I would avoid the question for as long as sanely possible... It seems like a pretty important question, politically speaking. Why weasel out of it?
My definition is this: anything that can function as a human in society is human. Yes, that excludes people in a coma, as long as they stay in the coma; it includes any possible artificial intelligences, assuming we ever manage to create one.
DNA is fine for one definition of human -- as in, me human, you big hairy ape-thing DNA is pretty bad for that, actually. The genetic difference between humans and big hairy ape-things is pretty small. I've seen 90-odd percent DNA similarity statistics quoted. Besides, there are things like chimeras with human cells, and human cell cultures on petri dishes.
They have places where you can find information about products now. They're called wikipedia, Consumer Reports, and trade mags. Or you could www.justfuckinggoogleit.com
The ultimate goal of these studies will take us further away from a "sheep-man" and closer to a sheepish organ incubator. Stem cells become their ultimate tissues/organs through many steps, some of which we can recognize and manipulate even today. Heck, right down the hall from me, they're making neurons and next to them, they're making bone cells hoping to find treatments for things like osteogenesis imperfecta...see Mr. Glass in "Unbreakable":D
I understand that the intentions are good, and that probably a lot of good will come out of these studies. I get more enthusiastic about stem cells with each passing year, as I age:)
... and of course the people who made these sheep would rather program the cells to JUST become a heart or JUST a liver for example. They don't want "sheep-human" freaks. That'd be a little hard to sell to their funding agency. hehe. So bottom line, a human liver does not a sheep-man make IMO.
The intentions of inventors and funding sources are not always honored. The Internet was originally developed to provide a redundant communication network for military installations-- not to bring you spam about v1g4a. A lot of the exotic materials developed for the space program have become commonplace in the commercial world. Dr. Albert Hofman originally believed that LSD would be used as a tool of psychotherapy.
I can already predict that some people will find the idea of animal implants to be pretty interesting. It might be possible to give someone claws like a cat or a tail, for example. You may find this outlandish and bizarre, but there's probably already some subculture of people on the internet who think it's just a great idea. Once we better understand how to create chimerism, the boundary between humans and animals will be weakened.
Of course, chimeras are small potatoes, really. Genetic engineering is where it's at. Eventually, we will have to somewhat rethink our ideas about what is human. There are a lot of philosophical details you could go into, but as a practical matter, I think eventually we will come to treat anything that can function effectively in society as human. It's the only "humane" solution, heh.
However, it is not easily predicted what a virus comfortable in sheep might do in a human host. Bottom line however, if the options are risk or death, go with risk. It would be an easy call if the risk was just to the person getting the transplant. Their body, their choice. However, communicable diseases like AIDS and bird flu got their start as zoonoses...
As a Christian, I've got no major problem with this and most Christians who really understand it will agree. We eat these animals and wear their skin and fur as clothing. I don't think this is such a small matter as you claim. The boundaries between two species have been weakened. Biologists have always pointed out the similarities between human DNA and that of animals... but now there is a concrete... creature embodying these similarities. Yes, I realize that the human cells have not interbred with the sheep cells, but this is still a partly human creature, a chimera.
I think the next big step is a sheep with a human brain. Someone will do it. Even if the research is banned in the US, you can count on Singapore or China to push the envelope. And that will raise a lot of questions that nobody likes to think about.
One sheep could save a person's life and save us all $$ that would be spent on anti-rejection drugs. I agree. This could be revolutionary for organ transplants.
For what it's worth, I'm also in the telecom industry.
A true production level environment has fully controlled endpoints. Even the VOIP cable companies provide don't use IP. Most use ATM and/or frame relay. ATM and frame relay are legacy technologies. They're being phased out in favor of IP just as quickly as carriers can.
Everything's not so free when it's YOUR equipment or network others want to leech, is it. Yeah, except that it was built with OUR tax dollars, and the current "owners" are doing little to invest their windfall profits back into the system. As a result, the United States is now falling behind the rest of the world in broadband internet.
