The way they do it is to implement a small real-time kernel, then run the linux kernel as a real-time process under that kernel. Then, userspace, non-RT linux process can communicate with the RT processes via either a mem-locked shared memory segment or a FIFO.
>This patent appears more obvious in hindsight >than it was when Amazon first began the practice. >Take the brick and mortar world >for instance. Show me a store anywhere that lets >you just walk in, pick up some merchandise, and >walk out the door (legally).
1) Our company has an account with the local electronic supply store. We go and ask for something, he gets it and gives it to us, we sign a notebook at leave. He sends us a bill. Not exactly the same thing, but close.
2) If somebody *did* do that offline, it wouldn't be patentable either.
This is bullshit. All the hardcore pseudo-capitalists who disclaim social responsibility can go to hell, as far as I am concerned. Whatever your economics 100 professer may have spewed, this is an unescapable fact. Corporations are responsible for acting in the best interests of their shareholders withing the boundaries set by laws, the government, and ethical behavior. It is (or should be) the responsibility of the government to set and enforce those boundaries in a manner which is optimial for the citizens at large.
That is why the patent system exists in the first place -- As Jeff and Tim agreed, it should prevent corporations from using their money and power to take the innovations of a smaller company or individual and smoother them with superior marketing etc, thus preventing further innovation by said small company.
Unfortunately, it has come to the point where large companies use patents and cross-licensing to prevent small startups to innovate without stepping on the patents of the "big boys".
It is somewhat refreshing seeing a relatively small company using patents to defend itself against larger corporations, the problem is that neither the 1-click or the affiliate programs are (or should be) patentable. Patents are supposed to prevent technical innovations, *not* marketing strategies. Amazons greatest, and fundamental innovations are all marketing, none techinical. The patent process never has, nor was intended to, protect such things, and is not intended to.
The real telling fact of these is that neither of them would be considered interesting, much less patentable offline. They are equivelant to patenting banner ads--people have been putting ads on every available surface for years now, there is no reason that putting them on a web-page should be patentable.
Basically, neither of those patents should have ever been allowed to be granted, and the USPO should be taking as much flak for this as Amazon.
Well, notice he did not provide the BIOS, nor advocate finding a friend to pull it from. Instead, he pointed you to where the BIOS was a free download on Promises website, and their own flash utility will update without making any modifications. Remeber, he says flash before you solder, so it isn't even as if you have to "fake it out" into flashing the RAID bios onto the UDMA66 board. In my book, that falls neatly into the category of "if you didn't want me to do that, don't give it to me".
A configuration that I have seen and used a lot (even in very high end HP systems) is to put system+software+swap on a pair of software mirrored internal disks, then use an extrnal hardware RAID for data.
I wouldn't use software RAID 5 on any non-Linux system, nor would I use it in a critical situation. But for internal web/mail/fileserver type tasks, I would use it in a minute.
I am looking at setting up 4 7200 RPM IDE disks with software RAID at home, just to play with it, but that is another issue entirely.
While the calculations can take time, a checksum is pretty fast, and hardware checksumming is faster. Where you take your hit is that when doing small random writes, you have to read from the other disks, calculate the checksum, and write back to the data and parity disks. This can turn 1 IO into 5 IO/s, and cause your system suck. Intelligent software/hardware can help this out considerably, both by buffering and by coalescing writes, but there is only so much you can do.
Uh... no, actually it does. When I buy a CD, revenues are divided between the producer and the retailer in some fashion. Then, the portion of the retailers revenue that is profit is divided among its owners. If one of those owners happens to be a record company, the record company gets another cut.
Better yet, pattent "EULAs that screw over customers and absolve the company from any responsibility"... Then software companies need to either a) license the technology from you, or B) open up their EULAs to be more reasonable...
I definately agree. I think people who learn procedural programming first make better OO programmers in the long run--they are much more likely to understand the benefits, limitation, and tradeoffs associated with using those OO features that people who start off in C++ overuse, or use poorly.
Personally, the thing I dislike about C++ (and Java even more so) is that it turns too many warnings into errors. Computers exist for me to tell what to do, not the other way around, and if I write something that is syntactically correct and well defined, a compiler has no business telling me I can't do it. Yes, I even thing that accessing a private member of a class should only be a warning... If I have a legitimate cause to use a classes private variables, don't make me "#define private public"
The HP J5000 Visualize workstation (dual 440 MHz CPUs) weigh 75 lbs and draw up to 1500 watts power supply rating, at least... I don't know how much they actually use). Makes sun look downright energy efficient.
