I've enjoyed Mr. Ebert's commentary on movies for years, but in this matter I feel he's revealing a certain degree of "fuddy-duddy-ism".
By way of disclosure, I work for Pixar Animation Studios, but I do not speak for them.
By means of an informal poll, I have yet to meet a single person who saw Toy Story 2 on a digital projector and thought it looked worse than ANY print of the same movie. The lack of dust, film scratches, sprocket jitter and generational loss are CLEARLY apparent, and quite striking. To be fair, Toy Story was mastered directly from the original digital source, and we have some pretty talented engineers and DPs who worked to make that look good. I have not had the chance to view other films like The Phantom Menace in such a theater, so I won't say that digital projection makes for a uniformly better picture, but TS2 is a good technology demonstration that remarkable things are possible.
Right now digital projection systems are rather expensive: of course, there are only a couple of dozen theater ready ones in the world, so perhaps that's not too surprising. The interesting this is that virtually everything in a digital projector will benefit from Moore's law kind of price/performance drops. So it takes 100gb of disk to store a movie. The disks are getting cheaper, and what's remarkable: they are REUSEABLE. Ever price what a 35mm film print costs? Try multiplying those by the 3000 theaters that a major release might have, and you run into some serious money. Digital projectors have MUCH fewer moving parts than a conventional projector, because they have no film transport and no shutter.
Costs aside, it is my firm belief that digital theaters will deliver a better LOOKING picture as well. The resolution of current projectors (1280x1024) is thought to be 'low' but in fact, when done properly, they can look amazingly good. The typical film projector has several pixels worth of sprocket jitter, so the imagined higher resolution of film isn't usually delivered to the screen, also being eaten up in generational loss, aging of film pigments, scratches and dust. Some of you may recall the television special detailing the remastering of Star Wars. The original film required LARGE amounts of work to correct for the terrible aging of the pigments involved. Digital projection means that 30 or 300 years from now, you will be able to show the film EXACTLY as it appears today. That's a pretty strong selling point in my book.
As a brief aside addressing the choice of resolution and compression ratios, let me just point out that first of all, costs for disks is dropping, while the resolution that people want to view things at is pretty much fixed. This means that the entire issue of compression will become a moot point anyway. If you assume 2M pixels per frame, 24fps, 12 bits per color channel, and a 90 minute movie, the uncompressed data storage is about 1 TB. That's admittedly a lot of storage, but it isn't infinite, and its becoming easier to do as time goes on. Current (lossy) compression technology can easily cut that down by a factor of more than ten, even without using frame/frame coherence. Forget MPEG, state of the art leaves that stuff in the dust. 100gb disk arrays aren't particularly expensive, and next year they'll be half as expensive as they are this year. In three to five years, we probably wouldn't even need to compress at all.
I saw Ebert on television claim that when he went to a theater, he wants to see light coming through celluloid. Well, he's entitled to watch whatever he likes however he likes, but I think I've seen the future, and it isn't far off. I'm waiting eagerly for the next installment of Star Wars. I suspect it will turn a few heads.
Hmmmm... I guess you'll have to write a compiler then. And how do you do that? Use assembly language
Let me guess, you don't write compilers for a living....
At the risk of making an overly general statement, NOBODY does what you describe anymore. A compiler is one of the more perfect examples of something that shouldn't be written in assembly language. It embodies some fairly complex algorithms but is itself not very speed critical. Most people don't care about how fast their compiler outputs code, merely how fast the resulting code is when run.
The two most available open source C compilers (gcc and lcc) take a similar approach by defining new machines via a rather compact machine description file. This allows the compiler to properly abstract the portions of the compiler which aren't machine dependant (the input parser, symbol table construction, parse tree assembly and the like) from the parts that are (code generation). The result is a compiler which is easier to port to new architectures, and has less errors.
Most of these compilers similarly use assemblers written in C. Initially these are cross-assemblers, but once the assembler is compiled by your new compiler, you can run the assembler in native mode. *voila*
To write machine descriptions for these compilers, you of course _do_ need to know something about the assembly language of the target machine, but this is vastly easier than writing the entire compiler in assembly.