I can vouch for this, having worked at a major router manufacturer. We made lots of sales around the world, to BT, France Telecom, some Russians, KT, India. There's a lot of growth around the world, but hardly any in the U.S. market. Of course, part of this is because the U.S. population has not been growing.
But the bigger issue is that the U.S. telecoms invest the absolute bare minimum to keep people from screaming. And if you live in a rural area, you might as well forget about it-- you'll be stuck paying 1940s era prices for 1940s era phone technology. (Here's a hint-- in 2007, that's not a good deal.)
Companies also try to maintain the farce that long distance calls cost them more, which today makes about as much sense as the droit de signor. And the related farces that caller ID and three-way calls should be hugely expensive. I could go on, but... what's the point? I could hardly think of an organization I hate more than Verizon, except maybe the Internal Revenue Service.
Usually, companies try to go for the "we buy the technology for a few million dollars and let it die" route, because it's the least amount of hassle and risk and keeps everybody happy.
Patents expire after 15 years. They're also a matter of public record. So I really don't see how any single company can suppress a given technology on a permanent basis.
The whole foods diet you recommend will kill most diabetics...
It would also kill most critically ill patients who have to be fed through a tube. And probably people who are allergic to the food. And probably 1-day old babies.
What's your point? The diet is not FOR any of these people. People who have special medical conditions should consult a nutritionist. Period. Full stop.
Maybe not, but these are long-range plans. Are you really certain that the next few decades won't see the US needing to fight a high-intensity war?
1945 called, they want their military strategy back.
These days, "high-intensity wars" aren't about soldiers. They're about the president pushing a nice shiny button marked "nukes" and turning us all into crispy chicken.
I wouldn't worry too much about stupid comments in a web forum. What you are doing sounds very much in the good old tradition of advancing science and human understanding through computation.
Whenever the word "computer science" appears in the description or the article, there's always a tedious pedant who shows up to inform everyone that "X isn't REAL computer science."
I take the position that "real computer science" is anything that the university puts in the computer science department. I went to Carnegie Mellon University and we had theory (algorithm analysis, graph theory, turing machines), discrete math (cryptography, compression), systems (software engineering, web caches, operating systems), artificial intelligence (bayesian networks, computer vision, machine learning), programming languages (proving soundness and compactness of various languages, etc), and human-computer interaction (psychology, quantifying human reaction times, etc.) There's a bunch of other sort of interdisciplinary stuff I left out like computational bio.
Just because you like one of these things more than the others doesn't mean that you are the only "real" computer scientist. There are tenured professors in all of these sub-disciplines of computer science. If you're good at what you do, I feel good for you. If not, I'm sorry for you. But don't fling poo at people who have different interests than yourself. It just makes you look like a monkey.
1. No 64 bit Flash. Or the lack of support in the X64 version of Firefox for 32 bit plug ins. That's a Macromedia problem. Hopefully they will release their 64-bit version soon, or the GNU flash will catch up...
3. Drivers specifically the fact that it is IMPOSIBLE for a manufacture to put a binary linux driver on a disk and stick in the box with his product. Driver disks suck. By the time you buy your hardware from WallyWorld, the driver disk has been gathering dust for a few months, and is no longer current. Bugs may have been discovered, or new features added, in a later release. Also, it's a huge pain to insert 10 different driver disks to set up your new computer.
It's much easier to just do "apt-get" or "yum," and get the drivers you need. But if you really feel nostalgic, you can ritually insert and remove a few floppies.
The other issue you alluded to, not having a fixed ABI for drivers, is actually an advantage for Linux. It can adapt and change over time much easier than any version of Windows, for exactly this reason. And the kernel maintainers keep all in-tree drivers up to date, anyway, so companies have no reason to complain.
And I get to thinking... that we have all this stuff that could be done around here. Roads, housing, bridges, new infrastructure that could be built, that sort of thing. Stuff that benefits the community.
