The default modem internal modem in the 7500 is a winmodem. I expect this to be ture of most of the Dell's as the 7500 is near the top of their line. Granted they told me this straight out when it was ordered and recommended the global modem for linux users. Works like a champ. The only easier laptop I've setup was the older Nec Versa's, but they weren't real reliable and the heat rash on my lap was never comfortable:-).
I'm using an almost identical system right now (slower CPU and DVD to reduce battery consumption and 18 GB drive (The 75 is only if you have no battery or other drive installed so you can use all three bays)). Runs linux great, and the higher resolution is nice. Dell has had a website up with drivers for the screen, sound, ethernet card, modem.
The pointing device is an option. They can have either a joystick thingy or a pad. In fact I have a pad, but looking at the keyboard I can see the hardware for the joystick, so it may support both at the same time.
The only advice is to get the global modem, only use the Dell drivers for X if the ones on the internet don't work, and the newer ethernet cards need to have a few lines added to the PCMCIA config file (email me if you need them). A few hours after doing all that I had it doing everything I needed.
Collections and/or compilations of facts/data is not copyrightable. Works produced with value added or some sort of processing of this data may be, but this is a hazy area. There is one example of this that works well in highlighting the issues involved. The ARRL collects information on all the repeaters in the US, and then compiles a list that is sorted among header for find a local repeater of a specific type (band, beacon, etc.). This data is obtained from open sources (local repeater coordinators, requests to be included, etc.)
Now the ARRL makes a good bit of money on the sale of these little pocket books. For that reason they protect what they feel is an exclusive right to that data by claiming a copyright. Some people have come along and attempted to put some of the data from these books into electronic databases for easier access, and have been forced by the ARRL to stop and retract all of the data. While the people have been told time and time again that the ARRL does not have a good legal foundation for their case it is awfully hard for a college student to take them to court and challenge the claims.
The reason this is a hazy area is that the ARRL adds formating that might be considered creative content. Normally formating is also not protected under a copyright, but when that formatting is techinical data that is added by the compiler where to draw the line is less clear.
So if the collection/compilation of facts can be protected with copyrights, or something similiar, then the only thing that can be written without using fair use (which may also go away under many proposals) will be fiction, opinion, and legal briefs. Will children have to pay a royalty to do a report when they have to look up an atomic weight in the CRC handbook? The border of compliation versus creative work has worked well for many decades, along with practices of fair use. Crossing or eliminating these will make the things they need to protect less valuable in the end as not being able to use them reduces their need and demand.
How would this relate to common facts (history, statistics, etc.), universal constants (the value of e, pi, avogadro's number, etc.) and there use once they are included in a protected work? We never remember them, we always look them up. I know this legislation is being pushed by people like stock exchanges and sports organizations to protect numbers that they spend money on, but the effects of protecting them become chilling on everything else.
That is one drawback of the slashdot forums. There is no interface to a spelling checker, or in my case my better half. To keep near topic, I need to record a dictionary into my brain.:-).
Actually I thought I was highlighting the differences. Granted the firing properties of all the different types of nuerons are different then the simple sigmoid function used in software nets, but there are hardware versions that get much closer. What is interesting is the work that shows that behavioral equivalence can be found between the two types(analog hardware and software). My point is that I think it might not be possible to extend this to the biological equivalent without taking into account many other factors.
Penrose was the first, and most well known, person to suggest this. There was some evidence that some quantum events could effect parts of the brain in a way that the brain could measure/discern. I'm not sure anyone ever put any hard evidence towards the theories. Also when Penrose came out with that it brought a large number of people out of the woodwork.
On the other hand I've seen some math that shows that any nueral net can be mapped to a Determinisitic Finite State Machine. This means that if we were just a nueral net (I don't think that is all there is to the mind, as I said in my previous posting) then given a suffcient number of states, a machine could perfectly emulate a human mind.
The interconnection filtering and chemical processes (and whatever other things might be true, like quantum interaction) though would break the nueral net model for all but the most gross simulations of intelligent control. Nueral net developers simplify this by putting filters on the inputs and outputs, or simulating them in the net themselves. The question is then if a interconnect filter can be mathematically placed outside the net and still find a net with behavioral equivalence.