Traditional PCs have a number of advantages, but also a number of glaring problems.
Undoutably one of their key advantages has been the ability to update and customize configurations dependant on customer needs.
This is also one of their key disadvantages. Try just looking at a PC sometimes. It's a virtual hornet's nest of cables and wires, most of which just serve to carry signals from one part of the computer to another. How much of the total complexity (read cost) of a PC is just to support this kind of legacy interconnects?
It seems rather obvious that most people in the world probably don't care that much about all the flexibility that a PC offers. They'd like to buy an inexpensive machine to access the net and play a few games. An internet appliance if you like. For the most part, the traditional "assemble-it-yourself" computer gets in the way of that.
This is classic "wooly thinking". First of all, where do you get this 99.9% figure? I say that 100% of all UFO sightings are fakes, crocks, and misidentified. And until a UFO comes down in the middle of Times Square, squashes a couple of SUVs, and purple haired maidens with six breasts emerge and asked to be taken to our leader, you can't prove that I am not correct.
I find it amazing that people can be so far separated from common sense and the principle of Occam's Razor that they can't see through crap like this. We have an anectdotal story about some young anonymous Mexican girl recovering a skull, at some unspecified location and time in the past. We have assertions (but curiously no pictures) that the skull shows bizarre deformities. We have assertions from that it has been viewed by a number of "experts", who of course couldn't "risk the ridicule of their scientific peers".
Uh huh.
Okay, so far it isn't so bad. But what really gets me is the fanciful leap into explanations. It "must" be a Gray, or a Gray-human hybrid. Uh huh. Why? Are their other possibilities. Say, well, FRAUD for one? Given that no proof that Grays exist, isn't it just a teensy bit premature to assign a skeleton to their particular family tree? Then we get into this crap about DNA testing. As several other posters noted, there is little reason to believe that DNA has any particular magical properties that make it the only way (or the most likely way, or even a way that has occurred more than once) for life to develop in the universe. Since no copy of Gray DNA has ever been recovered, it seems hard to know how DNA testing would ever result in an appropriate classification.
This is bunkery of the highest order. It tries to take on the trappings of science, but it is just a particularly odd form of creationism, with all its attendant perils.
By the way, if there is a.1% chance for any given report to be true, then the chances are 99.8% that one or more of them are true in a year (given 6000 sightings). It doesn't take a 30 page dissertation to work it out.
We are getting a bit astray from the topic of DVD encryption, but I'll merely update with a couple of specific comments on your musings, and then let this drop back to real discussions on the real topic at hand.
The film industry has invested large amounts of money in developing properties which have a rather sizeable commercial value. They wish to protect their investment by limiting the ordinary individuals ability to partake of their "experience" to those who have paid for the ticket, cd, video tape, or dvd.
There is only one problem with that, and it is a problem that is shared with most media: in order to show a movie or play a cd, it is necessary to read the information off a disk and convert it into a representation which people can watch or listen to. And if a person can listen to or watch a particular representation, then so can some form of machine, and therefore, copies can be made.
The film industry has so far used dvds because the decoders were mostly encoded in physical devices which would be difficult and/or expensive for the average joe to duplicate. But once software decoders became available, whether they encrypted keys or not, it seems like they let the cat out of the bag, because software IS cheap to duplicate and distribute.
Of course, the open source vs. trade secret arguments have nothing to do with any of this, which was my point originally. This point was not lost on a number of other people who commented similarly on your original posting.
The rest of your posting (ray tracing, radiosity and the Free Film project) are fascinating topics in themselves, but not really apropos to the topic at hand.
What is the point you are trying to make regarding Open Source?
You claim that the film industry has more to fear by not adopting Open Source than by embracing it.
You make the claim that an open source ideology reduces distribution costs because the technology required to mass produce is being developed faster. I'm not sure what one has to do with the other. Just how has open source influenced the distribution costs (ignoring the obvious "Gee, I can download The Phantom Menace off this website for nothing") CDs and DVDs made an enormous impact because they are cheap to reproduce and give great fidelity, not because of any revolution caused by Open Source.