Most people like to do only 30 minutes or an hour of exercising, then go back to whatever they had been doing. Building roads, housing, and bridges is not something that can really be done in that time frame. I agree with the other poster-- with people arriving on the construction site for 30 minutes, and then leaving, no work would get done. It would be much too hard to coordinate.
Also, you are talking about skilled labor, which means that there is a learning curve. Believe it or not, it's hard to manage skilled laborers, to ensure that everyone knows what they're supposed to know, and doesn't mess up. Some of the work could be dangerous for a newbie to do. We are talking about power tools and construction sites here.
Also, on a more practical level, construction workers and machinists have powerful unions that would oppose any kind of community involvement in the industry. They are well aware that it would cost them a lot of jobs, and they would play up the safety angle as much as they could, to ensure that the idea never got off the ground.
Probably the most practical idea for using people's exercise energy would be to attach electrical generators to certain pieces of equipment. Of course, then you get into the same problems that wind and solar energy has always had-- how do you store this energy without causing more pollution than you save? Batteries are highly toxic and relatively expensive.
There aren't a whole lot of businesses or governments eager to buy "surplus energy." Think about it. Would you sign a contract offering to power your house "only when it was convenient"? No. So perhaps you should think about this problem instead.
My 2c - pThreads suck ass. There's nothing worse than looking at 1000 Java 'processes' on your JSP server and wondering which ones will die if you kill one. It also makes it hard for the OS to figure out that App X hasn't had any CPU time for over 500ms.
You are confused. Pthreads is a cross-platform API with a bunch of different implementations, including ones for Linux, Windows, BSD, and some embedded platforms.
LinuxThreads is the crappy threading library that Linux used to use, which treats threads as processes. Luckily, new Linux systems now come with NPTL, which doesn't have this problem.
> > Basically, using flash memory for swap. Linux has been able to do this for a long time. > This is not what ReadyBoost is doing.
Ok. I am genuinely curious about this kernel stuff, and not just looking to bash Windows! So I looked up the paragraphs about ReadyBoost.
After the ReadyBoost service initializes caching, the Ecache.sys device driver intercepts all reads and writes to local hard disk volumes (C:\, for example), and copies any data being written into the caching file that the service created. Ecache.sys compresses data and typically achieves a 2:1 compression ratio so a 4GB cache file will usually contain 8GB of data. The driver encrypts each block it writes using Advanced Encryption Standard (AES) encryption with a randomly generated per-boot session key in order to guarantee the privacy of the data in the cache if the device is removed from the system.
When ReadyBoost sees random reads that can be satisfied from the cache, it services them from there, but because hard disks have better sequential read access than flash memory, it lets reads that are part of sequential access patterns go directly to the disk even if the data is in the cache.
You are technically correct-- Microsoft's ReadyBoost isn't quite swap. It's more like a hard drive cache to speed up random accesses to and from the disk. Typically these I/O caches are kept in memory, but I guess using a flash key would allow you to have a bigger (although slower) I/O cache.
This is a feature that, as far as I know, Linux does not have.
Thanks for the clarification. Is it also used during suspend / resume? Because, at least I personally do more suspending and resuming than cold boots.
It seems like you could create "ReadyBoost for Linux" by writing a driver for the H-HDD flash, and putting your/boot and/etc partitions on the flash. Maybe you could get parts of X on there too, I'm not sure.
You may not like them, but how many of them can Linux do? Specifically, can Linux cache to Flash devices? Does Linux automatically optimize application startup based on the last five logon sessions? (I'm not trying to be a smart ass here, I'm really curious. And if it doesn't, why not?)
ReadyBoot / ReadyBoost - Basically, using flash memory for swap. Linux has been able to do this for a long time. However, Linux does not do this "automatically" when you insert a flash device, nor does it set the priority of the flash swap space automatically. You could write a script that would do this pretty easily though.
Bootmgr - Microsoft replaced their NTLDR boot manager with this. NTLDR only booted variants of Windows, and refused to see any other partitions. Perhaps the new Bootmgr will be more permissive? It's unlikely to matter, since everyone who wanted to boot something other than Windows was already running GRUB or LILO anyway.
