This phrase is one of my pet peeves. Almost every time I see the phrase, it has been used incorrectly.
If you are granting any non-statutory rights at all you should be using "all other rights reserved".
It's pretty much like public presentations and white papers having "proprietary and confidential" in fine print at the bottom. It's obviously put on the documents without through or deliberation. If the person or company has freely distributed copies of white papers, or given presentations open to the public, I can only assume that the phrase does not have the plain meaning and ignore it every time and place I see you use it.
Again, if you are granting any non-statutory rights at all, you should list those right and use "all other rights reserved". Or just skip the latter phrase, since it's implied.
In the specific case here, if your web site instructs my browser to automatically get and execute a script, that's the normal and expected action (i.e. as described with formal or de facto standards), and it's published/served without password protection, I expect that I have the right to execute it. Thus your "all rights reserved" notice contradicts your explicit actions.
You miss an important point: idle people consume more resources.
I work long hours, and don't have time to spend money. If I had more free time, I would use the time on expensive (both in $ and natural resources) hobbies.
A company in the position of SGI needs press releases to keep saying "I'm still here". A large percentage of the press releases were about Maya and Alias.
With reduced pressure, the thermal capacity of the cooling air is proportionally reduced. Combined with changes in forced air flow and microgravity not creating convectional cooling, you can have long-term overheating issues with equipment that consumes very little power.
This is solved on satellites with conductive cold/hot plates, but that results in much heavier equipment.
Regarding leakage rates, it's very difficult to estimate leak flow rates. The flow might be proportional to pressure squared, or cubed. If it's in an elastic seal, it may completely re-seal when the pressure drops to a specific level.
Humans can function at elevations of about 5psi (see the other posts about mountain climbers), and survive on a little less. Since O2 is less than 20% of the atmosphere, you can theoretically live on 1psi of pure O2. But secondary effects are killers at that low pressure, such as keeping enough moisture in the lung tissue. When near-pure O2 is used, it is usually at about 3psi or nearly the partial pressure of O2 at sea level.
Hmmm, those words granting limited power to the federal government and further protecting against state government excesses may be ignored when dealing with non-citizens or U.S. citizens abroad.
Synthetic Aluminum? Synthetic Steel
on
The Diamond Age
·
· Score: 2, Insightful
This is so "bought"!
Aluminium, and to a lesser extent steel, used to be rare and expensive because only tiny quantities occur naturally. Especially in the case of a pure mineral, I can't see the justification for requiring a perjorative adjective.
"Synthetic" is intended to be perjorative, or else DeBeers wouldn't be pushing for it. It will never happen, but it would be sweet justice if the FTC rules that natural diamonds must be called "contaminated crystal carbon".
We have moved to different offices in Annapolis about a mile away from our original location. The view isn't as nice, but the new office space has a machine room that is about three times the size of the small office that we previously used.
Re:OpenMosix, really Distributed Shared Memory
on
Ask Donald Becker
·
· Score: 2, Insightful
Despite (or perhaps because of) long experience with shared memory parallel processors, I don't see distributed shared memory (DSM) as a useful approach. The programmer must write or rewrite the application to very carefully use the shared memory in a way that avoids the write lock from bouncing between systems. It ends up being simpler and faster, for most applications, to just write the direct message passing code.
Mosix is almost completely unrelated to DSM. While I think Mosix is a very interesting academic project, it's the wrong model to build scalable performance clusters. Cluster applications don't want transparent process migration with forwarded paging and I/O. They want to explicitly and quickly start up processes on remote machines, and have direct control over the performance-critical I/O and communication paths.
Yes, I reversed-engineered the Morris worm and wrote source code that compiled to the same binary. The original binaries did not have their symbols stripped, making the task somewhat easier. I also had the advantage of having studied the internals of the compiler that Morris used.
I was working on several parallel processing projects at Harris. The original Linux user-level NFS server was the NFS server I developed to support the Concert shared memory multiprocessor that was a joint MIT-Harris project.
The rtl8139 driver code keeps churning exactly because I don't have control over the changes. Look at who is making the modifications -- it's not me.
Many of the changes do come from my driver releases, but are incompletely or incorrectly copied over the driver in the kernel.