As for piracy protection, the problem with the DVD encryption as I understand it wasn't caused by a lack of proper encryption algorithms, but rather by the fact that some doofus shipped the keys in an unprotected form with an application. You can make things foolproof I guess, but you can't make them damn-fool proof. Perhaps some form of technical review should have been done before the offenders shipped their app, but again, your points seems to miss the mark. You can't write an open source DVD decryptor, because one key components of the system is a secret key that the film industry doesn't want people to have because it amounts to a carte blanche license for making illegal copies. You seem to be saying that Open Source can help, but I fail to see how any thing that would help the film industry would be interesting to the open source community, or vice versa.
With regards to advertising and promotion, it might surprise you to know that film companies generally MAKE money by advertising and promoting films. It's unclear how putting a few lines in a Credits section of a file would be better than that.
Lastly, your comments about BMRT and film making are so far off base that I brought your comments to the attention of Larry Gritz, author of BMRT. His comments regarding BMRT (which isn't open source by the way) will undoubtably follow. I'll merely state that your comments indicate a lack of understanding about how movies are made, and where the costs from them arise. It might shock you to find out just how much computing power a studio like Pixar uses, and yet how small a percentage that actually works out to over the cost of a production.
Now, regarding the whole DVD encryption-decryption idea: I am a big of fan of Open Source software as the next guy. I want to be able to play my DVD movies on my Linux and FreeBSD boxes as much as the next guy. But the bottom line is that to be able to play movies opens up the possibility that people could copy movies. Hence, it isn't in the cards that any movie distributor would like open source software, since it basically create a way to copy their protected media, which they have some significant interest in protecting.
Frankly, I don't see any good way out of this dilemma. Given that movie distributors want to protect their copyrighted media, it seems that nothing the open source community can do will help them in that regard. And anything which the movie distributors do to help out Open Source software is either useless or allows unrestricted copying.
I'm confused by people calling this a "tax". This is a fee, paid to the local telephone companies, not a tax.
Do the phone companies incur additional costs as the result of people using local phone lines to access ISPs? I have no doubt that they do. But the rules of the game have been "local service is free" virtually forever. If they change by introducing some kind of additional fee for using local networks, I suspect it will serve to drive those still using phonelines for internet access to other technologies such as cable modems.
In some sense, I agree with your attitude. The fact that a group of talented young engineers have built a picosat that will soon be launched should be the story.
But welcome back to the real world. Some of us know that women are good at math and science, but not every young woman knows that yet. Many are still (either subtly or overtly) steered away from productive and valuable careers in math or science. The reporting of such achievements can serve as examples to young people (especially young women) of what they can achieve.
On the achievment itself, it's _damn_ cool. I look forward to tuning in to the little bird.
In some sense, I agree with your attitude. The fact that a group of talented young engineers have built a picosat that will soon be launched should be the story.
But welcome back to the real world. Some of us know that women are good at math and science, but not every young woman knows that yet. Many are still (either subtly or overtly) steered away from productive and valuable careers in math or science. The reporting of such achievements can serve as examples to young people (especially young women) of what they can achieve.
On the achievment itself, it's _damn_ cool. I look forward to tuning in to the little bird.
I suspect that you are correct. ESR is trying to get people to buy into Open Source Software by romanticizing the entire process and stirring up the backlash against "Cathedral" organziations like those in Redmond.
Unfortunately, there is little difference between this romanticizing, and the romanticizing done by companies such as Micro$oft to promote their own products. You can't even make the argument that money isn't made by such advocacy, as the rather robust stock evaluations of RHAT and others will attest to.
No amount of commenting can make bad code any better.
Comments also often contain bugs, or are out of date.
When code is properly designed and implemented, it's intent is usually pretty clear. Code that requires explanation can nearly always be rewritten so that it's intent is clear even without comments.
