movies are made there and a lot of computers are made there, and so anything which involves computers and movies are fair game.
Don't forget a certain President of Panama, still being held (technically as a prisoner of war, I believe).
Could someone who has downloaded the song please post a brief review? What kind of music is it? Similar to the works of what other artists?
I know that US$1 isn't much, but I'm not going to spend that on every artist who chooses to sell this way without at least some idea what I'm buying.
It's often pretty painful, though. Off the top of my head:
That doesn't work with, say, movies. At least the producer, the director, the screenwriter, the editor, the composer, the production designer, the costume designer and possibly even the actor might all have some claim on the copyright of a certain shot in a film. Securing rights to use that scene, and ensuring appropriate copyright attributions, would be nothing less than a nightmare.
I think the original comment ignores the fact that nowadays, creativity is almost always a team effort.
Let me guess... you've never seen The Producers , right?