Yeah, but this is just like CSS in that any one of (potentially) hundreds of keys can decrypt the signal. Every display would have one of these codes hard-wired into it, and presumably the FBI would have a key to use with their tempest gear.
Also, like CSS, this is one of those technologies that makes it easy to "record, store, playback to authorized device" without having to break the encryption, but difficult for a small non-licensee to build or write a display system that can use the format.
Though every rev. of the chip may have a slightly different underlying archetecture. In fact, the two chips they announced *do* have different archetectures, and therefore different code-morphing engines. For the simple reason that if they released that, people would write "to the hardware" and they would be stuck in the same old backwards hardware compatability morass that caused the x86 mess in the first place is enough to leave the code-morphing proprietary.
It takes about as much time to drive out to CompUSA and buy the game as it does to download over my 256K link, and I can surf/. while I do it. It less time to order the game, but it takes longer to get it. I, of course, downloaded the demo AND ordered the game (hadn't been paying enough attention to see that it was released)
HOMMII was one of the very, very few games I found it worth rebooting into W95 to play.
Of course, Loki has open sourced quite a bit of software, including their installer and libraries they use to port games. ie. code that someone else actually could use (Unreal Tournament for Linux uses the Loki installer).
But yeah, there is a double standard for games:) Games rock:)
Loki games are licensed by Loki from the companies that write the original, and thus are not Open Source. Their Linux based installer, which they wrote, is (I belive) Open Source, and was used by the Linux version of Unreal Tournament.
That is simply untrue. The cost of cycles is not 0. Consider that it takes longer to launch word 95 on a PII XYZ (whatever they are at these days) than it took to launch wordstar 5 on a 286. It also takes longer to load a file, save a file, reformat a file, and they typing lag is worse. And I will let you in on a little secret: There is almost no functionality missing in Wordstar -- just some of it is harder to get to (but faster to use once you know how!)
Now, these days, people (evidentally) accept the increased latency in exchange for the ease of use of Word, but my point is, it isn't free. It isn't even cheap.
Memory and disk are essentially free, but their bandwidth is not, and usually large files == long load times.
What they can do is effectively open development of LiViD for the purposes of creating a high-quality, free UNIX based DVD player. It just the illegal uses that are harder to curtail.
Since NTSC is an analog standard it has no defined horizontal resolution (it doesn't even have pixels) -- it just depends on how often you sample a given scan line. However, bandwidth, broadcaster, and CRT limitations keep the practical limit in the 500-700 range. The number of scanlines is, however, fixed so you always have the same number of scanlines (otherwise the TV wouldn't know where to display them).
Not only that, but since nothign prevents you from making a bit-by-bit copy of the original, encrypted DVD which can then be played by existing software and (if you managed to put it on another DVD) hardware, cracking the DVD is only good for either A) converting it to another format, or B) playing it (just a specialized form of A where the target format is a video stream). Anyone who can manage the 6-8 GB storage requirements can pirate DVD w/o DeCSS.
Yes, exactly. You have hit the nail on the head. It explicitly falls under fair use to reverse engineer/modify software for the purpose of making it work with your system. This was decided years ago in some of the early computer software copyright lawsuits.
Netscape on UNIX has this option ("Only accept cookies originating from the same server as the page being viewed"), and has since at least version 3.0.
Netscape on Windows has an option in the same place called "Only accept cookies that are sent to originating server" -- I don't know if this means don't accept.co.uk or other idiotic domains (which would break things like sharing a login between www.yahoo.com and quote.yahoo.com), or if it is just dumbspeak for the same thing NS UNIX supports (I suspect it is the latter). I don't use windows much, so I haven't investigated.
On Linux, using netscape, I haven't seen a cookie from doubleclick in over a year (I prune my cookie file regularly as well)
Slashdot Mirror
The way they do it is to implement a small real-time kernel, then run the linux kernel as a real-time process under that kernel. Then, userspace, non-RT linux process can communicate with the RT processes via either a mem-locked shared memory segment or a FIFO.
