Fair enough, you apparently know more about kidneys than I do. I was under the impression that every cell that's part of your body has an identical copy of your DNA (there are probably copying errors in a few places, and I'm excluding bacteria that live inside us as well). I did a quick Google search trying to confirm that kidneys do, in fact, have DNA that's significantly different than the rest of the body but didn't find anything in the brief time I looked. So I can avoid making an uninformed argument in the future, do you have a reference that says kidney DNA is different from the rest of the body?
Turning the question around, why does a fetus that has been expelled through the vagina in a process we like to call birth have anymore rights than a finger that was severed in an accident with a table saw?
There could be one, but to avoid classifying my kidney, semen, or pet as a person and making it illegal to kill them, it would have to do some absurd convolutions.
Your kidney has only your own human DNA. Your semen has half of your own human DNA, and it is missing the other half. Your pet does not have any human DNA. A fetus has a complete set of human DNA, half of which is different from that of the woman carrying him or her. Which of these convolutions were terribly absurd?
I played a lot of SimCity as a kid. I don't know of any way that I could quantify the educational value of the time I spent playing with it, but I wouldn't be surprised if it was about as educational to me as playing with Legos was. My gut feeling is this is worth it, especially given that by today's standards, the resources it requires are practically nothing.
I'm sorry, I can't use that word processor. It doesn't support my video card?
That's the issue of I was thinking about this. It sounds an awful lot like the old DOS days, where every application or game had to include its own drivers for whatever hardware it was going to use. Luckily, these days real operating systems abstract all this so there can be one driver that works for all applications that run on that OS. How is this OS-less computer dealing with these issues?
Maxing out an ethernet connection doesn't take much CPU.
This is actually less true than most people think. Let's say a 3GHz CPU and a gigabit ethernet connection is fairly common right now. In this setup, to max out the network connection, you have to send one bit every three CPU cycles. Granted, people don't send data one bit at a time, but even to send 1500 bytes, which is the normal MTU for ethernet, you currently get about a 36,000 CPU cycle budget. By the time you factor in the TCP/IP overhead, possible compression and encryption, most non-trivial network protocols will end up being CPU-limited, rather than network-limited.
Historically, the Cycle/Bit ratio oscillates from less than 0.5 to around 10. When you're at the 10 cycle/bit range, it's trivial to fill a network pipe, but it's incredibly difficult to max out the network connection. This is why to get really high network performance, people start talking about TCP/IP offload engines and zero-copy network support.
Keep in mind that when you start to get into the 10GBit ethernet range, your network bandwidth is actually approaching the bandwidth of your RAM.
This seems like a fairly good idea, if it could be integrated with the computer classes and such that most students already take. When I was in school, we had to sign some paper saying we wouldn't do naughty things in order to use the school's internet service. This could be an extension of that same idea.
To me, it seems like educating people is the best way to stop the spread of computer viruses and other malware. While some of these things spread through security flaws in operating systems, I would guess the majority are spread through people clicking things like "FREE PONIEZ" links. How many of us slashdotters have antivirus or other anti-malware software on our computers? I've found that in many cases, you can achieve an acceptable level of safety more effectively just by knowing what you're doing, rather than having all this extra stuff running in the background to protect you. The best way to protect people isn't through technological solutions, but through education, so this seems like the right approach, if done well.
You have choices, you just may not like them better, which is why they can get away with charging you $20 extra. Personally, I don't even have a TV, let alone cable. I decided I do like to be able to watch certain shows and such, so I recently got a Netflix subscription and I'm having them ship me a lot of TV series and things. It seems to work well for me, and for less than the $20 a month Comcast just raised your rate. Of course, you don't get the TV shows until they come out on DVD, and you don't get the evening news or sports and things, which may be important for you, but I make do without them.
I had almost the opposite thought on this. "Wait, you mean that was only in 1997?" 1997 still doesn't seem that long ago to me. I thought this happened at least 15 years ago, and possibly 20.
I thought almost this exact same thing a few weeks ago, that I'll just be glad when I don't have to hear people complain about the president anymore. Of course, people will still complain, so not much will change. Anyway, I decided I'm pretty disillusioned with both parties right now, so I'm going to try to find a third party candidate worth voting for.
RISC binaries tend to be larger than CISC binaries. The reason is the complex instruction set, like in the x86 architecture, were made complex to save memory. Most of the common instructions are represented in only one byte, while the rarer instructions can be much much longer. RISC instruction sets, on the other hand, typically have a fixed instruction size for all instructions, and the average instruction length ends up being longer.
