Actually the original ATX spec called for the the power supply fan to be an intake (rather than an exhaust) it was supposed to blow air over the processor below it which is what the grill in the bottom of the PSU is for.
OEMs soon realised that INTEL had taken leave of it's senses and started designing 2 or 3 fan systems with an intake at the bottom front and having the PSU fan as an exhaust along with an optional secondary exhuast.
Of course you can only do so much with air, to really cool things down vapour phase systems are needed. I suppose the ultimate would be a liquid helium system but that would cost significatly more than the computer:)
I beg your pardon, FORTH is anything but dead. It has just evolved for example the plugin-and-play ROMs on all your PCI cards are written in a FORTH derivative, the original aim being to achieve platform independence.
Just because something isn't on view anymore doesn't mean that it doesn't exist.
This is nothing to do with XFree4, the Linux kernel or anything else in the system. The plain fact is that the SMP optimisations in Q3A for Linux don't work and ID haven't got around to fixing them. (Although to be fair writing good threaded code is very difficult.)
Much as I hate to say it Win2K or NT is probably the best platform for SMP Q3A. However I have yet to see any benchmarks that show that Q3A is faster using SMP under NT than uniprocessor under 98. The limiting factor in your system is going to be the PII 333 cpu.
On my machine (PIII 500) large FFA matches slow to a crawl (30-40fps) as they become CPU bound. Duels and the like stay firmly above 100fps.
> I've always considered MP3 lossy and distorted, > even at ultra-high bitrates, so I'd like to see > something a bit better. Do such formats exist, > or should I stick with CDs and huge.wav files?
Try Monkey audio lossless compression. Does about 50% reduction on most audio.
> A little more care required before you post > inaccurate stuff like that.
Ahem.
The nCipher device contains slightly customised RISC processors which just happen to be running an embedded OS and running dedicated software to do the encryption. I'll agree that the line between software and hardware is getting a little fuzzy here.
Many communications companies have the same problem. For example the IPsec implementation in Lucent Pipeline routers (formally Ascend) is done in software on a gereral purpose processor for the same reason.
If the rules have been relaxed in France it must have been very recently (last 6 months). I have been involved in several companies which have had to go to great expense to install _unencrypted_ leased lines into France as they couldn't just add the French offices to a worldwide VPN for encryption reasons.
I don't see how this is going to be accomplished. In some european contries, like France, encryption is illegal (unless you are the government of course) and in others restrictions are placed on it's use. For example in the UK it is actually illegal to do encryption in hardware. This dates back to the days where the implementation in software were too slow to be useful.
Thus if we have restrictions on internal use of encryption I don't see how we are going to develop and export strong encryption.
France has more oustanding european court cases against it that any other nation
However it is throttled to 8 Mb/s upstream. The vast majority of downstream bandwidth in Cambridge goes to feed personal computers in student rooms. If Oxford have a more draconian policy on bandwidth allocation to students then they may well not need anything faster than 34 Mb/s. I suspect that a 34 Mb/s connect is vastly cheaper than a 155 Mb/s connect.
Macrovision level 1 as used on VHS video is indeed part of the video signal.
Macrovision level 2 is implemented as a burst during the vertical blanking signal between PAL (or NTSC) frames which confuses the auto gain control circuit on a VCR. Since the vob files on the disk only contain an MPEG program stream. You couldn't physically put macrovison in the vob. The macrovison is generated by the digital to analogue converter (RAMDAC in the video world) attached to the MPEG decoder. This is why it is possible to hack many players not produce macrovison.
If this was not true I would not be able to watch DVD's on my LCD projector which is just as susceptable to macrovision (level 2) as VCR's are.
It is certainly true that the flag to tell the player to produce macrovision is in the vob but macrovison itself can't be.
It looks like disks recorded on this will not be playable on legacy DVD players. The article says:
Pioneer, as well as other major hardware manufacturers, will introduce Video Recording Format-ready DVD-Video players, and it will be possible to replay DVD-RW recorded discs on those players too
It also mentions that the Video Recording Format could be made compliant with DVD video but hasn't, you can probably guess the reason why.
As for copy protection macrovision is added on output and is not encoded onto the disk. Thus using the analogue input (scrubbed of macrovision) it should be possible to copy disks. As far as I know no one has incorporated an analogue Copy Generation Management System into DVD. I assume that the digital system will let you make one digital copy of a disk then stop any further copies. The disks that have CSS would probably have the appropriate flags to stop digital output working on a player with this facility (non exist yet).
The article goes further:
Furthermore, the DVR-1000 implements secure media ID detection functions, which form a key element in disabling playback of discs containing unauthorized copies.
I assume that this means that each DVD-RW contains a ID so that only their disks can be used. However I don't see how this could be used to prevent unorthorized copies (you just need to use their recorders and media).