The overall problem though is that without a way of handling complex systems like the mind I'm not sure how much our simple abstractions can model the mind accuratly. In my opinion tools for handling/understanding complex systems is where the next step is in this field.
At the moment we don't really understand the storage and formation of concisnous and knowledge in the human brain. While the nuerons are well understood in a behavioral sense (firing under these conditions, etc.), the nuerons themselves are not the whole story. Each connection between them is a web upon itself with feedback onto the connection (not nueron feedback, that is different) which may be a filter, control, or even a crude analog storage using a filter. This doesn't even include how the chemical interactions might effect all of this.
While I don't think the brain is using anything esoteric (multi dimensions, quantum mechanics as has been expressed in some dubios theories), it does appear to be a collection of systems that evolved at different levels (filtering connections, nuerons, chemical interaction, etc...) to form one whole. I'm not sure much research is being done in handling the shear complexity of problems like this. Without being to handle complex systems we probably have no hope of understanding the brain other than our current crude approximations.
Actually reading the article I would say they gave things a far shake, and did highlight where there testing may have not been acurate. Conclusions are always subjective based each of our requirements. They are trying to fit theirs to the bais of an IT manager.
I know here in the US we have a number of organizations that try to protect the individual against abuses of power, but who helps in these cases (open source and right to reverse engineer, etc.) and what about outside the US? With the growth of profile for the open source movement this is going to happen a lot more before it gets better. Large companies often would rather lobby, legislate and litigate then change a flawed technology (like the one in this case or scanners for cell phones) or bussiness plan.
Is there a list of organizations that can be supported, promoted, and/or contacted for these issues?
Actually I've never read "The Code Book" ( I've actually never heard of it actually), but have been following quantum computing in the mainstream science press (Science News, NPR, etc). Currently there are demonstrated algorithms for factoring small numbers (five bits has been done and 11 bits might have been by now). The base algorithms are in place and extensions to them due to the inherent parallelism of qubits for problems like this makes the factoring of larger number just a matter of being able to handle more qubits.
There are people out there using all sorts of esoteric machines to make quantum gates. At the moment a hundred gates is a large infrastructure, but with advances like the quantum resivor from Lucent these could be done in a LSI type circuit in the not to distant future.
Fortunatly no one has shown equivalent work for Fiestel (sp?) networks that most symmetric block ciphers are based on, and stream ciphers tend also be safe if used properly. I just haven't seen a PKI that doesn't have something on the horizon that break it.
Judging from the advances in quantum computing and the algorithms for factoring using qubits (sp?) I can see it expiring and then just as people are adopting it wholesale we see a set of factoring breakthroughs. It not like you can increase the key length all that much either. How long would it take to encrypt a session key with a 65535 RSA key:-)?
Maybe we should start looking at that IBM algorithm that they claim is provably difficult.
Except we called our domicile the lab/office, our monthly salary never doubled, and we couldn't afford food let alone to have it delivered. Ok, so we left to go happy hour to eat, but that was it, honest.
In ten years I wouldn't doubt seeing a CPU with just a handful of pins (more then a dozen, but not much more). We've just about pushed parallel buses to the limit of DC signalling. Hell, my motherboards now have a spread spectrum signalling option to reduce RFI. We can maybe double the clock speed, but that would require a good bit of work.
With serial we can actually use RF. Layout the motherboard in stripline to filter the signal between components. The IC runs parallel to an interface section, which is a high speed shift register. If you wanted parallel busses you could then adda pin or two, modulate the signal up to another frequency, and/or spread it out with a Direct Sequence spreader.
Then when we max out the viable speed of the serial bus we will aggregate them together having learned as much about serial bus implementation as we had parallel. In the end a happy mix between the two will be found.
The client can't control anything like that. Basically the laws of physics have to be implemented on the server, or the game play will look like something from "The Matrix".
I've been thinking a lot of how to make an aimbot impratical. It reall boils down to limiting the control responses (movement of the players) and increasing the external stimulous to the system. If the movement was limited to what a star player should be expected to do, then maybe we could design the response to external disruptions (being hit) such that they would throw a classic linear control system off enough to make it non competitive. This is an inverse control theory problem. Find a settling time and characteristic for classical control systems that makes the aimbot suck. Then change the game enough to make sure this happens.