ReadyDrive - This seems to be similar to ReadyBoost, except that the flash memory in question is on the hard disk itself. The summary mentions something called an H-HDD, which is a hard disk that has NVRAM integrated into it.
I think they are hoping to use this to improve startup time by preloading parts of the OS into this NVRAM. Other than that, the summary doesn't really give me a clear idea of how this is different from regular swap.
Actually most hard disks do have NVRAM on them even now. What's new about H-HDDs is that they expose that fact to the operating system. Whether or not this will yield important performance gains remains to be seen.
SuperFetch - This is a virtual memory system enhancement. Basically, VM systems are being tweaked all the time. There isn't really any "right" way to do it, there's just a bunch of heuristics that work better or worse depending on your usage patterns.
The paragraphs about SuperFetch spend a long time talking about the "after lunch syndrome" whereby users have a bunch of foreground apps open (think MS Word, MS Excel, whatever...), and then they go to lunch. Then, their screensaver, or SETI@home, or whatever, begins to run, and the foreground apps are swapped out to disk. When the user comes back, he finds that the each time he tries to click on a former foreground app, Windows has to spend a long time loading the app back into RAM from swap.
In many cases, it took longer to load the application back into RAM than it did to restart the app entirely. This is not that surprising, actually. Apps tend to increase their memory use over time, by means of incremental allocations. Being hit with the total cost of those allocations all at once can be overwhelming. Many Windows apps can use gigs of memory, so it was like doing a huge file copy.
If you know that desktop computers tend to have an "on / off" usage pattern (i.e., the monkey is either at the keyboard, or not,) you can try to make your VM "clever." This might help desktop users, although server users would want to stay as far away from it as possible.
I haven't ever observed "after lunch syndrome" on my Linux desktops. I seldom even have to use virtual memory, actually. Currently, the biggest memory hogs I run are GNOME and firefox, but even they can only about half-fill my 1 gigabyte of memory. Even OpenOffice doesn't seem to take much of a bite out of it. The only app I have that consistently thrashes the cache is the Gimp.
Not much of a solution for me, I refuse to run script or bytecode.
I hope someday you change your mind. Then you'll be able to run MS Windows, Linux, or BSD, all of which execute scripts when they start up, and to accomplish many routine tasks.
If he did the surgery expertly, and the poor outcome was for reasons other than what was in his control - then that's simply not his fault... If she had died unexpectedly as a result of the anaesthetic - statistically a 1 in 250 000 chance - then that's out of his control.
Absolutely. There's a reason why people fear "unecessary surgery."
Unfortunately, no matter how many times you warn people about the risks, they still get mad when the dice come up snake eyes.
Disclaimer: I have no idea whether the doctor in question is a good one or a bad one.
There are lots of reasons why a car accident might not be the fault of the driver. Perhaps the driver's car was assembled improperly in the factory. Maybe someone else on the road was driving like an idiot. Maybe there was a medical emergency that nobody could have predicted.
If you have a stroke, and lose control of your body, and crash your car into a bus full of schoolchildren, you are not guilty of manslaughter. Assuming that there is no pre-existing medical condition that you were aware of, it is considered "an act of god" in legal terminology.
You cannot be held responsible for any action unless you knowingly consent to performing that action. Without intent, there can be no guilt. You aren't responsible just because "you happened to be there and something went wrong." Anything else makes a mockery out of morality.
"Citing sources" is ALL an encyclopedia can ever be. If you just "make stuff up" because you're a smart guy in general, then it's not wikipedia, it's something else-- a blog, perhaps. The goal is to provide the very minimum amount of interpretation and bias possible-- just the facts.
Yes, some sources are wrong, and some become obsolete. That's why we have this button called "edit," that allows you to change things after they've been entered in.
We can get all postmodernist and quantum mechanical, and argue about the distinction between facts and opinion, but the reality is, most people acknowlege that there is such a distinction, and find it useful.
Slashdot Mirror
I would avoid the question for as long as sanely possible...