It is not at all clear if you could record digital TV (cable or satallite) in a digital manor onto these devices. Would the cable company for you to pay for the privilege of recording their programs or would they prevent it and force you to buy a device such as a ReplayTV ?
In short this is a technology that has been hamstrung by the movie studios.
The problem is that Napster only has one use, transfering MP3 files. I think that this is why it can't hide behind a "We just tell them where to go" defence. A more general system such as gnutella does not have these kind of problems. The problem with gnutella is that it doesn't preserve anonymity.
It is actually possible to transfer data between two points and to stop each end knowing what is going where but protect against it being spied upon or changed while in transit. However this tends to involve very hairy discrete mathematics and codeing a good implementation is beyond the vast majority of people.
Lets face it whatever we do industry are going to turn round and say you might be infringing our copyright so we are going to sue you. This is a fight that I do not think that we can win.
1) It is slow 2) It is proprietary 3) It only runs on windows 4) It isn't a real programming language 5) Did I mention that it is so slooooooooow
Although programs may look simpler to read and maintain in VB in reality it is just hiding the grottyness under a high level API. Because the libraries that it depends upon are precompiled you can't go and look what a function actually does when it doesn't act as expected.
It is slow, the language only recently began to be compiled. Pprevious attemps included embeding code+interpreter in a Windows PE format executable.
In a real language you can write a compiler for your language in that language and have the compiler compile itself. (This can be done with gcc) VB can't do this.
PERL is only hard to read if you don't know PERL, I find VB hard to read as it is verbose where as PERL is rather terse.
I wonder how DRE is planning on identifying napster users? I imagine placing random taps on ethernet sockets in colleges and regions that have cable modems would have civil liberties implications.
If napster is supid enough to keep logs of it users then they and napster deserve to have there ass sued off.
So are they planning on conducting dawn raids on any room with a computer?
I think that this is little more than a show suit designed to raise publicise his cause and scare few people.
This is slightly inaccurate. The StongARM was not an ARM design it was an evolution of the ARM designed in cooperation between Digital and ARM.
When Digital was bought by Compaq the the rights to the processor reverted to ARM. Intel bought Digital's fabrication facilities unfortuately for them that entailed also having to make StrongARM and Alpha even though Intel have competing processors.
As for ARM being the next Intel this won't happen for two reason. First ARM is fabless and will remain that way. Second ARM's stated aim is to make money for it's employies in the best way it can rather than to dominate the processor market which is Intel's mission:)
DRI being availible under XFree 4 won't make a scrap of difference to support for the Geforce under Linux as Nvidia has stated that they won't use DRI but something "similar" for their OpenGL support.
This sucks and nvidia should be publically slaped for not supporting Linux.
I predict that the world+dog will post a preview/review at about 00:10 on wednesday morning. All these reviews at anandtech, firingsquad etc will be virtually identical being based upon the same press releases from nvidia:)
I wish people would put some thought in to there jounalism rather just spouting what the manufacturers tell them to.
The problem with private companes is they almost always look at the short term and try to make the most profit.
NASA has been forced to do this kind of thing in the past and what happened? The Challenger disaster.
Challenger blew up happened because the sub contractor for the solid fuel boosters decided to make them in several pieces instead of one like they had been before. The seals between the sections failed and the main fuel tank blew up.
This is what happens when private companys are involved on space, people die.
Turning hubble towards earth would blind it. The instuments on board are so sensitive that the astronomers have to be very careful they don't look at either the earth or the sun.
The show article mentions that it is going to tour to promote driving safety.
I can imagine a car that has so obvious connections with being driven dangerously (with the possible exception of the bat mobile). Then again that might be the point.
Nearly all program files downloaded from the Internet have the suffix.zip, meaning they are compressed in the format Katz developed.
Hum. What about.bz ? Bzip makes use of the Burrows-Wheeler transform on blocks of data which are then are then compressed using more standard methods such as Huffman, Run Length Encoding and Arithmethic encoding.
The resulting alogrithm compresses much better than LZ77 + Shannon. It is so good that almost ever kernel image in the linux world is a bzImage.
Katz should be praised for his work but the world of compression has moved on and zip is no longer state of the art.
When Hubble's mirror was ground the piece of glass was so large that the earths gravity affected its shape significantly. When the calculations for the mirror were made they had to take into account the difference in gravitential potential between oribit and ground level and correct for it.
The problem was that the correction was applied twice. The maths for this is actually quite complicated so don't be two hard on them.
Slashdot Mirror
Actually the original ATX spec called for the the power supply fan to be an intake (rather than an exhaust) it was supposed to blow air over the processor below it which is what the grill in the bottom of the PSU is for.
:)
OEMs soon realised that INTEL had taken leave of it's senses and started designing 2 or 3 fan systems with an intake at the bottom front and having the PSU fan as an exhaust along with an optional secondary exhuast.