The was all part of the kernel development fall-out about two years ago. I only wanted to release working drivers into the kernel, while Linus wanted all development moved into the regular kernel releases. Linus just started accepting patches from anyone, which resulted in much lower quality drivers.
Check out the 27Bz-7 Scyld Beowulf CD distributed at the NYC LinuxWorld show. You'll see how Wake-On-LAN is now moving into the cluster control and node scheduling tools.
If a compute node takes two minutes to boot, you can't risk it being powered off when you need it. If it only takes a few seconds for a compute node to rejoin the cluster you can freely power down idle nodes. If you look at the average utilization for medium sized clusters, you'll see that this could save a lot of power.
We are doing the public roll-out of the new Beowulf system that we've been working on for the past two years.
Drop by our booth if you would like to see a demo.
The Scyld Beowulf distribution has been available since August. The system was demoed and CDs were widely distributed at the October Atlanta Linux Showcase.
Our distribution is considerably more sophisticated than what OSCAR is attempting to do. The advances have lead to a simpler system to install and run.
Our distribution CD may be used as an install disk for the cluster master, or after 20 seconds it automatically boots the machine as a cluster slave and tries to contact a master. Once the master has been installed (about 20 minutes), each slave takes only a few seconds to join the cluster.
Look for our demos at LinuxWorld Expo next week, and our latest distribution is now available for $2 from LinuxCentral.com. You don't have to wait weeks or months for a less capable OSCAR system.
It takes a long time for the technical reports to make their way through the system and turn into policy. Beowulf clusters were mentioned by name as a technology that makes the old hardware-based scheme irrelevant.
It doesn't matter if the device driver has its own virtual address space. The hardware is still writing to physical memory, and thus the hardware can still corrupt unrelated memory.
You don't address my main point: if the Hurd is so much easier to write and extend, why is it still incomplete and unusable after more than a dozen years?
1) There are "servers" instead of "drivers". Why is this important? Servers have their own address space. So, what happens if my network driver decides to go ballistic? Just restart it. If that happens in Linux it's bye-bye system.
Uhmmm, no. Under Linux, if the device driver dereferences NULL the kernel reports the problem and moves on. It has been this way since 1992.
Under either system if the hardware writes a random chunck of physical memory, you are hosed. Putting the device driver code in its own address space doesn't make any difference.
) Any user can write kernel code without creating a security risk. This includes user filesystems and the like.
If there is no security issue, why is that code in the kernel? Oh, because it's part of the file system security model? Hmmmm.
I offered to contribute to the Hurd in 1988, soon after I left MIT. My offer was rejected, with the claim that the design work was already done and that only people physically near MIT would be able to effectively work on the nearly finished code. That makes sense -- it would be impossible to develop a working OS with developers scattered around the world communicating only with email and FTP.
Six years later I remember RMS urging us to stop working on Linux, since it was distracting people from the Hurd, the only OS that the FSF would endorse. If anyone would have written "GNU/Linux" the FSF would probably have sued for trademark misuse! Hmmm, and why isn't it GNU/Hurd?
Anyway, by my count the Hurd is well over a dozen years old. It started from cribbed software, and has used software from the many working OSes that have surpassed it, yet it still hasn't produced a usable system. The same old arguments about how monolithic kernels are too painful to develop for are still being repeated. Give it up, the evidence is clear: the Hurd is a dead-end path.
We have created a much easier to use and maintain cluster system!
Once the software has been installed on the front-end machine (which adds three Beowulf-specific questions to a standard install), adding a new compute node takes only a few seconds more than booting up the new machine. We even provide a button on the "beosetup" GUI to make boot floppies if your system won't boot from the network or CD-ROM.
Booting is fast because relatively little is initially transferred. The compute nodes typically run with about 40MB of cached library and configuration files.
This is not a NFS root scheme, which has even more complexity than setting up disk file systems. With BProc we actually eliminate most of the complexity rather just than moving it to some other place in the system.
The Scyld Beowulf-2 distribution you can buy from LinuxCentral.com is x86 only, but we support other architectures.
The only x86 specific feature is the cool "Two Kernel Monte", a kernel module which allows you load a new kernel(!). T-K-M is useful for any Linux system, not just for Beowulf.