It seems like a pretty important question, politically speaking. Why weasel out of it?
My definition is this: anything that can function as a human in society is human. Yes, that excludes people in a coma, as long as they stay in the coma; it includes any possible artificial intelligences, assuming we ever manage to create one.
DNA is fine for one definition of human -- as in, me human, you big hairy ape-thing
DNA is pretty bad for that, actually. The genetic difference between humans and big hairy ape-things is pretty small. I've seen 90-odd percent DNA similarity statistics quoted. Besides, there are things like chimeras with human cells, and human cell cultures on petri dishes.
Well, nothing in life is free, especially information.
I would have more sympathy for you if I didn't already know about your history with support forums.
The stark truth of slashdot is that 100% of the people who "make up statistics" on the spot are jackasses, whether they realize it or not.
Kindly find a study that seems to show what you want, and cherry-pick some factoids out of context, like the rest of the politicos.
They have places where you can find information about products now. They're called wikipedia, Consumer Reports, and trade mags. Or you could www.justfuckinggoogleit.com
I understand that the intentions are good, and that probably a lot of good will come out of these studies. I get more enthusiastic about stem cells with each passing year, as I age :)
The intentions of inventors and funding sources are not always honored. The Internet was originally developed to provide a redundant communication network for military installations-- not to bring you spam about v1g4a. A lot of the exotic materials developed for the space program have become commonplace in the commercial world. Dr. Albert Hofman originally believed that LSD would be used as a tool of psychotherapy.
I can already predict that some people will find the idea of animal implants to be pretty interesting. It might be possible to give someone claws like a cat or a tail, for example. You may find this outlandish and bizarre, but there's probably already some subculture of people on the internet who think it's just a great idea. Once we better understand how to create chimerism, the boundary between humans and animals will be weakened.
Of course, chimeras are small potatoes, really. Genetic engineering is where it's at. Eventually, we will have to somewhat rethink our ideas about what is human. There are a lot of philosophical details you could go into, but as a practical matter, I think eventually we will come to treat anything that can function effectively in society as human. It's the only "humane" solution, heh.
No, you idiot! Reroute warp core power to the hull shields!
Then reverse the polarity FTW!
However, it is not easily predicted what a virus comfortable in sheep might do in a human host. Bottom line however, if the options are risk or death, go with risk.
It would be an easy call if the risk was just to the person getting the transplant. Their body, their choice. However, communicable diseases like AIDS and bird flu got their start as zoonoses...
As a Christian, I've got no major problem with this and most Christians who really understand it will agree. We eat these animals and wear their skin and fur as clothing.
I don't think this is such a small matter as you claim. The boundaries between two species have been weakened.
Biologists have always pointed out the similarities between human DNA and that of animals... but now there is a concrete... creature embodying these similarities. Yes, I realize that the human cells have not interbred with the sheep cells, but this is still a partly human creature, a chimera.
I think the next big step is a sheep with a human brain. Someone will do it. Even if the research is banned in the US, you can count on Singapore or China to push the envelope. And that will raise a lot of questions that nobody likes to think about.
One sheep could save a person's life and save us all $$ that would be spent on anti-rejection drugs.
I agree. This could be revolutionary for organ transplants.
For what it's worth, I'm also in the telecom industry.
A true production level environment has fully controlled endpoints. Even the VOIP cable companies provide don't use IP. Most use ATM and/or frame relay.
ATM and frame relay are legacy technologies. They're being phased out in favor of IP just as quickly as carriers can.
Everything's not so free when it's YOUR equipment or network others want to leech, is it.
Yeah, except that it was built with OUR tax dollars, and the current "owners" are doing little to invest their windfall profits back into the system.
As a result, the United States is now falling behind the rest of the world in broadband internet.
I can vouch for this, having worked at a major router manufacturer. We made lots of sales around the world, to BT, France Telecom, some Russians, KT, India. There's a lot of growth around the world, but hardly any in the U.S. market. Of course, part of this is because the U.S. population has not been growing.
