First, didn't we all *hate* the CD caddies that came with the first cdrom drives? We'll, it's back to something not unlike them with these flexible CDs. Read the "Adapter" section of the technical description paper.
Second, they are also planning to have some lame "content protection system" as a part of flexcds. See the "flexRights" section of the features page.
Summary -- it's less durable than existing CDs, requires a caddy-style adapter to use in your drive, and is tied up with nasty intellectual property cartel ambitions. Oh, and no sign of a version that's writable in our existing CDR or CDRW drives, either.
The "XP" technology refers at least to the incorporation of the SSE instruction set---my 1GHz duron has that. I think there are other non-marketing differences between "XP" and older K7s, but find someone more technical than me to explain them.
Let's see.. playing a media stream at varying speeds. In MPEG video, it's moderately complex to skip frames, but I would be surprised if it were not discussed in mpeg literature for use in systems which cannot decode the full stream in realtime, but must instead decode every Nth frame (N=2-4). Playing more slowly just involves showing a frame for more than one retrace. Any decent MPEG player should already have frameskip control, and wiring a timebase multiplier to a UI knob is not rocket science.
Playing and recording at the same time is a simple matter of having a multitasking OS, a disk fast enough to handle the bandwidth of two streams, and separate encoder/decoder hardware.
As for "connecting DVRs to a network in the home", DVRs are just another piece of network hardware. Streaming media technology is probably the subject of patents that precede DVRs. Besides, the hard parts of streaming are when bandwidth is scarse, which isn't the case over ethernet (2mbps wireless excepted)
Playing backwards is a little more complex than playing forward at variable rate, but again most DVD players have this capability. This patent has a April 1998 application date, but DVDs date to 1995 ("December 09, 1995: The final DVD format is originally announced.") Since DVDs are streams of video, the capabilities of DVDs to manipulate the order in which the stream is presented seem relevant. Surely "play in reverse" wasn't missing from DVD for their first two years of existence..
Other posters have discussed how SonicBlue and TiVO will probably cross-license, and the patents wouldn't stand up to scrutiny anyway, so the only thing they'll be good for is to raise the bar against additional participants in the DVR market (those who don't have deep-enough pokets to withstand a lawsuit, which means any startup...) and maybe to furnish C&D-letter fodder for OpenDVR software projects.
The story seems to imply that it's something about storage that is the limiting factor in today's digital cameras. But that's far from true, IMO. The first two problems with digital cameras are resolution (6+ megapixel professional cameras notwithstanding) and battery life (again, the pros aren't sticking 4AAs in their cameras).. With solid state storage for 1000 pictures under $200 and 2oz, and portable hard-drive storage for 10000+ pictures under $300 and 12oz, it's a distant third as far as I'm concerned.
Another post suggests that the write-once capability of the 3D memory chip will be some kind of magic to prove that a photo hasn't been retouched. But why couldn't I get a writer for these memories in my PC, and store an image to it after I run gimp/photoshop? This problem can be solved by having the camera digitally sign its images---after you load it in your photo editor, you can't write out another file with a valid signature. (But is Kodak or anybody else prepared to create a tamper-proof digital camera and adequately protect the digital signing key that would be built into such a camera?)
Now that there are high-quality digital photographic printing processes, 300-600dpi onto photo paper, is somebody working on 3000-6000dpi printing onto photographic negatives? The same technology that makes our digital prints look good should also let us make awesome fake negatives! So much for the faith of people in the photographic negative...
The problem with this sort of rationalization of the bible is that there's no support in the text for any of these odd interpretations of "this generation". If I were to go willy-nilly through your Sacred Text, and add qualifiers like "this kind of thing", who knows what weird religion you'd soon be following!
On the other hand, it's quite common for religious leaders to predict the End of the World within their lifetime---and they're invariably wrong! Some of them just slink off into a hole somewhere, others convince their followers that they were actually predicting a different End of the World all along.
I think the person we're talking about here is one of the more slick-tongued variety.
Let's look at it this way: If a human-animal cross is possible, then this theological view that humans have souls and animals don't is in need of revision.
If theology were science, then an experiment crossing a human and an animal would be one way to disprove the thesis "god exists".
Of course, theology isn't science, so religious people don't want you to perform the experiment in the first place!
If the mere result from some scientific experiment or other has the power to "shake your beliefs to the core", it's time to get a more resilient set of beliefs!
The EFF DES machine was breaking DES quite quickly even without special information about the key--three years ago. 56 hours. Since you can speed it up by adding transistors or cranking speed, I bet an implementation today could reach 10 hours. If you have something that can tell you half the bits in the key, setup time would begin to dominate over solve time.