If the cheater then wants to use a non linear control system, or some really good linear multivariable control system let them develop it. With the right plant to design for this could be made into something that if someone implemented the program for it I'd be willing to let them use it just from the impressivness of the work.
I only meant using encryption for the object tracking and inventory problem. The signature would be used to verify from the server, back to the server that the client really does have what he says he has.
When it comes to computerized aiming, and target tracking, I'm not sure there is a way around this other then sending a spoiler input to the aiming control system, or recognizing it based on the response.
Maybe this is something more like what postal chess does. Have everyone register to get on a server by giving personal info, and then boot them of if there is a problem. To make it work might be slow and intrusive. Then again I always liked just playing against the computer much better.
I can think of a number of ways around that one, but if the motion (rotation included) is controlled by the server then all the client could do is implement a control system to track the target.
The better solution for that may be logging and fingerprinting. While I haven't seen this data, almost all other I've seen you can easily tell the difference between human and machine control signals. Develop a fingerprint for control signals. If it looks to much like a control system output, send it spoiler data that would be out of control bounds. Or you can just log them off:-).
While I realize this is a record/record player problem (as in GEB), but the breaking record can be made unattractivly hard to make compared to the effort for the player.
I have spent much of the last year developing systems/protocols for hostile client to trusted server connections. There are ways to do this, but it requires that the base protocol be designed with it in mind. I don't know the quake protocol, but I will try to describe a possible method.
The client connects to the server with a request for a unique ID. What comes back is a two part ID, one public, one private. The client then makes universe change requests (movement of client in the universe, firing, etc...) with the pub ID, request, and a hash of the above with the priv ID. The server then can verify it's the same person that requested the ID (PKI can be used to send the intial ID back if snooping is a problem). The server sends back universe updates along with the verification. If you want these can be signed with the priv ID. If you want to do object tracking then the objects could have seperate signatures so they are unique and verifiable. Ammo could also be tracked with a sub object or something like that.
Basically any rule that you don't want broken need to handled on the server in this model.
Encryption could be cut down to a low level to prevent computaional slowness and export problems. Even a mild algorithm would be acceptable so long as the key secure until the end of the game against a single PC.
If you want to do peer to peer, or move more handling back to the client, then one would have to look into one of the many blind poker algorithms, but it to should be doable.
Closed source doesn't make any code less hackable. All it does is make a protocol not designed against a malicous client effective against those not willing to go through any effort in hacking.
The real solution for this is to make the protocol in a way so the client can only make requests to the server. Any time the client describes itself to the server, those things that can be described can not be trusted. In this case a safer protocol is to have the client request motion. The server will then provide updated info back to the client. If you want the client to track objects, then you can cryptologically sign them, so theywould be unique to the game session and non repeatable. The crypto could use very small keys to keep the performance managable, and the game exportable. 32 bits would probably be enough.
They "tuned the cache" to the point that none of the servers went to disk for their workload. That right there gave NT an advantage. In a real world scenario I'll be out to the disk all the time. I'll be there to get my data during misses, to write my state info for transactions, to log things. How much logging was the NT server doing? If it wasn't a lot then they took out the disk subsystem from the equation. The whole reason they used the 2.0 kernel was fishy at best also. 2.2 TCP stack broke communications with their win 95 clients? Finally, why on earth would I want to do hundreds or thousands of SSL transactions in software. If you are doing more then a few a second you really need a hardware SSL brick or card, which works with all the tested platforms. These people obviously don't understand but one set of solutions.
If I need more comp power then a 68HC11/12 or a 68332, then I'm going to have a good bit of memory (read over 2 MB). The 68HC11/12 was never designed as a computational platform (the 6502 and family was actually not a Motorola part and in no way related to 68xx family or 68K family). That's were you move up to the CORE32 family like the 68332. It can then address 1 MB on up to 16 MB (and more on ceratin versions). There are tons of OSs that run in that space, even Linux. Linux is also quite useful for embedded applications in that space.
The 386EX and other IA-32 embedded solutions are almost always implemented with a meg or more of memory. This could easily run LinuxRT, or one of the many other RTOSs. Plus the embedded IA-32 equipment is expensive by comparison to more traditional microcontrollers.