While most people have plenty of hard drive space now, and RAM isn't as scarce as it used to be, CPU caches are still pretty small. Using a more compact instruction set can make the code caches more effective, which can dramatically improve performance.
I doubt this completely justifies choosing CISC over RISC, but it is at least one piece of information in favor of sticking with CISC.
It's probably a matter of perspective. As developers, we certainly should try for bug-free software. As people that care about the security of our machine, you should assume that any piece of software has unknown security flaws in it. Absolutes like "unbreakable" and "perfectly secure" have no business being used. Try to write bug-free, secure software, but it is unwise to ever claim you've accomplished it.
Andy Wilson, one of the guys at Microsoft Research who seems to have done a lot of the research work that let to this table product, got his PhD at MIT, so it's not unreasonable to think that Andy Wilson was probably involved on similar multitouch interfaces at MIT. At any rate, this is taking a research prototype and bringing a product based on it to market. I'm guessing the work at MIT was not ready to be commercialized yet, so it's unfair to complete discount what Microsoft has done here.
It's an interesting idea, but I'm not sure it's any better than switching to IPv6. If I understand your proposal right, when you ask a DNS server for an address, it gives you the IP address, and an address for the router, which will act like a proxy server? To open a connection, you'd connect to the router, then encapsulate all the packets destined for the foreign country in traffic for that router, which would forward it on.
Somehow the client has to keep track of whether 137.17.37.43 means the Australian version, or the German version, so it knows which router to send packets destined to 137.17.37.43. In essence, the clients will have to add a few more bits on to the address space. Eight bits should be enough to get all of the countries, so we've moved from a 32-bit address space to a 40-bit address space.
I'm not sure how you can do this without changing or introducing a new protocol. If we're going to be changing the protocol anyway, we might as well go all the way to IPv6. IPv6 isn't that far away anyway. Windows Vista supports it, I think MacOS X supports it, and various other Unix flavors have supported it for a very long time. I'd imagine most routers support it too, so really we're just waiting for the ISPs to start assigning everyone IPv6 addresses instead of IPv4 addresses. The infrastructure is probably mostly there already.
Check out the TouchLight video on Andy Wilson's web page. It's strikingly similar to the Minority Report displays, and it doesn't even need any crazy glowing gloves to use it.
I think the reason why we don't all start accelerating at the same time when the light turns green is that we normally stop closer to the car in front of us than we are comfortable driving. I frequently stop less than a car length behind the one in front of me at a light, but I don't feel comfortable driving that close behind a car. The reason for the lag is to leave a comfortable distance between you and the person in front of you.
A few months ago I was thinking about doing something like this, but in heavy traffic situations. What if the course had way more cars than should ever actually fit, and the cars independently tried to minimize their travel time around the course. I wonder if the computer could get a better overall throughput than people seem to do on the crowded highways of, say, Seattle.
In other news, Microsoft researchers have discovered that the sky is blue and that water is wet.
Discovering that the sky is blue is quite a discovery for a company based near Seattle. They should have known about water though, given all the rain they get.
Regarding Scheme being inefficient, I heard an anecdote once where a class was given a competition to make the fastest implementation of a particular multiplication algorithm. Most of the class decided to start coding in assembly right away, since obviously assembly is the best way to write really fast code. About two weeks into the assignment, most of the students still didn't have their algorithm working. There was one student, on the other hand, who recognized that the recursive nature of this particular algorithm was particularly well suited for a language like Scheme, so he coded his solution in Scheme and had it working perfectly within a few days. He then devised a correctness-preserving transform that turned his program into a really simple subset of C, which the optimizing compiler was able to generate really fast code for. He ended up getting second place, being just a little behind first place, and far ahead of third place. Sure, it's not pure Scheme, but this type of programming is not inherently slow.
DysFunctional Programming has practical benefits? I thought it was only useful for lazy-evaluated sigs.
Functional programming does have practical benefits. If your code doesn't have any side effects (pure-functional code doesn't) then everything is trivially parallelizable. Google took advantage of this to make MapReduce. I think this kind of thing is going to become more and more important as we move into a world where we have hundreds of CPU cores instead of just one.
So why don't we all just keep the clocks an hour ahead, and get that "extra hour" all year round?
I live in Seattle, and up here daylight savings time should be reversed if we keep it at all. The winter days are so short anyway that it's quite easy to go to work while it's dark and go home while it's dark. Unless you have a window office, it's easy to go the whole day without even seeing the sunlight. It's usually totally dark by about 4:30 in the winter. In the summer time, however, the days are really long, so it will still be light out as late as 10pm. Personally I'd rather have it get dark at 9pm in the summer and stay light out until 5:30 in the winter.