Slashdot Mirror
I've enjoyed Mr. Ebert's commentary on movies for years, but in this matter I feel he's revealing a certain degree of "fuddy-duddy-ism".
By way of disclosure, I work for Pixar Animation Studios, but I do not speak for them.
By means of an informal poll, I have yet to meet a single person who saw Toy Story 2 on a digital projector and thought it looked worse than ANY print of the same movie. The lack of dust, film scratches, sprocket jitter and generational loss are CLEARLY apparent, and quite striking. To be fair, Toy Story was mastered directly from the original digital source, and we have some pretty talented engineers and DPs who worked to make that look good. I have not had the chance to view other films like The Phantom Menace in such a theater, so I won't say that digital projection makes for a uniformly better picture, but TS2 is a good technology demonstration that remarkable things are possible.
Right now digital projection systems are rather expensive: of course, there are only a couple of dozen theater ready ones in the world, so perhaps that's not too surprising. The interesting this is that virtually everything in a digital projector will benefit from Moore's law kind of price/performance drops. So it takes 100gb of disk to store a movie. The disks are getting cheaper, and what's remarkable: they are REUSEABLE. Ever price what a 35mm film print costs? Try multiplying those by the 3000 theaters that a major release might have, and you run into some serious money. Digital projectors have MUCH fewer moving parts than a conventional projector, because they have no film transport and no shutter.
Costs aside, it is my firm belief that digital theaters will deliver a better LOOKING picture as well. The resolution of current projectors (1280x1024) is thought to be 'low' but in fact, when done properly, they can look amazingly good. The typical film projector has several pixels worth of sprocket jitter, so the imagined higher resolution of film isn't usually delivered to the screen, also being eaten up in generational loss, aging of film pigments, scratches and dust. Some of you may recall the television special detailing the remastering of Star Wars. The original film required LARGE amounts of work to correct for the terrible aging of the pigments involved. Digital projection means that 30 or 300 years from now, you will be able to show the film EXACTLY as it appears today. That's a pretty strong selling point in my book.
As a brief aside addressing the choice of resolution and compression ratios, let me just point out that first of all, costs for disks is dropping, while the resolution that people want to view things at is pretty much fixed. This means that the entire issue of compression will become a moot point anyway. If you assume 2M pixels per frame, 24fps, 12 bits per color channel, and a 90 minute movie, the uncompressed data storage is about 1 TB. That's admittedly a lot of storage, but it isn't infinite, and its becoming easier to do as time goes on. Current (lossy) compression technology can easily cut that down by a factor of more than ten, even without using frame/frame coherence. Forget MPEG, state of the art leaves that stuff in the dust. 100gb disk arrays aren't particularly expensive, and next year they'll be half as expensive as they are this year. In three to five years, we probably wouldn't even need to compress at all.
I saw Ebert on television claim that when he went to a theater, he wants to see light coming through celluloid. Well, he's entitled to watch whatever he likes however he likes, but I think I've seen the future, and it isn't far off. I'm waiting eagerly for the next installment of Star Wars. I suspect it will turn a few heads.
Hmmmm... I guess you'll have to write a compiler then. And how do you do that? Use assembly language
Let me guess, you don't write compilers for a living....
At the risk of making an overly general statement, NOBODY does what you describe anymore. A compiler is one of the more perfect examples of something that shouldn't be written in assembly language. It embodies some fairly complex algorithms but is itself not very speed critical. Most people don't care about how fast their compiler outputs code, merely how fast the resulting code is when run.
The two most available open source C compilers (gcc and lcc) take a similar approach by defining new machines via a rather compact machine description file. This allows the compiler to properly abstract the portions of the compiler which aren't machine dependant (the input parser, symbol table construction, parse tree assembly and the like) from the parts that are (code generation). The result is a compiler which is easier to port to new architectures, and has less errors.
Most of these compilers similarly use assemblers written in C. Initially these are cross-assemblers, but once the assembler is compiled by your new compiler, you can run the assembler in native mode. *voila*
To write machine descriptions for these compilers, you of course _do_ need to know something about the assembly language of the target machine, but this is vastly easier than writing the entire compiler in assembly.