The Hubble and most of the other space applications are more involved in why they choose those CPUs. First, they have to be RAD hard. ULSI CPUs (pentium and above) aren't there yet. Second, power consumption plays a huge roll. Running a CPU that takes 9W standby is pricy in space. Third, the low voltage stuff doesn't work as well in those environments due to grounds and radiation. NASA chooses Intel over Motorola often (atleast the people I've worked with) because of development tools investment (pSOS, QNX, Wind River) that they have, and that all the boards are custom makes anyway, so the hardware requirements are that it only needs to be survivable in those environments.
As far as the GPL goes, they can probably do what the Tivo does, release it on a web page.
I used the first version of Linux that Linus put on the net, and was really early adopter of Minix 386 (remember the arguement between Tannebaum and Linus? Funny how that worked out.) Linux was much rougher then, but it was much more closer to an OS the V2OS. Linux has always had the layer of abstraction that allowed it to grow into what it was today. In fact I believe it is the fact that linux had some key features missing in Minix that made almost everyone switch development over.
V2OS is really a real time exec. Granted it looks like it has some very nice features, and a cleanly thought out design, but there are a lot of these out there. Plus, if I wanted a stripped RTE I'd want it to run on something smaller then a IA-32 (I've got one on a 6811 that I've used for years). Plus I can get Linux on a 68332 if I need a boomer of a microcontroller.
I'm missing what the target audience is supposed to be I guess. It seems to be sitting in the same place as some of QNX and Wind River's products, but it is a long way away from being as useful.
This OS will be of no real use to a beowulf. It's lacking high bandwidth networking (bonding, fast NICs, etc...) and the virtualization needed to port PVM or MPI (the heart of any beowulf) to it. It's targetd more towards the embeded application. Media center is one thing that comes to mind. It should be good with MP3s or MPEGs due to the simple server architecture, but I don't know what the switching latency is so that may be all you can do on it at one time. Of course adding a little more hardware and you can do the same thing on a full OS.
That was true until a few weeks ago. I got the notice that NT4 with SP6, and 2000 were now certified for both orange and red book. So you can now add the NIC back in. I know there were probably exceptions made, and problems overlooked, but they have the certification, which is a problem for linux, because that can now be used by the NT camp.
Slashdot Mirror
The default modem internal modem in the 7500 is a winmodem. I expect this to be ture of most of the Dell's as the 7500 is near the top of their line. Granted they told me this straight out when it was ordered and recommended the global modem for linux users. Works like a champ. The only easier laptop I've setup was the older Nec Versa's, but they weren't real reliable and the heat rash on my lap was never comfortable :-).
I'm using an almost identical system right now (slower CPU and DVD to reduce battery consumption and 18 GB drive (The 75 is only if you have no battery or other drive installed so you can use all three bays)). Runs linux great, and the higher resolution is nice. Dell has had a website up with drivers for the screen, sound, ethernet card, modem.
The pointing device is an option. They can have either a joystick thingy or a pad. In fact I have a pad, but looking at the keyboard I can see the hardware for the joystick, so it may support both at the same time.
The only advice is to get the global modem, only use the Dell drivers for X if the ones on the internet don't work, and the newer ethernet cards need to have a few lines added to the PCMCIA config file (email me if you need them). A few hours after doing all that I had it doing everything I needed.
Collections and/or compilations of facts/data is not copyrightable. Works produced with value added or some sort of processing of this data may be, but this is a hazy area. There is one example of this that works well in highlighting the issues involved. The ARRL collects information on all the repeaters in the US, and then compiles a list that is sorted among header for find a local repeater of a specific type (band, beacon, etc.). This data is obtained from open sources (local repeater coordinators, requests to be included, etc.)
Now the ARRL makes a good bit of money on the sale of these little pocket books. For that reason they protect what they feel is an exclusive right to that data by claiming a copyright. Some people have come along and attempted to put some of the data from these books into electronic databases for easier access, and have been forced by the ARRL to stop and retract all of the data. While the people have been told time and time again that the ARRL does not have a good legal foundation for their case it is awfully hard for a college student to take them to court and challenge the claims.
The reason this is a hazy area is that the ARRL adds formating that might be considered creative content. Normally formating is also not protected under a copyright, but when that formatting is techinical data that is added by the compiler where to draw the line is less clear.