Of course you can only do so much with air, to really cool things down vapour phase systems are needed. I suppose the ultimate would be a liquid helium system but that would cost significatly more than the computer
-dp
I beg your pardon, FORTH is anything but dead. It has just evolved for example the plugin-and-play ROMs on all your PCI cards are written in a FORTH derivative, the original aim being to achieve platform independence.
Just because something isn't on view anymore doesn't mean that it doesn't exist.
-dp
This is nothing to do with XFree4, the Linux kernel or anything else in the system. The plain fact is that the SMP optimisations in Q3A for Linux don't work and ID haven't got around to fixing them. (Although to be fair writing good threaded code is very difficult.)
Much as I hate to say it Win2K or NT is probably the best platform for SMP Q3A. However I have yet to see any benchmarks that show that Q3A is faster using SMP under NT than uniprocessor under 98. The limiting factor in your system is going to be the PII 333 cpu.
On my machine (PIII 500) large FFA matches slow to a crawl (30-40fps) as they become CPU bound. Duels and the like stay firmly above 100fps.
-dp
> I've always considered MP3 lossy and distorted, > even at ultra-high bitrates, so I'd like to see > something a bit better. Do such formats exist, > or should I stick with CDs and huge .wav files?
Try Monkey audio lossless compression. Does about 50% reduction on most audio.
-dp
AFAIK This refers to an old hardware joke which gave the reason why any particular chip didn't work as "The Blue Smoke has escaped".
Hence a smokeless CPU = a CPU that is broken.
-p
> A little more care required before you post > inaccurate stuff like that.
Ahem.
The nCipher device contains slightly customised RISC processors which just happen to be running an embedded OS and running dedicated software to do the encryption. I'll agree that the line between software and hardware is getting a little fuzzy here.
Many communications companies have the same problem. For example the IPsec implementation in Lucent Pipeline routers (formally Ascend) is done in software on a gereral purpose processor for the same reason.
If the rules have been relaxed in France it must have been very recently (last 6 months). I have been involved in several companies which have had to go to great expense to install _unencrypted_ leased lines into France as they couldn't just add the French offices to a worldwide VPN for encryption reasons.
-dp
I don't see how this is going to be accomplished. In some european contries, like France, encryption is illegal (unless you are the government of course) and in others restrictions are placed on it's use. For example in the UK it is actually illegal to do encryption in hardware. This dates back to the days where the implementation in software were too slow to be useful.
Thus if we have restrictions on internal use of encryption I don't see how we are going to develop and export strong encryption.
France has more oustanding european court cases against it that any other nation
-dp
However it is throttled to 8 Mb/s upstream. The vast majority of downstream bandwidth in Cambridge goes to feed personal computers in student rooms.
If Oxford have a more draconian policy on bandwidth allocation to students then they may well not need anything faster than 34 Mb/s. I suspect that a 34 Mb/s connect is vastly cheaper than a 155 Mb/s connect.
-dp
I accuse you sir of being a chatter (of the ucam variety). Or is Mr. R. Henson more wide spread than we believed.
You are confused.
Macrovision level 1 as used on VHS video is indeed part of the video signal.
Macrovision level 2 is implemented as a burst during the vertical blanking signal between PAL (or NTSC) frames which confuses the auto gain control circuit on a VCR. Since the vob files on the disk only contain an MPEG program stream. You couldn't physically put macrovison in the vob. The macrovison is generated by the digital to analogue converter (RAMDAC in the video world) attached to the MPEG decoder. This is why it is possible to hack many players not produce macrovison.
If this was not true I would not be able to watch DVD's on my LCD projector which is just as susceptable to macrovision (level 2) as VCR's are.
It is certainly true that the flag to tell the player to produce macrovision is in the vob but macrovison itself can't be.
-dp
It looks like disks recorded on this will not be playable on legacy DVD players. The article says:
It also mentions that the Video Recording Format could be made compliant with DVD video but hasn't, you can probably guess the reason why.
As for copy protection macrovision is added on output and is not encoded onto the disk. Thus using the analogue input (scrubbed of macrovision) it should be possible to copy disks. As far as I know no one has incorporated an analogue Copy Generation Management System into DVD. I assume that the digital system will let you make one digital copy of a disk then stop any further copies. The disks that have CSS would probably have the appropriate flags to stop digital output working on a player with this facility (non exist yet).
The article goes further:
I assume that this means that each DVD-RW contains a ID so that only their disks can be used. However I don't see how this could be used to prevent unorthorized copies (you just need to use their recorders and media).It is not at all clear if you could record digital TV (cable or satallite) in a digital manor onto these devices. Would the cable company for you to pay for the privilege of recording their programs or would they prevent it and force you to buy a device such as a ReplayTV ?
In short this is a technology that has been hamstrung by the movie studios.