Slashdot Mirror
Fair enough, you apparently know more about kidneys than I do. I was under the impression that every cell that's part of your body has an identical copy of your DNA (there are probably copying errors in a few places, and I'm excluding bacteria that live inside us as well). I did a quick Google search trying to confirm that kidneys do, in fact, have DNA that's significantly different than the rest of the body but didn't find anything in the brief time I looked. So I can avoid making an uninformed argument in the future, do you have a reference that says kidney DNA is different from the rest of the body?
Turning the question around, why does a fetus that has been expelled through the vagina in a process we like to call birth have anymore rights than a finger that was severed in an accident with a table saw?
Your kidney has only your own human DNA. Your semen has half of your own human DNA, and it is missing the other half. Your pet does not have any human DNA. A fetus has a complete set of human DNA, half of which is different from that of the woman carrying him or her. Which of these convolutions were terribly absurd?
I played a lot of SimCity as a kid. I don't know of any way that I could quantify the educational value of the time I spent playing with it, but I wouldn't be surprised if it was about as educational to me as playing with Legos was. My gut feeling is this is worth it, especially given that by today's standards, the resources it requires are practically nothing.
That's the issue of I was thinking about this. It sounds an awful lot like the old DOS days, where every application or game had to include its own drivers for whatever hardware it was going to use. Luckily, these days real operating systems abstract all this so there can be one driver that works for all applications that run on that OS. How is this OS-less computer dealing with these issues?
Depending on how much counts as "the kernel," I wouldn't be surprised if even the kernel has changed more over the last 20 years than Paint has.
I'd mod you up if I had mod points. I get so tired of hearing about "Bush's ban on stem cell research," because the thing truly does not exist.
Mark Russinovich posted a fairly in-depth and technical explanation of this problem. It's worth taking a look at.
2 007/08/27/1833290.aspx
http://blogs.technet.com/markrussinovich/archive/
This is actually less true than most people think. Let's say a 3GHz CPU and a gigabit ethernet connection is fairly common right now. In this setup, to max out the network connection, you have to send one bit every three CPU cycles. Granted, people don't send data one bit at a time, but even to send 1500 bytes, which is the normal MTU for ethernet, you currently get about a 36,000 CPU cycle budget. By the time you factor in the TCP/IP overhead, possible compression and encryption, most non-trivial network protocols will end up being CPU-limited, rather than network-limited.
Historically, the Cycle/Bit ratio oscillates from less than 0.5 to around 10. When you're at the 10 cycle/bit range, it's trivial to fill a network pipe, but it's incredibly difficult to max out the network connection. This is why to get really high network performance, people start talking about TCP/IP offload engines and zero-copy network support.
Keep in mind that when you start to get into the 10GBit ethernet range, your network bandwidth is actually approaching the bandwidth of your RAM.
This seems like a fairly good idea, if it could be integrated with the computer classes and such that most students already take. When I was in school, we had to sign some paper saying we wouldn't do naughty things in order to use the school's internet service. This could be an extension of that same idea.
To me, it seems like educating people is the best way to stop the spread of computer viruses and other malware. While some of these things spread through security flaws in operating systems, I would guess the majority are spread through people clicking things like "FREE PONIEZ" links. How many of us slashdotters have antivirus or other anti-malware software on our computers? I've found that in many cases, you can achieve an acceptable level of safety more effectively just by knowing what you're doing, rather than having all this extra stuff running in the background to protect you. The best way to protect people isn't through technological solutions, but through education, so this seems like the right approach, if done well.
You have choices, you just may not like them better, which is why they can get away with charging you $20 extra. Personally, I don't even have a TV, let alone cable. I decided I do like to be able to watch certain shows and such, so I recently got a Netflix subscription and I'm having them ship me a lot of TV series and things. It seems to work well for me, and for less than the $20 a month Comcast just raised your rate. Of course, you don't get the TV shows until they come out on DVD, and you don't get the evening news or sports and things, which may be important for you, but I make do without them.
I had almost the opposite thought on this. "Wait, you mean that was only in 1997?" 1997 still doesn't seem that long ago to me. I thought this happened at least 15 years ago, and possibly 20.
I thought almost this exact same thing a few weeks ago, that I'll just be glad when I don't have to hear people complain about the president anymore. Of course, people will still complain, so not much will change. Anyway, I decided I'm pretty disillusioned with both parties right now, so I'm going to try to find a third party candidate worth voting for.
RISC binaries tend to be larger than CISC binaries. The reason is the complex instruction set, like in the x86 architecture, were made complex to save memory. Most of the common instructions are represented in only one byte, while the rarer instructions can be much much longer. RISC instruction sets, on the other hand, typically have a fixed instruction size for all instructions, and the average instruction length ends up being longer.