>This patent appears more obvious in hindsight >than it was when Amazon first began the practice. >Take the brick and mortar world
>for instance. Show me a store anywhere that lets >you just walk in, pick up some merchandise, and >walk out the door (legally).
1) Our company has an account with the local electronic supply store. We go and ask for something, he gets it and gives it to us, we sign a notebook at leave. He sends us a bill. Not exactly the same thing, but close.
2) If somebody *did* do that offline, it wouldn't be patentable either.
This is bullshit. All the hardcore pseudo-capitalists who disclaim social responsibility can go to hell, as far as I am concerned. Whatever your economics 100 professer may have spewed, this is an unescapable fact. Corporations are responsible for acting in the best interests of their shareholders withing the boundaries set by laws, the government, and ethical behavior. It is (or should be) the responsibility of the government to set and enforce those boundaries in a manner which is optimial for the citizens at large.
That is why the patent system exists in the first place -- As Jeff and Tim agreed, it should prevent corporations from using their money and power to take the innovations of a smaller company or individual and smoother them with superior marketing etc, thus preventing further innovation by said small company.
Unfortunately, it has come to the point where large companies use patents and cross-licensing to prevent small startups to innovate without stepping on the patents of the "big boys".
It is somewhat refreshing seeing a relatively small company using patents to defend itself against larger corporations, the problem is that neither the 1-click or the affiliate programs are (or should be) patentable. Patents are supposed to prevent technical innovations, *not* marketing strategies. Amazons greatest, and fundamental innovations are all marketing, none techinical. The patent process never has, nor was intended to, protect such things, and is not intended to.
The real telling fact of these is that neither of them would be considered interesting, much less patentable offline. They are equivelant to patenting banner ads--people have been putting ads on every available surface for years now, there is no reason that putting them on a web-page should be patentable.
Basically, neither of those patents should have ever been allowed to be granted, and the USPO should be taking as much flak for this as Amazon.
Well, notice he did not provide the BIOS, nor advocate finding a friend to pull it from. Instead, he pointed you to where the BIOS was a free download on Promises website, and their own flash utility will update without making any modifications. Remeber, he says flash before you solder, so it isn't even as if you have to "fake it out" into flashing the RAID bios onto the UDMA66 board. In my book, that falls neatly into the category of "if you didn't want me to do that, don't give it to me".
A configuration that I have seen and used a lot (even in very high end HP systems) is to put system+software+swap on a pair of software mirrored internal disks, then use an extrnal hardware RAID for data.
I wouldn't use software RAID 5 on any non-Linux system, nor would I use it in a critical situation. But for internal web/mail/fileserver type tasks, I would use it in a minute.
I am looking at setting up 4 7200 RPM IDE disks with software RAID at home, just to play with it, but that is another issue entirely.
While the calculations can take time, a checksum is pretty fast, and hardware checksumming is faster. Where you take your hit is that when doing small random writes, you have to read from the other disks, calculate the checksum, and write back to the data and parity disks. This can turn 1 IO into 5 IO/s, and cause your system suck. Intelligent software/hardware can help this out considerably, both by buffering and by coalescing writes, but there is only so much you can do.
Uh... no, actually it does. When I buy a CD, revenues are divided between the producer and the retailer in some fashion. Then, the portion of the retailers revenue that is profit is divided among its owners. If one of those owners happens to be a record company, the record company gets another cut.
Well, it is slow because they use 1024x768 full color jpegs, aparently compressed at a relatively high quality setting.
Better yet, pattent "EULAs that screw over customers and absolve the company from any responsibility"... Then software companies need to either a) license the technology from you, or B) open up their EULAs to be more reasonable...
I definately agree. I think people who learn procedural programming first make better OO programmers in the long run--they are much more likely to understand the benefits, limitation, and tradeoffs associated with using those OO features that people who start off in C++ overuse, or use poorly.
Personally, the thing I dislike about C++ (and Java even more so) is that it turns too many warnings into errors. Computers exist for me to tell what to do, not the other way around, and if I write something that is syntactically correct and well defined, a compiler has no business telling me I can't do it. Yes, I even thing that accessing a private member of a class should only be a warning... If I have a legitimate cause to use a classes private variables, don't make me "#define private public"
The HP J5000 Visualize workstation (dual 440 MHz CPUs) weigh 75 lbs and draw up to 1500 watts power supply rating, at least... I don't know how much they actually use). Makes sun look downright energy efficient.