The Alpha AXP is supported only with custom distributions because the Alpha requires a kernel matched to the specific motherboard type. That would mean two dozen CD-ROMs instead of just one.
We previously had Sparc-32 support, but that has been dropped. Beowulf is focused on price/performance. Sparcs are expensive and slow.
PowerPC support is planned. The Beowulf-2 system is based around BProc, which requires processor specific modifications to the kernel. For instance we add a new executable type to the kernel and "VMA dump" to save an executing program to a file or network stream. So it's way more than just a recompile to support a new architecture.
This situation started over a month ago. RMS was Cc:ed on the email, but since the Linux kernel is not a FSF project (the FSF requires that copyright be assigned to the FSF) they did not want to become directly involved.
Sun restricted access to the porting kit a few days ago, before this story was published.
Bruce Perens has nothing to do with this, beside making rash public statements that the GPL has a loophole that permits any abuse -- statements have little basis in reality.
Take a look at 3c509.c or 8390.c to see who I was assigning the copyright to in 1992 and 1993. As far as I can tell, I was the first person to get a GPL release of code through that system. I'm proud of that landmark!
Going through a six month external release review cycle does make you very careful when selecting the next position.
In 1994 I moved to USRA-CESDIS. USRA usually hosts visiting scientists that are working with NASA. Their home institutions usually want to control any inventions and copyright, so it's possible for people in research positions to retain their own copyright as long as the work product is made available. The GPL is significantly better in the long run than the usual limited benefit the government gets from sponsored research.
One of the benefits of my current position working at Scyld is that we support projects released under GPL. Don't question how I'm able to keep my copyright, question why you haven't applied to work at Scyld or one of the other similar employers!
The network drivers are not stand-alone executables or libraries. They are sections of code that are part of a larger program ("executable"). The drivers are not "run on" the Solaris system like an executable, but become part of the kernel itself.
The situation becomes clear when stated as: Sun is extracting useful code from the Linux kernel and including it into the Solaris kernel.
The legal test for "derivative work" does not require the inclusion of the original in the distribution. (We can even ignore that Sun did include some Linux kernel code in their kit.) The Sun kit could not have been written without the source code of tulip.c, eepro100.c and the other Linux network drivers. The kit has no value or meaning without the original code, and is thus a derivate work covered by the original copyright and license.
Using this derivative work results in a license conflict. This situation is especially clear, since Sun itself is distributing the kit. They can resolve the license conflict by releasing the Solaris kernel under the GPL.
Slashdot Mirror
This phrase is one of my pet peeves. Almost every time I see the phrase, it has been used incorrectly.
If you are granting any non-statutory rights at all you should be using "all other rights reserved".
It's pretty much like public presentations and white papers having "proprietary and confidential" in fine print at the bottom. It's obviously put on the documents without through or deliberation. If the person or company has freely distributed copies of white papers, or given presentations open to the public, I can only assume that the phrase does not have the plain meaning and ignore it every time and place I see you use it.
Again, if you are granting any non-statutory rights at all, you should list those right and use "all other rights reserved". Or just skip the latter phrase, since it's implied.
In the specific case here, if your web site instructs my browser to automatically get and execute a script, that's the normal and expected action (i.e. as described with formal or de facto standards), and it's published/served without password protection, I expect that I have the right to execute it. Thus your "all rights reserved" notice contradicts your explicit actions.
It's difficult to tell if it has distributed loading or mid-point loading,
You miss an important point: idle people consume more resources.
I work long hours, and don't have time to spend money. If I had more free time, I would use the time on expensive (both in $ and natural resources) hobbies.
A company in the position of SGI needs press releases to keep saying "I'm still here". A large percentage of the press releases were about Maya and Alias.
With reduced pressure, the thermal capacity of the cooling air is proportionally reduced. Combined with changes in forced air flow and microgravity not creating convectional cooling, you can have long-term overheating issues with equipment that consumes very little power.
This is solved on satellites with conductive cold/hot plates, but that results in much heavier equipment.
Regarding leakage rates, it's very difficult to estimate leak flow rates. The flow might be proportional to pressure squared, or cubed. If it's in an elastic seal, it may completely re-seal when the pressure drops to a specific level.