But the bigger issue is that the U.S. telecoms invest the absolute bare minimum to keep people from screaming.
And if you live in a rural area, you might as well forget about it-- you'll be stuck paying 1940s era prices for 1940s era phone technology. (Here's a hint-- in 2007, that's not a good deal.)
Companies also try to maintain the farce that long distance calls cost them more, which today makes about as much sense as the droit de signor. And the related farces that caller ID and three-way calls should be hugely expensive. I could go on, but... what's the point? I could hardly think of an organization I hate more than Verizon, except maybe the Internal Revenue Service.
Just put in a heating element and power it off the battery. Or run the motor a little bit.
"How to use energy to make stuff hot" is hardly an unsolved problem in engineering.
Usually, companies try to go for the "we buy the technology for a few million dollars and let it die" route, because it's the least amount of hassle and risk and keeps everybody happy.
Patents expire after 15 years. They're also a matter of public record. So I really don't see how any single company can suppress a given technology on a permanent basis.
Now network effects... those are real, and huge.
The whole foods diet you recommend will kill most diabetics...
It would also kill most critically ill patients who have to be fed through a tube. And probably people who are allergic to the food. And probably 1-day old babies.
What's your point? The diet is not FOR any of these people. People who have special medical conditions should consult a nutritionist. Period. Full stop.
Maybe not, but these are long-range plans. Are you really certain that the next few decades won't see the US needing to fight a high-intensity war?
1945 called, they want their military strategy back.
These days, "high-intensity wars" aren't about soldiers. They're about the president pushing a nice shiny button marked "nukes" and turning us all into crispy chicken.
I wouldn't worry too much about stupid comments in a web forum.
What you are doing sounds very much in the good old tradition of advancing science and human understanding through computation.
Whenever the word "computer science" appears in the description or the article, there's always a tedious pedant who shows up to inform everyone that "X isn't REAL computer science."
I take the position that "real computer science" is anything that the university puts in the computer science department. I went to Carnegie Mellon University and we had theory (algorithm analysis, graph theory, turing machines), discrete math (cryptography, compression), systems (software engineering, web caches, operating systems), artificial intelligence (bayesian networks, computer vision, machine learning), programming languages (proving soundness and compactness of various languages, etc), and human-computer interaction (psychology, quantifying human reaction times, etc.) There's a bunch of other sort of interdisciplinary stuff I left out like computational bio.
Just because you like one of these things more than the others doesn't mean that you are the only "real" computer scientist. There are tenured professors in all of these sub-disciplines of computer science. If you're good at what you do, I feel good for you. If not, I'm sorry for you. But don't fling poo at people who have different interests than yourself. It just makes you look like a monkey.
1. No 64 bit Flash. Or the lack of support in the X64 version of Firefox for 32 bit plug ins.
That's a Macromedia problem. Hopefully they will release their 64-bit version soon, or the GNU flash will catch up...
3. Drivers specifically the fact that it is IMPOSIBLE for a manufacture to put a binary linux driver on a disk and stick in the box with his product.
Driver disks suck. By the time you buy your hardware from WallyWorld, the driver disk has been gathering dust for a few months, and is no longer current. Bugs may have been discovered, or new features added, in a later release. Also, it's a huge pain to insert 10 different driver disks to set up your new computer.
It's much easier to just do "apt-get" or "yum," and get the drivers you need. But if you really feel nostalgic, you can ritually insert and remove a few floppies.
The other issue you alluded to, not having a fixed ABI for drivers, is actually an advantage for Linux. It can adapt and change over time much easier than any version of Windows, for exactly this reason. And the kernel maintainers keep all in-tree drivers up to date, anyway, so companies have no reason to complain.
And I get to thinking... that we have all this stuff that could be done around here. Roads, housing, bridges, new infrastructure that could be built, that sort of thing. Stuff that benefits the community.
Most people like to do only 30 minutes or an hour of exercising, then go back to whatever they had been doing. Building roads, housing, and bridges is not something that can really be done in that time frame. I agree with the other poster-- with people arriving on the construction site for 30 minutes, and then leaving, no work would get done. It would be much too hard to coordinate.