So, yeah, it sucks that these people found a weakness that lets them guess key bits, but DES should have been tossed years ago. At least for 3DES, which doubles the effective key size. But isn't the AES standard finalized now?
Problem is, banks don't want to replace outdated hardware and networks, as long as their customers don't know they should be scared where their money is going.
On the one hand, a register-based VM can have lower (virtual) instruction counts, since
add r1, r2, r3
would be
load r2
load r3
add
store r1
in a stack-based machine.
On the other hand, it doesn't really help to have registers in the virtual instruction set, because the targets all have different register counts---For instance, none on a stack machine, 8 registers on x86, ~32 registers on many RISC, ~128 registers on ia64, ~16 registers on x86-64 (and the number of floating-point registers is similarly variable)
So you can pick a stack-based vm and let the native code translator optimize stack actions into register actions (not hard) or pick a large enough number of registers in the virtual instruction set that code will typically never spill registers to the stack (say, 1024 registers) and let the native code translator optimze register actions into spill-fill operations (not hard)
*OR* you could pick a number of registers that is small enough to require spill-fill in the virtual machine code, because that makes it easy to write the translator for one of your architectures (x86) at the expense of either performance in other architectures (Only using 4 of your 128 integer registers?) or complexity there (turning memory operations into register operations is a tougher problem than either of the alternatives I presented above---GNU Lightning, upon which I believe IVM may be based, doesn't do it)
That's right -- the "internet virtual machine" seems designed to map well onto a single CPU, the x86.
The standard library seems designed to map well onto a single OS, Unix.
You can run it on your big fancy RISC machine (well, if they've gotten around to writing a stable back-end), but it'll still approximate the performance of a register-emasculated ia32 processor.
It's a bit more portable than an Linux ELF binary, but not much. And about as revolutionary, given that a Linux runtime also exists on the modern BSDs (for example)
There is an alternative to reverse engineering. With open specifications, multiple implementations can coexist. This is true of internet standards such as TCP/IP, SMTP, HTTP, and others. It's true of languages like C, C++, Pascal, and Ecma/JavaScript.
To say that all interoperability is thanks to reverse engineering is overstating the case.
I must admit I'm having trouble fathoming just how this works. But I know why Phoenix has done this.
Remember the Microsoft anti-trust trial? One detail that emerged was that Microsoft does not permit OEMs to perform modifications to the desktop, startup sequence, etc. This means that the OEMs can't give the user a "custom experience" or differentiate their machine from others using Microsoft's software.
This Phoenix BIOS trick lets OEMs skirt the Microsoft OEM license by performing the customization after the user has the machine.
So, in one way, I say "kudos" to Phoenix for figuring out how to subvert Microsoft's restrictive OEM licensing agreements in this way.
On the other hand, I'd like to understand more technical details of the feature, whether it could bite me while I'm trying to use Linux, etc. Has anybody turned up relevant patents?
The part of the consitution being alluded to is not "free speech" or "free expression".
The section of the constitution in question is Article I Section 8, enumerating the power of the Congress. One of the powers is
To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries
The abuse is that these "global monopolies" are attempting to use a combination of technological controls (DVD CSS, SDMI) and laws (DMCA, UCITA, Sunny Bono Copyright Act) to extend their control over content they did not create and in perpetuity. (In fact, even if copyrights are not extended indefinitely, you will still not be able to use decss to view or make a copy of your old 20th century DVDs, because the program which could do so will violate the DMCA, even if it has some uses which are legitimate)
The concept of fair use may not be found in the Constitution or in the Bill of Rights, but does have a long history in the American legal system, as well as in other countries the world over. It is legal to excerpt images and audio clips under certain situations (for instance, in a review, or for purposes of satire or parody) Again, between technological controls like SDMI and CSS and laws such as DMCA, it may be legal to excerpt images or audio clips from a DVD or a watermarked digital audio file, but creation or ownership of software to do so will be illegal under the DMCA.
In other words, content covered by the DMCA plus a system access controls remove both your constitutional right to eventually use the work when it has passed into the public domain, as well as to make use of it under the fair use provisions of US Federal law while it is still under copyright.
I am not against globalization, but I am against any law which erodes both my constitutional rights as well as rights that have been traditionally afforded under the laws of the United States, even if they come under the guise of being necessary for "global business".
I also think it's a great pity that it's a case like Napster which is getting all the public's attention. I view this as a big victory for the RIAA and MPAA, because I would much rather see publicity about the way DMCA+access controls make it impossible to make fair use of the content on DVDs, or to make free use of the content once it passes into the public domain.