While most people have plenty of hard drive space now, and RAM isn't as scarce as it used to be, CPU caches are still pretty small. Using a more compact instruction set can make the code caches more effective, which can dramatically improve performance.
I doubt this completely justifies choosing CISC over RISC, but it is at least one piece of information in favor of sticking with CISC.
It's probably a matter of perspective. As developers, we certainly should try for bug-free software. As people that care about the security of our machine, you should assume that any piece of software has unknown security flaws in it. Absolutes like "unbreakable" and "perfectly secure" have no business being used. Try to write bug-free, secure software, but it is unwise to ever claim you've accomplished it.
Andy Wilson, one of the guys at Microsoft Research who seems to have done a lot of the research work that let to this table product, got his PhD at MIT, so it's not unreasonable to think that Andy Wilson was probably involved on similar multitouch interfaces at MIT. At any rate, this is taking a research prototype and bringing a product based on it to market. I'm guessing the work at MIT was not ready to be commercialized yet, so it's unfair to complete discount what Microsoft has done here.
It's an interesting idea, but I'm not sure it's any better than switching to IPv6. If I understand your proposal right, when you ask a DNS server for an address, it gives you the IP address, and an address for the router, which will act like a proxy server? To open a connection, you'd connect to the router, then encapsulate all the packets destined for the foreign country in traffic for that router, which would forward it on.
Somehow the client has to keep track of whether 137.17.37.43 means the Australian version, or the German version, so it knows which router to send packets destined to 137.17.37.43. In essence, the clients will have to add a few more bits on to the address space. Eight bits should be enough to get all of the countries, so we've moved from a 32-bit address space to a 40-bit address space.
I'm not sure how you can do this without changing or introducing a new protocol. If we're going to be changing the protocol anyway, we might as well go all the way to IPv6. IPv6 isn't that far away anyway. Windows Vista supports it, I think MacOS X supports it, and various other Unix flavors have supported it for a very long time. I'd imagine most routers support it too, so really we're just waiting for the ISPs to start assigning everyone IPv6 addresses instead of IPv4 addresses. The infrastructure is probably mostly there already.
Check out the TouchLight video on Andy Wilson's web page. It's strikingly similar to the Minority Report displays, and it doesn't even need any crazy glowing gloves to use it.
I think the reason why we don't all start accelerating at the same time when the light turns green is that we normally stop closer to the car in front of us than we are comfortable driving. I frequently stop less than a car length behind the one in front of me at a light, but I don't feel comfortable driving that close behind a car. The reason for the lag is to leave a comfortable distance between you and the person in front of you.
A few months ago I was thinking about doing something like this, but in heavy traffic situations. What if the course had way more cars than should ever actually fit, and the cars independently tried to minimize their travel time around the course. I wonder if the computer could get a better overall throughput than people seem to do on the crowded highways of, say, Seattle.
Discovering that the sky is blue is quite a discovery for a company based near Seattle. They should have known about water though, given all the rain they get.
Regarding Scheme being inefficient, I heard an anecdote once where a class was given a competition to make the fastest implementation of a particular multiplication algorithm. Most of the class decided to start coding in assembly right away, since obviously assembly is the best way to write really fast code. About two weeks into the assignment, most of the students still didn't have their algorithm working. There was one student, on the other hand, who recognized that the recursive nature of this particular algorithm was particularly well suited for a language like Scheme, so he coded his solution in Scheme and had it working perfectly within a few days. He then devised a correctness-preserving transform that turned his program into a really simple subset of C, which the optimizing compiler was able to generate really fast code for. He ended up getting second place, being just a little behind first place, and far ahead of third place. Sure, it's not pure Scheme, but this type of programming is not inherently slow.
Functional programming does have practical benefits. If your code doesn't have any side effects (pure-functional code doesn't) then everything is trivially parallelizable. Google took advantage of this to make MapReduce. I think this kind of thing is going to become more and more important as we move into a world where we have hundreds of CPU cores instead of just one.
Finally, a way to keep track of the positions of all of my sharks with lasers on their heads!
I live in Seattle, and up here daylight savings time should be reversed if we keep it at all. The winter days are so short anyway that it's quite easy to go to work while it's dark and go home while it's dark. Unless you have a window office, it's easy to go the whole day without even seeing the sunlight. It's usually totally dark by about 4:30 in the winter. In the summer time, however, the days are really long, so it will still be light out as late as 10pm. Personally I'd rather have it get dark at 9pm in the summer and stay light out until 5:30 in the winter.
When I camp, a lot of times I don't even bring a clock.