Traditional PCs have a number of advantages, but also a number of glaring problems.
Undoutably one of their key advantages has been the ability to update and customize configurations dependant on customer needs.
This is also one of their key disadvantages. Try just looking at a PC sometimes. It's a virtual hornet's nest of cables and wires, most of which just serve to carry signals from one part of the computer to another. How much of the total complexity (read cost) of a PC is just to support this kind of legacy interconnects?
It seems rather obvious that most people in the world probably don't care that much about all the flexibility that a PC offers. They'd like to buy an inexpensive machine to access the net and play a few games. An internet appliance if you like. For the most part, the traditional "assemble-it-yourself" computer gets in the way of that.
*sigh*
.1% chance for any given report to be true, then the chances are 99.8% that one or more of them are true in a year (given 6000 sightings). It doesn't take a 30 page dissertation to work it out.
This is classic "wooly thinking". First of all, where do you get this 99.9% figure? I say that 100% of all UFO sightings are fakes, crocks, and misidentified. And until a UFO comes down in the middle of Times Square, squashes a couple of SUVs, and purple haired maidens with six breasts emerge and asked to be taken to our leader, you can't prove that I am not correct.
I find it amazing that people can be so far separated from common sense and the principle of Occam's Razor that they can't see through crap like this. We have an anectdotal story about some young anonymous Mexican girl recovering a skull, at some unspecified location and time in the past. We have assertions (but curiously no pictures) that the skull shows bizarre deformities. We have assertions from that it has been viewed by a number of "experts", who of course couldn't "risk the ridicule of their scientific peers".
Uh huh.
Okay, so far it isn't so bad. But what really gets me is the fanciful leap into explanations. It "must" be a Gray, or a Gray-human hybrid. Uh huh. Why? Are their other possibilities. Say, well, FRAUD for one? Given that no proof that Grays exist, isn't it just a teensy bit premature to assign a skeleton to their particular family tree? Then we get into this crap about DNA testing. As several other posters noted, there is little reason to believe that DNA has any particular magical properties that make it the only way (or the most likely way, or even a way that has occurred more than once) for life to develop in the universe. Since no copy of Gray DNA has ever been recovered, it seems hard to know how DNA testing would ever result in an appropriate classification.
This is bunkery of the highest order. It tries to take on the trappings of science, but it is just a particularly odd form of creationism, with all its attendant perils.
By the way, if there is a
We are getting a bit astray from the topic of DVD encryption, but I'll merely update with a couple of specific comments on your musings, and then let this drop back to real discussions on the real topic at hand.
The film industry has invested large amounts of money in developing properties which have a rather sizeable commercial value. They wish to protect their investment by limiting the ordinary individuals ability to partake of their "experience" to those who have paid for the ticket, cd, video tape, or dvd.
There is only one problem with that, and it is a problem that is shared with most media: in order to show a movie or play a cd, it is necessary to read the information off a disk and convert it into a representation which people can watch or listen to. And if a person can listen to or watch a particular representation, then so can some form of machine, and therefore, copies can be made.
The film industry has so far used dvds because the decoders were mostly encoded in physical devices which would be difficult and/or expensive for the average joe to duplicate. But once software decoders became available, whether they encrypted keys or not, it seems like they let the cat out of the bag, because software IS cheap to duplicate and distribute.
Of course, the open source vs. trade secret arguments have nothing to do with any of this, which was my point originally. This point was not lost on a number of other people who commented similarly on your original posting.
The rest of your posting (ray tracing, radiosity and the Free Film project) are fascinating topics in themselves, but not really apropos to the topic at hand.
What is the point you are trying to make regarding Open Source?
You claim that the film industry has more to fear by not adopting Open Source than by embracing it.
You make the claim that an open source ideology reduces distribution costs because the technology required to mass produce is being developed faster. I'm not sure what one has to do with the other. Just how has open source influenced the distribution costs (ignoring the obvious "Gee, I can download The Phantom Menace off this website for nothing") CDs and DVDs made an enormous impact because they are cheap to reproduce and give great fidelity, not because of any revolution caused by Open Source.