So if the collection/compilation of facts can be protected with copyrights, or something similiar, then the only thing that can be written without using fair use (which may also go away under many proposals) will be fiction, opinion, and legal briefs. Will children have to pay a royalty to do a report when they have to look up an atomic weight in the CRC handbook? The border of compliation versus creative work has worked well for many decades, along with practices of fair use. Crossing or eliminating these will make the things they need to protect less valuable in the end as not being able to use them reduces their need and demand.
How would this relate to common facts (history, statistics, etc.), universal constants (the value of e, pi, avogadro's number, etc.) and there use once they are included in a protected work? We never remember them, we always look them up. I know this legislation is being pushed by people like stock exchanges and sports organizations to protect numbers that they spend money on, but the effects of protecting them become chilling on everything else.
That is one drawback of the slashdot forums. There is no interface to a spelling checker, or in my case my better half. To keep near topic, I need to record a dictionary into my brain. :-).
Actually I thought I was highlighting the differences. Granted the firing properties of all the different types of nuerons are different then the simple sigmoid function used in software nets, but there are hardware versions that get much closer. What is interesting is the work that shows that behavioral equivalence can be found between the two types(analog hardware and software). My point is that I think it might not be possible to extend this to the biological equivalent without taking into account many other factors.
Penrose was the first, and most well known, person to suggest this. There was some evidence that some quantum events could effect parts of the brain in a way that the brain could measure/discern. I'm not sure anyone ever put any hard evidence towards the theories. Also when Penrose came out with that it brought a large number of people out of the woodwork.
On the other hand I've seen some math that shows that any nueral net can be mapped to a Determinisitic Finite State Machine. This means that if we were just a nueral net (I don't think that is all there is to the mind, as I said in my previous posting) then given a suffcient number of states, a machine could perfectly emulate a human mind.
The interconnection filtering and chemical processes (and whatever other things might be true, like quantum interaction) though would break the nueral net model for all but the most gross simulations of intelligent control. Nueral net developers simplify this by putting filters on the inputs and outputs, or simulating them in the net themselves. The question is then if a interconnect filter can be mathematically placed outside the net and still find a net with behavioral equivalence.
The overall problem though is that without a way of handling complex systems like the mind I'm not sure how much our simple abstractions can model the mind accuratly. In my opinion tools for handling/understanding complex systems is where the next step is in this field.
At the moment we don't really understand the storage and formation of concisnous and knowledge in the human brain. While the nuerons are well understood in a behavioral sense (firing under these conditions, etc.), the nuerons themselves are not the whole story. Each connection between them is a web upon itself with feedback onto the connection (not nueron feedback, that is different) which may be a filter, control, or even a crude analog storage using a filter. This doesn't even include how the chemical interactions might effect all of this.
While I don't think the brain is using anything esoteric (multi dimensions, quantum mechanics as has been expressed in some dubios theories), it does appear to be a collection of systems that evolved at different levels (filtering connections, nuerons, chemical interaction, etc...) to form one whole. I'm not sure much research is being done in handling the shear complexity of problems like this. Without being to handle complex systems we probably have no hope of understanding the brain other than our current crude approximations.
Is this the launch of a new line of books,
:-).
"Benchmarks from Dummies"
Actually reading the article I would say they gave things a far shake, and did highlight where there testing may have not been acurate. Conclusions are always subjective based each of our requirements. They are trying to fit theirs to the bais of an IT manager.
I know here in the US we have a number of organizations that try to protect the individual against abuses of power, but who helps in these cases (open source and right to reverse engineer, etc.) and what about outside the US? With the growth of profile for the open source movement this is going to happen a lot more before it gets better. Large companies often would rather lobby, legislate and litigate then change a flawed technology (like the one in this case or scanners for cell phones) or bussiness plan.
Is there a list of organizations that can be supported, promoted, and/or contacted for these issues?
Actually I've never read "The Code Book" ( I've actually never heard of it actually), but have been following quantum computing in the mainstream science press (Science News, NPR, etc). Currently there are demonstrated algorithms for factoring small numbers (five bits has been done and 11 bits might have been by now). The base algorithms are in place and extensions to them due to the inherent parallelism of qubits for problems like this makes the factoring of larger number just a matter of being able to handle more qubits.
There are people out there using all sorts of esoteric machines to make quantum gates. At the moment a hundred gates is a large infrastructure, but with advances like the quantum resivor from Lucent these could be done in a LSI type circuit in the not to distant future.