Yeah, but this is just like CSS in that any one of (potentially) hundreds of keys can decrypt the signal. Every display would have one of these codes hard-wired into it, and presumably the FBI would have a key to use with their tempest gear.
Also, like CSS, this is one of those technologies that makes it easy to "record, store, playback to authorized device" without having to break the encryption, but difficult for a small non-licensee to build or write a display system that can use the format.
Though every rev. of the chip may have a slightly different underlying archetecture. In fact, the two chips they announced *do* have different archetectures, and therefore different code-morphing engines. For the simple reason that if they released that, people would write "to the hardware" and they would be stuck in the same old backwards hardware compatability morass that caused the x86 mess in the first place is enough to leave the code-morphing proprietary.
It takes about as much time to drive out to CompUSA and buy the game as it does to download over my 256K link, and I can surf /. while I do it. It less time to order the game, but it takes longer to get it. I, of course, downloaded the demo AND ordered the game (hadn't been paying enough attention to see that it was released)
HOMMII was one of the very, very few games I found it worth rebooting into W95 to play.
Of course, Loki has open sourced quite a bit of software, including their installer and libraries they use to port games. ie. code that someone else actually could use (Unreal Tournament for Linux uses the Loki installer).
:) Games rock :)
But yeah, there is a double standard for games
Loki games are licensed by Loki from the companies that write the original, and thus are not Open Source. Their Linux based installer, which they wrote, is (I belive) Open Source, and was used by the Linux version of Unreal Tournament.
That is simply untrue. The cost of cycles is not 0. Consider that it takes longer to launch word 95 on a PII XYZ (whatever they are at these days) than it took to launch wordstar 5 on a 286. It also takes longer to load a file, save a file, reformat a file, and they typing lag is worse. And I will let you in on a little secret: There is almost no functionality missing in Wordstar -- just some of it is harder to get to (but faster to use once you know how!)
Now, these days, people (evidentally) accept the increased latency in exchange for the ease of use of Word, but my point is, it isn't free. It isn't even cheap.
Memory and disk are essentially free, but their bandwidth is not, and usually large files == long load times.
What they can do is effectively open development of LiViD for the purposes of creating a high-quality, free UNIX based DVD player. It just the illegal uses that are harder to curtail.
Since NTSC is an analog standard it has no defined horizontal resolution (it doesn't even have pixels) -- it just depends on how often you sample a given scan line. However, bandwidth, broadcaster, and CRT limitations keep the practical limit in the 500-700 range. The number of scanlines is, however, fixed so you always have the same number of scanlines (otherwise the TV wouldn't know where to display them).
newspaper pages per second is about the most ass-backward bandwidth unit I have ever heard...
I just got 3.9.17 built and started. The fonts are beautiful. I have a Millenium II, and it has always been reasonably fast under XF86, though.
Try changing "perhaps" to only with in the above statement.
or read the statement "If you had verification and security, perhaps you could even do this remotely"
Now don't you wish English was LALR(1)?
Not only that, but since nothign prevents you from making a bit-by-bit copy of the original, encrypted DVD which can then be played by existing software and (if you managed to put it on another DVD) hardware, cracking the DVD is only good for either A) converting it to another format, or B) playing it (just a specialized form of A where the target format is a video stream). Anyone who can manage the 6-8 GB storage requirements can pirate DVD w/o DeCSS.
Yes, exactly. You have hit the nail on the head. It explicitly falls under fair use to reverse engineer/modify software for the purpose of making it work with your system. This was decided years ago in some of the early computer software copyright lawsuits.
Netscape on UNIX has this option ("Only accept cookies originating from the same server as the page being viewed"), and has since at least version 3.0.
.co.uk or other idiotic domains (which would break things like sharing a login between www.yahoo.com and quote.yahoo.com), or if it is just dumbspeak for the same thing NS UNIX supports (I suspect it is the latter). I don't use windows much, so I haven't investigated.
Netscape on Windows has an option in the same place called "Only accept cookies that are sent to originating server" -- I don't know if this means don't accept
On Linux, using netscape, I haven't seen a cookie from doubleclick in over a year (I prune my cookie file regularly as well)