Also, you are talking about skilled labor, which means that there is a learning curve. Believe it or not, it's hard to manage skilled laborers, to ensure that everyone knows what they're supposed to know, and doesn't mess up. Some of the work could be dangerous for a newbie to do. We are talking about power tools and construction sites here.
Also, on a more practical level, construction workers and machinists have powerful unions that would oppose any kind of community involvement in the industry. They are well aware that it would cost them a lot of jobs, and they would play up the safety angle as much as they could, to ensure that the idea never got off the ground.
Probably the most practical idea for using people's exercise energy would be to attach electrical generators to certain pieces of equipment. Of course, then you get into the same problems that wind and solar energy has always had-- how do you store this energy without causing more pollution than you save? Batteries are highly toxic and relatively expensive.
There aren't a whole lot of businesses or governments eager to buy "surplus energy." Think about it. Would you sign a contract offering to power your house "only when it was convenient"? No. So perhaps you should think about this problem instead.
My 2c - pThreads suck ass. There's nothing worse than looking at 1000 Java 'processes' on your JSP server and wondering which ones will die if you kill one. It also makes it hard for the OS to figure out that App X hasn't had any CPU time for over 500ms.
You are confused. Pthreads is a cross-platform API with a bunch of different implementations, including ones for Linux, Windows, BSD, and some embedded platforms.
LinuxThreads is the crappy threading library that Linux used to use, which treats threads as processes. Luckily, new Linux systems now come with NPTL, which doesn't have this problem.
> > Basically, using flash memory for swap. Linux has been able to do this for a long time.
> This is not what ReadyBoost is doing.
Ok. I am genuinely curious about this kernel stuff, and not just looking to bash Windows!
So I looked up the paragraphs about ReadyBoost.
After the ReadyBoost service initializes caching, the Ecache.sys device driver intercepts all reads and writes to local hard disk volumes (C:\, for example), and copies any data being written into the caching file that the service created. Ecache.sys compresses data and typically achieves a 2:1 compression ratio so a 4GB cache file will usually contain 8GB of data. The driver encrypts each block it writes using Advanced Encryption Standard (AES) encryption with a randomly generated per-boot session key in order to guarantee the privacy of the data in the cache if the device is removed from the system.
When ReadyBoost sees random reads that can be satisfied from the cache, it services them from there, but because hard disks have better sequential read access than flash memory, it lets reads that are part of sequential access patterns go directly to the disk even if the data is in the cache.
You are technically correct-- Microsoft's ReadyBoost isn't quite swap. It's more like a hard drive cache to speed up random accesses to and from the disk. Typically these I/O caches are kept in memory, but I guess using a flash key would allow you to have a bigger (although slower) I/O cache.
This is a feature that, as far as I know, Linux does not have.
What is ReadyBoost doing, then?
Thanks for the clarification. Is it also used during suspend / resume? Because, at least I personally do more suspending and resuming than cold boots.
/boot and /etc partitions on the flash. Maybe you could get parts of X on there too, I'm not sure.
It seems like you could create "ReadyBoost for Linux" by writing a driver for the H-HDD flash, and putting your
You may not like them, but how many of them can Linux do? Specifically, can Linux cache to Flash devices? Does Linux automatically optimize application startup based on the last five logon sessions? (I'm not trying to be a smart ass here, I'm really curious. And if it doesn't, why not?)
ReadyBoot / ReadyBoost - Basically, using flash memory for swap. Linux has been able to do this for a long time.
However, Linux does not do this "automatically" when you insert a flash device, nor does it set the priority of the flash swap space automatically.
You could write a script that would do this pretty easily though.
Bootmgr - Microsoft replaced their NTLDR boot manager with this. NTLDR only booted variants of Windows, and refused to see any other partitions. Perhaps the new Bootmgr will be more permissive?