Fortunatly no one has shown equivalent work for Fiestel (sp?) networks that most symmetric block ciphers are based on, and stream ciphers tend also be safe if used properly. I just haven't seen a PKI that doesn't have something on the horizon that break it.
Judging from the advances in quantum computing and the algorithms for factoring using qubits (sp?) I can see it expiring and then just as people are adopting it wholesale we see a set of factoring breakthroughs. It not like you can increase the key length all that much either. How long would it take to encrypt a session key with a 65535 RSA key :-)?
Maybe we should start looking at that IBM algorithm that they claim is provably difficult.
Except we called our domicile the lab/office, our monthly salary never doubled, and we couldn't afford food let alone to have it delivered. Ok, so we left to go happy hour to eat, but that was it, honest.
In ten years I wouldn't doubt seeing a CPU with just a handful of pins (more then a dozen, but not much more). We've just about pushed parallel buses to the limit of DC signalling. Hell, my motherboards now have a spread spectrum signalling option to reduce RFI. We can maybe double the clock speed, but that would require a good bit of work.
With serial we can actually use RF. Layout the motherboard in stripline to filter the signal between components. The IC runs parallel to an interface section, which is a high speed shift register. If you wanted parallel busses you could then adda pin or two, modulate the signal up to another frequency, and/or spread it out with a Direct Sequence spreader.
Then when we max out the viable speed of the serial bus we will aggregate them together having learned as much about serial bus implementation as we had parallel. In the end a happy mix between the two will be found.
The client can't control anything like that. Basically the laws of physics have to be implemented on the server, or the game play will look like something from "The Matrix".
I've been thinking a lot of how to make an aimbot impratical. It reall boils down to limiting the control responses (movement of the players) and increasing the external stimulous to the system. If the movement was limited to what a star player should be expected to do, then maybe we could design the response to external disruptions (being hit) such that they would throw a classic linear control system off enough to make it non competitive. This is an inverse control theory problem. Find a settling time and characteristic for classical control systems that makes the aimbot suck. Then change the game enough to make sure this happens.
If the cheater then wants to use a non linear control system, or some really good linear multivariable control system let them develop it. With the right plant to design for this could be made into something that if someone implemented the program for it I'd be willing to let them use it just from the impressivness of the work.
I only meant using encryption for the object tracking and inventory problem. The signature would be used to verify from the server, back to the server that the client really does have what he says he has.
When it comes to computerized aiming, and target tracking, I'm not sure there is a way around this other then sending a spoiler input to the aiming control system, or recognizing it based on the response.
Maybe this is something more like what postal chess does. Have everyone register to get on a server by giving personal info, and then boot them of if there is a problem. To make it work might be slow and intrusive. Then again I always liked just playing against the computer much better.
I can think of a number of ways around that one, but if the motion (rotation included) is controlled by the server then all the client could do is implement a control system to track the target.
:-).
The better solution for that may be logging and fingerprinting. While I haven't seen this data, almost all other I've seen you can easily tell the difference between human and machine control signals. Develop a fingerprint for control signals. If it looks to much like a control system output, send it spoiler data that would be out of control bounds. Or you can just log them off
While I realize this is a record/record player problem (as in GEB), but the breaking record can be made unattractivly hard to make compared to the effort for the player.
I have spent much of the last year developing systems/protocols for hostile client to trusted server connections. There are ways to do this, but it requires that the base protocol be designed with it in mind. I don't know the quake protocol, but I will try to describe a possible method.
The client connects to the server with a request for a unique ID. What comes back is a two part ID, one public, one private. The client then makes universe change requests (movement of client in the universe, firing, etc...) with the pub ID, request, and a hash of the above with the priv ID. The server then can verify it's the same person that requested the ID (PKI can be used to send the intial ID back if snooping is a problem). The server sends back universe updates along with the verification. If you want these can be signed with the priv ID.
If you want to do object tracking then the objects could have seperate signatures so they are unique and verifiable. Ammo could also be tracked with a sub object or something like that.
Basically any rule that you don't want broken need to handled on the server in this model.
Encryption could be cut down to a low level to prevent computaional slowness and export problems. Even a mild algorithm would be acceptable so long as the key secure until the end of the game against a single PC.