Humans can function at elevations of about 5psi (see the other posts about mountain climbers), and survive on a little less. Since O2 is less than 20% of the atmosphere, you can theoretically live on 1psi of pure O2. But secondary effects are killers at that low pressure, such as keeping enough moisture in the lung tissue. When near-pure O2 is used, it is usually at about 3psi or nearly the partial pressure of O2 at sea level.
Hmmm, those words granting limited power to the federal government and further protecting against state government excesses may be ignored when dealing with non-citizens or U.S. citizens abroad.
This is so "bought"!
Aluminium, and to a lesser extent steel, used to be rare and expensive because only tiny quantities occur naturally. Especially in the case of a pure mineral, I can't see the justification for requiring a perjorative adjective.
"Synthetic" is intended to be perjorative, or else DeBeers wouldn't be pushing for it. It will never happen, but it would be sweet justice if the FTC rules that natural diamonds must be called "contaminated crystal carbon".
We have moved to different offices in Annapolis about a mile away from our original location. The view isn't as nice, but the new office space has a machine room that is about three times the size of the small office that we previously used.
Despite (or perhaps because of) long experience with shared memory parallel processors, I don't see distributed shared memory (DSM) as a useful approach. The programmer must write or rewrite the application to very carefully use the shared memory in a way that avoids the write lock from bouncing between systems. It ends up being simpler and faster, for most applications, to just write the direct message passing code.
Mosix is almost completely unrelated to DSM. While I think Mosix is a very interesting academic project, it's the wrong model to build scalable performance clusters. Cluster applications don't want transparent process migration with forwarded paging and I/O. They want to explicitly and quickly start up processes on remote machines, and have direct control over the performance-critical I/O and communication paths.
Yes, I reversed-engineered the Morris worm and wrote source code that compiled to the same binary. The original binaries did not have their symbols stripped, making the task somewhat easier. I also had the advantage of having studied the internals of the compiler that Morris used.
I was working on several parallel processing projects at Harris. The original Linux user-level NFS server was the NFS server I developed to support the Concert shared memory multiprocessor that was a joint MIT-Harris project.
The netdriver update package at ftp://ftp.scyld.com/pub/network supports the 2.0 kernels. The 1.2 kernel support was dropped about a year ago.
It's already a switching power supply -- a linear power supply would be far larger, heavier and more expensive.
Reproduce the bad grammer correctly ;->.
The rtl8139 driver code keeps churning exactly because I don't have control over the changes. Look at who is making the modifications -- it's not me.
Many of the changes do come from my driver releases, but are incompletely or incorrectly copied over the driver in the kernel.
The was all part of the kernel development fall-out about two years ago. I only wanted to release working drivers into the kernel, while Linus wanted all development moved into the regular kernel releases. Linus just started accepting patches from anyone, which resulted in much lower quality drivers.
If a compute node takes two minutes to boot, you can't risk it being powered off when you need it. If it only takes a few seconds for a compute node to rejoin the cluster you can freely power down idle nodes. If you look at the average utilization for medium sized clusters, you'll see that this could save a lot of power.
Scyld Beowulf
We are doing the public roll-out of the new Beowulf system that we've been working on for the past two years. Drop by our booth if you would like to see a demo.
Our distribution is considerably more sophisticated than what OSCAR is attempting to do. The advances have lead to a simpler system to install and run.
Our distribution CD may be used as an install disk for the cluster master, or after 20 seconds it automatically boots the machine as a cluster slave and tries to contact a master. Once the master has been installed (about 20 minutes), each slave takes only a few seconds to join the cluster.
Look for our demos at LinuxWorld Expo next week, and our latest distribution is now available for $2 from LinuxCentral.com. You don't have to wait weeks or months for a less capable OSCAR system.
It takes a long time for the technical reports to make their way through the system and turn into policy. Beowulf clusters were mentioned by name as a technology that makes the old hardware-based scheme irrelevant.
You don't address my main point: if the Hurd is so much easier to write and extend, why is it still incomplete and unusable after more than a dozen years?
Uhmmm, no. Under Linux, if the device driver dereferences NULL the kernel reports the problem and moves on. It has been this way since 1992.
Under either system if the hardware writes a random chunck of physical memory, you are hosed. Putting the device driver code in its own address space doesn't make any difference.