It's unlikely to matter, since everyone who wanted to boot something other than Windows was already running GRUB or LILO anyway.
ReadyDrive - This seems to be similar to ReadyBoost, except that the flash memory in question is on the hard disk itself. The summary mentions something called an H-HDD, which is a hard disk that has NVRAM integrated into it.
I think they are hoping to use this to improve startup time by preloading parts of the OS into this NVRAM. Other than that, the summary doesn't really give me a clear idea of how this is different from regular swap.
Actually most hard disks do have NVRAM on them even now. What's new about H-HDDs is that they expose that fact to the operating system. Whether or not this will yield important performance gains remains to be seen.
SuperFetch - This is a virtual memory system enhancement.
Basically, VM systems are being tweaked all the time. There isn't really any "right" way to do it, there's just a bunch of heuristics that work better or worse depending on your usage patterns.
The paragraphs about SuperFetch spend a long time talking about the "after lunch syndrome" whereby users have a bunch of foreground apps open (think MS Word, MS Excel, whatever...), and then they go to lunch. Then, their screensaver, or SETI@home, or whatever, begins to run, and the foreground apps are swapped out to disk. When the user comes back, he finds that the each time he tries to click on a former foreground app, Windows has to spend a long time loading the app back into RAM from swap.
In many cases, it took longer to load the application back into RAM than it did to restart the app entirely. This is not that surprising, actually. Apps tend to increase their memory use over time, by means of incremental allocations. Being hit with the total cost of those allocations all at once can be overwhelming. Many Windows apps can use gigs of memory, so it was like doing a huge file copy.
If you know that desktop computers tend to have an "on / off" usage pattern (i.e., the monkey is either at the keyboard, or not,) you can try to make your VM "clever." This might help desktop users, although server users would want to stay as far away from it as possible.
I haven't ever observed "after lunch syndrome" on my Linux desktops. I seldom even have to use virtual memory, actually. Currently, the biggest memory hogs I run are GNOME and firefox, but even they can only about half-fill my 1 gigabyte of memory. Even OpenOffice doesn't seem to take much of a bite out of it. The only app I have that consistently thrashes the cache is the Gimp.
Not much of a solution for me, I refuse to run script or bytecode.
I hope someday you change your mind. Then you'll be able to run MS Windows, Linux, or BSD, all of which execute scripts when they start up, and to accomplish many routine tasks.
If he did the surgery expertly, and the poor outcome was for reasons other than what was in his control - then that's simply not his fault...
If she had died unexpectedly as a result of the anaesthetic - statistically a 1 in 250 000 chance - then that's out of his control.
Absolutely. There's a reason why people fear "unecessary surgery."
Unfortunately, no matter how many times you warn people about the risks, they still get mad when the dice come up snake eyes.
Disclaimer: I have no idea whether the doctor in question is a good one or a bad one.
This is insightful? More like stupid.
There are lots of reasons why a car accident might not be the fault of the driver. Perhaps the driver's car was assembled improperly in the factory. Maybe someone else on the road was driving like an idiot. Maybe there was a medical emergency that nobody could have predicted.
If you have a stroke, and lose control of your body, and crash your car into a bus full of schoolchildren, you are not guilty of manslaughter. Assuming that there is no pre-existing medical condition that you were aware of, it is considered "an act of god" in legal terminology.
You cannot be held responsible for any action unless you knowingly consent to performing that action. Without intent, there can be no guilt. You aren't responsible just because "you happened to be there and something went wrong." Anything else makes a mockery out of morality.
"Citing sources" is ALL an encyclopedia can ever be. If you just "make stuff up" because you're a smart guy in general, then it's not wikipedia, it's something else-- a blog, perhaps. The goal is to provide the very minimum amount of interpretation and bias possible-- just the facts.
Yes, some sources are wrong, and some become obsolete. That's why we have this button called "edit," that allows you to change things after they've been entered in.
We can get all postmodernist and quantum mechanical, and argue about the distinction between facts and opinion, but the reality is, most people acknowlege that there is such a distinction, and find it useful.