As for piracy protection, the problem with the DVD encryption as I understand it wasn't caused by a lack of proper encryption algorithms, but rather by the fact that some doofus shipped the keys in an unprotected form with an application. You can make things foolproof I guess, but you can't make them damn-fool proof. Perhaps some form of technical review should have been done before the offenders shipped their app, but again, your points seems to miss the mark. You can't write an open source DVD decryptor, because one key components of the system is a secret key that the film industry doesn't want people to have because it amounts to a carte blanche license for making illegal copies. You seem to be saying that Open Source can help, but I fail to see how any thing that would help the film industry would be interesting to the open source community, or vice versa.
With regards to advertising and promotion, it might surprise you to know that film companies generally MAKE money by advertising and promoting films. It's unclear how putting a few lines in a Credits section of a file would be better than that.
Lastly, your comments about BMRT and film making are so far off base that I brought your comments to the attention of Larry Gritz, author of BMRT. His comments regarding BMRT (which isn't open source by the way) will undoubtably follow. I'll merely state that your comments indicate a lack of understanding about how movies are made, and where the costs from them arise. It might shock you to find out just how much computing power a studio like Pixar uses, and yet how small a percentage that actually works out to over the cost of a production.
Now, regarding the whole DVD encryption-decryption idea: I am a big of fan of Open Source software as the next guy. I want to be able to play my DVD movies on my Linux and FreeBSD boxes as much as the next guy. But the bottom line is that to be able to play movies opens up the possibility that people could copy movies. Hence, it isn't in the cards that any movie distributor would like open source software, since it basically create a way to copy their protected media, which they have some significant interest in protecting.
Frankly, I don't see any good way out of this dilemma. Given that movie distributors want to protect their copyrighted media, it seems that nothing the open source community can do will help them in that regard. And anything which the movie distributors do to help out Open Source software is either useless or allows unrestricted copying.
Open source isn't always the answer.
I'm confused by people calling this a "tax". This is a fee, paid to the local telephone companies, not a tax.
Do the phone companies incur additional costs as the result of people using local phone lines to access ISPs? I have no doubt that they do. But the rules of the game have been "local service is free" virtually forever. If they change by introducing some kind of additional fee for using local networks, I suspect it will serve to drive those still using phonelines for internet access to other technologies such as cable modems.
In some sense, I agree with your attitude. The fact that a group of talented young engineers have built a picosat that will soon be launched should be the story.
But welcome back to the real world. Some of us know that women are good at math and science, but not every young woman knows that yet. Many are still (either subtly or overtly) steered away from productive and valuable careers in math or science. The reporting of such achievements can serve as examples to young people (especially young women) of what they can achieve.
On the achievment itself, it's _damn_ cool. I look forward to tuning in to the little bird.
In some sense, I agree with your attitude. The fact that a group of talented young engineers have built a picosat that will soon be launched should be the story.
But welcome back to the real world. Some of us know that women are good at math and science, but not every young woman knows that yet. Many are still (either subtly or overtly) steered away from productive and valuable careers in math or science. The reporting of such achievements can serve as examples to young people (especially young women) of what they can achieve.
On the achievment itself, it's _damn_ cool. I look forward to tuning in to the little bird.
I suspect that you are correct. ESR is trying to get people to buy into Open Source Software by romanticizing the entire process and stirring up the backlash against "Cathedral" organziations like those in Redmond.
Unfortunately, there is little difference between this romanticizing, and the romanticizing done by companies such as Micro$oft to promote their own products. You can't even make the argument that money isn't made by such advocacy, as the rather robust stock evaluations of RHAT and others will attest to.
No amount of commenting can make bad code any better.
Comments also often contain bugs, or are out of date.
When code is properly designed and implemented, it's intent is usually pretty clear. Code that requires explanation can nearly always be rewritten so that it's intent is clear even without comments.