What does the "Requested Connections" item under "Benchmark Configuration" mean? On the various benchmarks linked, this number is equal to or just higher than the the SPECweb99 score. Do the benchmarkers select the optimum value for this, where increasing the value doesn't increase the number of "conforming connections" (or decreases it)? That's how I read it, but even after looking at spec's white paper I'm not 100% sure. Can someone else clue me in?
Slashdot Mirror
First, didn't we all *hate* the CD caddies that came with the first cdrom drives? We'll, it's back to something not unlike them with these flexible CDs. Read the "Adapter" section of the technical description paper.
Second, they are also planning to have some lame "content protection system" as a part of flexcds. See the "flexRights" section of the features page.
Summary -- it's less durable than existing CDs, requires a caddy-style adapter to use in your drive, and is tied up with nasty intellectual property cartel ambitions. Oh, and no sign of a version that's writable in our existing CDR or CDRW drives, either.
The "XP" technology refers at least to the incorporation of the SSE instruction set---my 1GHz duron has that. I think there are other non-marketing differences between "XP" and older K7s, but find someone more technical than me to explain them.
Let's see .. playing a media stream at varying speeds. In MPEG video, it's moderately complex to skip frames, but I would be surprised if it were not discussed in mpeg literature for use in systems which cannot decode the full stream in realtime, but must instead decode every Nth frame (N=2-4). Playing more slowly just involves showing a frame for more than one retrace. Any decent MPEG player should already have frameskip control, and wiring a timebase multiplier to a UI knob is not rocket science.
Playing and recording at the same time is a simple matter of having a multitasking OS, a disk fast enough to handle the bandwidth of two streams, and separate encoder/decoder hardware.
As for "connecting DVRs to a network in the home", DVRs are just another piece of network hardware. Streaming media technology is probably the subject of patents that precede DVRs. Besides, the hard parts of streaming are when bandwidth is scarse, which isn't the case over ethernet (2mbps wireless excepted)
Playing backwards is a little more complex than playing forward at variable rate, but again most DVD players have this capability. This patent has a April 1998 application date, but DVDs date to 1995 ("December 09, 1995: The final DVD format is originally announced.") Since DVDs are streams of video, the capabilities of DVDs to manipulate the order in which the stream is presented seem relevant. Surely "play in reverse" wasn't missing from DVD for their first two years of existence..
Other posters have discussed how SonicBlue and TiVO will probably cross-license, and the patents wouldn't stand up to scrutiny anyway, so the only thing they'll be good for is to raise the bar against additional participants in the DVR market (those who don't have deep-enough pokets to withstand a lawsuit, which means any startup...) and maybe to furnish C&D-letter fodder for OpenDVR software projects.
The story seems to imply that it's something about storage that is the limiting factor in today's digital cameras. But that's far from true, IMO. The first two problems with digital cameras are resolution (6+ megapixel professional cameras notwithstanding) and battery life (again, the pros aren't sticking 4AAs in their cameras) .. With solid state storage for 1000 pictures under $200 and 2oz, and portable hard-drive storage for 10000+ pictures under $300 and 12oz, it's a distant third as far as I'm concerned.
Another post suggests that the write-once capability of the 3D memory chip will be some kind of magic to prove that a photo hasn't been retouched. But why couldn't I get a writer for these memories in my PC, and store an image to it after I run gimp/photoshop? This problem can be solved by having the camera digitally sign its images---after you load it in your photo editor, you can't write out another file with a valid signature. (But is Kodak or anybody else prepared to create a tamper-proof digital camera and adequately protect the digital signing key that would be built into such a camera?)
Now that there are high-quality digital photographic printing processes, 300-600dpi onto photo paper, is somebody working on 3000-6000dpi printing onto photographic negatives? The same technology that makes our digital prints look good should also let us make awesome fake negatives! So much for the faith of people in the photographic negative...
The problem with this sort of rationalization of the bible is that there's no support in the text for any of these odd interpretations of "this generation". If I were to go willy-nilly through your Sacred Text, and add qualifiers like "this kind of thing", who knows what weird religion you'd soon be following!
On the other hand, it's quite common for religious leaders to predict the End of the World within their lifetime---and they're invariably wrong! Some of them just slink off into a hole somewhere, others convince their followers that they were actually predicting a different End of the World all along.
I think the person we're talking about here is one of the more slick-tongued variety.
Let's look at it this way: If a human-animal cross is possible, then this theological view that humans have souls and animals don't is in need of revision.
If theology were science, then an experiment crossing a human and an animal would be one way to disprove the thesis "god exists".