-dp
The problem is that Napster only has one use, transfering MP3 files. I think that this is why it can't hide behind a "We just tell them where to go" defence. A more general system such as gnutella does not have these kind of problems. The problem with gnutella is that it doesn't preserve anonymity.
It is actually possible to transfer data between two points and to stop each end knowing what is going where but protect against it being spied upon or changed while in transit. However this tends to involve very hairy discrete mathematics and codeing a good implementation is beyond the vast majority of people.
Lets face it whatever we do industry are going to turn round and say you might be infringing our copyright so we are going to sue you. This is a fight that I do not think that we can win.
-dp
The problems with VB are roughly as follows
1) It is slow
2) It is proprietary
3) It only runs on windows
4) It isn't a real programming language
5) Did I mention that it is so slooooooooow
Although programs may look simpler to read and maintain in VB in reality it is just hiding the grottyness under a high level API. Because the libraries that it depends upon are precompiled you can't go and look what a function actually does when it doesn't act as expected.
It is slow, the language only recently began to be compiled. Pprevious attemps included embeding code+interpreter in a Windows PE format executable.
In a real language you can write a compiler for your language in that language and have the compiler compile itself. (This can be done with gcc) VB can't do this.
PERL is only hard to read if you don't know PERL, I find VB hard to read as it is verbose where as PERL is rather terse.
-dp
I wonder how DRE is planning on identifying napster users? I imagine placing random taps on ethernet sockets in colleges and regions that have cable modems would have civil liberties implications.
If napster is supid enough to keep logs of it users then they and napster deserve to have there ass sued off.
So are they planning on conducting dawn raids on any room with a computer?
I think that this is little more than a show suit designed to raise publicise his cause and scare few people.
-dp
This is slightly inaccurate. The StongARM was not an ARM design it was an evolution of the ARM designed in cooperation between Digital and ARM.
:)
When Digital was bought by Compaq the the rights to the processor reverted to ARM. Intel bought Digital's fabrication facilities unfortuately for them that entailed also having to make StrongARM and Alpha even though Intel have competing processors.
As for ARM being the next Intel this won't happen for two reason. First ARM is fabless and will remain that way. Second ARM's stated aim is to make money for it's employies in the best way it can rather than to dominate the processor market which is Intel's mission
-dp
DRI being availible under XFree 4 won't make a scrap of difference to support for the Geforce under Linux as Nvidia has stated that they won't use DRI but something "similar" for their OpenGL support.
This sucks and nvidia should be publically slaped for not supporting Linux.
-p
The NV15 NDA will lift on wednesday.
:)
I predict that the world+dog will post a preview/review at about 00:10 on wednesday morning. All these reviews at anandtech, firingsquad etc will be virtually identical being based upon the same press releases from nvidia
I wish people would put some thought in to there jounalism rather just spouting what the manufacturers tell them to.
-dp
The problem with private companes is they almost always look at the short term and try to make the most profit.
NASA has been forced to do this kind of thing in the past and what happened? The Challenger disaster.
Challenger blew up happened because the sub contractor for the solid fuel boosters decided to make them in several pieces instead of one like they had been before. The seals between the sections failed and the main fuel tank blew up.
This is what happens when private companys are involved on space, people die.
-dp
Wouldn't work.
Turning hubble towards earth would blind it. The instuments on board are so sensitive that the astronomers have to be very careful they don't look at either the earth or the sun.
-p
The show article mentions that it is going to tour to promote driving safety.
I can imagine a car that has so obvious connections with being driven dangerously (with the possible exception of the bat mobile). Then again that might be the point.
-dp
Hum. What about .bz ? Bzip makes use of the Burrows-Wheeler transform on blocks of data which are then are then compressed using more standard methods such as Huffman, Run Length Encoding and Arithmethic encoding.
The resulting alogrithm compresses much better than LZ77 + Shannon. It is so good that almost ever kernel image in the linux world is a bzImage.
Katz should be praised for his work but the world of compression has moved on and zip is no longer state of the art.
-dp
Oh Intel are keeping there tier one customers, such as Dell, happy. This is part of the reason that they can't supply the channel :)
-p
AMD are having supply problems for 1 month while getting Thunderbird and spitfire out the door.
This hardly compares with the months of supply problems compounded by bad architecture decisions that Intel has been having.
-dp
Um no.
When Hubble's mirror was ground the piece of glass was so large that the earths gravity affected its shape significantly. When the calculations for the mirror were made they had to take into account the difference in gravitential potential between oribit and ground level and correct for it.
The problem was that the correction was applied twice. The maths for this is actually quite complicated so don't be two hard on them.
-dp
Oh well we seem to have /.'d recourse.com
:)
500 Server Error
The hard access limit for this user has been reached
Hit him where it hurts, his companies wallet
-dp