) Any user can write kernel code without creating a security risk. This includes user filesystems and the like.
If there is no security issue, why is that code in the kernel? Oh, because it's part of the file system security model? Hmmmm.
I offered to contribute to the Hurd in 1988, soon after I left MIT. My offer was rejected, with the claim that the design work was already done and that only people physically near MIT would be able to effectively work on the nearly finished code. That makes sense -- it would be impossible to develop a working OS with developers scattered around the world communicating only with email and FTP.
Six years later I remember RMS urging us to stop working on Linux, since it was distracting people from the Hurd, the only OS that the FSF would endorse. If anyone would have written "GNU/Linux" the FSF would probably have sued for trademark misuse! Hmmm, and why isn't it GNU/Hurd?
Anyway, by my count the Hurd is well over a dozen years old. It started from cribbed software, and has used software from the many working OSes that have surpassed it, yet it still hasn't produced a usable system. The same old arguments about how monolithic kernels are too painful to develop for are still being repeated. Give it up, the evidence is clear: the Hurd is a dead-end path.
We have created a much easier to use and maintain cluster system!
Once the software has been installed on the front-end machine (which adds three Beowulf-specific questions to a standard install), adding a new compute node takes only a few seconds more than booting up the new machine. We even provide a button on the "beosetup" GUI to make boot floppies if your system won't boot from the network or CD-ROM.
Booting is fast because relatively little is initially transferred. The compute nodes typically run with about 40MB of cached library and configuration files.
This is not a NFS root scheme, which has even more complexity than setting up disk file systems. With BProc we actually eliminate most of the complexity rather just than moving it to some other place in the system.
The Scyld Beowulf-2 distribution you can buy from LinuxCentral.com is x86 only, but we support other architectures.
The only x86 specific feature is the cool "Two Kernel Monte", a kernel module which allows you load a new kernel(!). T-K-M is useful for any Linux system, not just for Beowulf.
The Alpha AXP is supported only with custom distributions because the Alpha requires a kernel matched to the specific motherboard type. That would mean two dozen CD-ROMs instead of just one.
We previously had Sparc-32 support, but that has been dropped. Beowulf is focused on price/performance. Sparcs are expensive and slow.
PowerPC support is planned. The Beowulf-2 system is based around BProc, which requires processor specific modifications to the kernel. For instance we add a new executable type to the kernel and "VMA dump" to save an executing program to a file or network stream. So it's way more than just a recompile to support a new architecture.
This situation started over a month ago. RMS was Cc:ed on the email, but since the Linux kernel is not a FSF project (the FSF requires that copyright be assigned to the FSF) they did not want to become directly involved.
Sun restricted access to the porting kit a few days ago, before this story was published.
Bruce Perens has nothing to do with this, beside making rash public statements that the GPL has a loophole that permits any abuse -- statements have little basis in reality.
Going through a six month external release review cycle does make you very careful when selecting the next position.
In 1994 I moved to USRA-CESDIS. USRA usually hosts visiting scientists that are working with NASA. Their home institutions usually want to control any inventions and copyright, so it's possible for people in research positions to retain their own copyright as long as the work product is made available. The GPL is significantly better in the long run than the usual limited benefit the government gets from sponsored research.
One of the benefits of my current position working at Scyld is that we support projects released under GPL. Don't question how I'm able to keep my copyright, question why you haven't applied to work at Scyld or one of the other similar employers!
The network drivers are not stand-alone executables or libraries. They are sections of code that are part of a larger program ("executable"). The drivers are not "run on" the Solaris system like an executable, but become part of the kernel itself.
The situation becomes clear when stated as: Sun is extracting useful code from the Linux kernel and including it into the Solaris kernel.
The legal test for "derivative work" does not require the inclusion of the original in the distribution. (We can even ignore that Sun did include some Linux kernel code in their kit.) The Sun kit could not have been written without the source code of tulip.c, eepro100.c and the other Linux network drivers. The kit has no value or meaning without the original code, and is thus a derivate work covered by the original copyright and license.
Using this derivative work results in a license conflict. This situation is especially clear, since Sun itself is distributing the kit. They can resolve the license conflict by releasing the Solaris kernel under the GPL.