Of course, theology isn't science, so religious people don't want you to perform the experiment in the first place!
If the mere result from some scientific experiment or other has the power to "shake your beliefs to the core", it's time to get a more resilient set of beliefs!
So, yeah, it sucks that these people found a weakness that lets them guess key bits, but DES should have been tossed years ago. At least for 3DES, which doubles the effective key size. But isn't the AES standard finalized now?
Problem is, banks don't want to replace outdated hardware and networks, as long as their customers don't know they should be scared where their money is going.
On the one hand, a register-based VM can have lower (virtual) instruction counts, since
add r1, r2, r3
would be
load r2
load r3
add
store r1
in a stack-based machine.
On the other hand, it doesn't really help to have registers in the virtual instruction set, because the targets all have different register counts---For instance, none on a stack machine, 8 registers on x86, ~32 registers on many RISC, ~128 registers on ia64, ~16 registers on x86-64 (and the number of floating-point registers is similarly variable)
So you can pick a stack-based vm and let the native code translator optimize stack actions into register actions (not hard) or pick a large enough number of registers in the virtual instruction set that code will typically never spill registers to the stack (say, 1024 registers) and let the native code translator optimze register actions into spill-fill operations (not hard)
*OR* you could pick a number of registers that is small enough to require spill-fill in the virtual machine code, because that makes it easy to write the translator for one of your architectures (x86) at the expense of either performance in other architectures (Only using 4 of your 128 integer registers?) or complexity there (turning memory operations into register operations is a tougher problem than either of the alternatives I presented above---GNU Lightning, upon which I believe IVM may be based, doesn't do it)
That's right -- the "internet virtual machine" seems designed to map well onto a single CPU, the x86.
The standard library seems designed to map well onto a single OS, Unix.
You can run it on your big fancy RISC machine (well, if they've gotten around to writing a stable back-end), but it'll still approximate the performance of a register-emasculated ia32 processor.
It's a bit more portable than an Linux ELF binary, but not much. And about as revolutionary, given that a Linux runtime also exists on the modern BSDs (for example)
There is an alternative to reverse engineering. With open specifications, multiple implementations can coexist. This is true of internet standards such as TCP/IP, SMTP, HTTP, and others. It's true of languages like C, C++, Pascal, and Ecma/JavaScript. To say that all interoperability is thanks to reverse engineering is overstating the case.
Remember the Microsoft anti-trust trial? One detail that emerged was that Microsoft does not permit OEMs to perform modifications to the desktop, startup sequence, etc. This means that the OEMs can't give the user a "custom experience" or differentiate their machine from others using Microsoft's software.
This Phoenix BIOS trick lets OEMs skirt the Microsoft OEM license by performing the customization after the user has the machine.
So, in one way, I say "kudos" to Phoenix for figuring out how to subvert Microsoft's restrictive OEM licensing agreements in this way.
On the other hand, I'd like to understand more technical details of the feature, whether it could bite me while I'm trying to use Linux, etc. Has anybody turned up relevant patents?
The abuse is that these "global monopolies" are attempting to use a combination of technological controls (DVD CSS, SDMI) and laws (DMCA, UCITA, Sunny Bono Copyright Act) to extend their control over content they did not create and in perpetuity. (In fact, even if copyrights are not extended indefinitely, you will still not be able to use decss to view or make a copy of your old 20th century DVDs, because the program which could do so will violate the DMCA, even if it has some uses which are legitimate)
The concept of fair use may not be found in the Constitution or in the Bill of Rights, but does have a long history in the American legal system, as well as in other countries the world over. It is legal to excerpt images and audio clips under certain situations (for instance, in a review, or for purposes of satire or parody) Again, between technological controls like SDMI and CSS and laws such as DMCA, it may be legal to excerpt images or audio clips from a DVD or a watermarked digital audio file, but creation or ownership of software to do so will be illegal under the DMCA.
In other words, content covered by the DMCA plus a system access controls remove both your constitutional right to eventually use the work when it has passed into the public domain, as well as to make use of it under the fair use provisions of US Federal law while it is still under copyright.
I am not against globalization, but I am against any law which erodes both my constitutional rights as well as rights that have been traditionally afforded under the laws of the United States, even if they come under the guise of being necessary for "global business".
I also think it's a great pity that it's a case like Napster which is getting all the public's attention. I view this as a big victory for the RIAA and MPAA, because I would much rather see publicity about the way DMCA+access controls make it impossible to make fair use of the content on DVDs, or to make free use of the content once it passes into the public domain.
Patents last 17 years. Copyrights last much longer.
God, can't we at least keep our facts straight?
b0rken