If you want to do peer to peer, or move more handling back to the client, then one would have to look into one of the many blind poker algorithms, but it to should be doable.
Closed source doesn't make any code less hackable. All it does is make a protocol not designed against a malicous client effective against those not willing to go through any effort in hacking.
The real solution for this is to make the protocol in a way so the client can only make requests to the server. Any time the client describes itself to the server, those things that can be described can not be trusted. In this case a safer protocol is to have the client request motion. The server will then provide updated info back to the client. If you want the client to track objects, then you can cryptologically sign them, so theywould be unique to the game session and non repeatable. The crypto could use very small keys to keep the performance managable, and the game exportable. 32 bits would probably be enough.
They "tuned the cache" to the point that none of the servers went to disk for their workload. That right there gave NT an advantage. In a real world scenario I'll be out to the disk all the time. I'll be there to get my data during misses, to write my state info for transactions, to log things.
How much logging was the NT server doing? If it wasn't a lot then they took out the disk subsystem from the equation.
The whole reason they used the 2.0 kernel was fishy at best also. 2.2 TCP stack broke communications with their win 95 clients?
Finally, why on earth would I want to do hundreds or thousands of SSL transactions in software. If you are doing more then a few a second you really need a hardware SSL brick or card, which works with all the tested platforms. These people obviously don't understand but one set of solutions.
If I need more comp power then a 68HC11/12 or a 68332, then I'm going to have a good bit of memory (read over 2 MB). The 68HC11/12 was never designed as a computational platform (the 6502 and family was actually not a Motorola part and in no way related to 68xx family or 68K family). That's were you move up to the CORE32 family like the 68332. It can then address 1 MB on up to 16 MB (and more on ceratin versions). There are tons of OSs that run in that space, even Linux. Linux is also quite useful for embedded applications in that space.
The 386EX and other IA-32 embedded solutions are almost always implemented with a meg or more of memory. This could easily run LinuxRT, or one of the many other RTOSs. Plus the embedded IA-32 equipment is expensive by comparison to more traditional microcontrollers.
The Hubble and most of the other space applications are more involved in why they choose those CPUs. First, they have to be RAD hard. ULSI CPUs (pentium and above) aren't there yet. Second, power consumption plays a huge roll. Running a CPU that takes 9W standby is pricy in space. Third, the low voltage stuff doesn't work as well in those environments due to grounds and radiation. NASA chooses Intel over Motorola often (atleast the people I've worked with) because of development tools investment (pSOS, QNX, Wind River) that they have, and that all the boards are custom makes anyway, so the hardware requirements are that it only needs to be survivable in those environments.
As far as the GPL goes, they can probably do what the Tivo does, release it on a web page.
I used the first version of Linux that Linus put on the net, and was really early adopter of Minix 386 (remember the arguement between Tannebaum and Linus? Funny how that worked out.) Linux was much rougher then, but it was much more closer to an OS the V2OS. Linux has always had the layer of abstraction that allowed it to grow into what it was today. In fact I believe it is the fact that linux had some key features missing in Minix that made almost everyone switch development over.
V2OS is really a real time exec. Granted it looks like it has some very nice features, and a cleanly thought out design, but there are a lot of these out there. Plus, if I wanted a stripped RTE I'd want it to run on something smaller then a IA-32 (I've got one on a 6811 that I've used for years). Plus I can get Linux on a 68332 if I need a boomer of a microcontroller.
I'm missing what the target audience is supposed to be I guess. It seems to be sitting in the same place as some of QNX and Wind River's products, but it is a long way away from being as useful.
This OS will be of no real use to a beowulf. It's lacking high bandwidth networking (bonding, fast NICs, etc...) and the virtualization needed to port PVM or MPI (the heart of any beowulf) to it.
It's targetd more towards the embeded application. Media center is one thing that comes to mind. It should be good with MP3s or MPEGs due to the simple server architecture, but I don't know what the switching latency is so that may be all you can do on it at one time. Of course adding a little more hardware and you can do the same thing on a full OS.
That was true until a few weeks ago. I got the notice that NT4 with SP6, and 2000 were now certified for both orange and red book. So you can now add the NIC back in. I know there were probably exceptions made, and problems overlooked, but they have the certification, which is a problem for linux, because that can now be used by the NT camp.