The definition of the kilogram was originally made in terms of a particular volume of water, but was later changed to the weight of a particluar ingot of irridium-platinum.
So when they sequence the human genome, how do they handle the variations?
For the purposes of creating the reference sequence they essentially ignored them. In the public human genome project the DNA from a handful of individuals was used. The Celera project used mostly the DNA of one individual, Craig Venter, the head of Celera. This does make the reference sequence arbitrary, but so was the block of platinum that was used to define the kilogram. The idea is that you measure differences from the standard.
The rule of thumb is that the sequence of any two individualss differ in about 1 base in 1000. This ignores complications like that fact that women have of two copies of the X chromosome and men have 1 X and 1 Y chromosome, and that whole sections of sequence can sometimes get shifted from one chromosome to another. As the other responder pointed out the variations are a major focus of research, particularly Single Nucleotide Polymorphisms (SNPs) where 90% of the population have an 'A' in a particulary position and 10% have a 'G'.
Yes, but since I mentioned that I've been working with UNIX systems of various flavors since 1976, a reasonable person might give my perception some consideration. I stand open to correction, but you haven't provided any further information.
Let me try to make my question more clear: Is there some feature set, some API, that OS X doesn't implement that makes it 'NOT UNIX'. Or perhaps you are refering to the fact that it is not 'UNIX(TM)'? If so, do you also correct people when they refer to "Kleenex(TM)" or "Xerox(TM)"? If Leopard Server is certified by the Open Group will it become UNIX with the stroke of a pen?
Sorry, all equivocation aside, it's NOT UNIX. It's an ugly hack on BSD.
OK, I'll bite: in what sense is it not UNIX? I've been working with various flavors of UNIX since v7. OS X cetainly feels like UNIX to me. I switched to a iBook as my laptop when I moved from a Windows shop to a Linux shop where I do 'C' and Python development. I move my source back and forth between Linux and OS X daily. My understanding is that the threading model on OS X sucks, so it makes a poor server platform, buy what makes it 'NOT UNIX'?
I do beleave they do have a nuke, as in all honesty a nuke is first year stough in university.
You have been watching too many movies and taking them way too seriously. This is like saying that a 1st year physics student could build a fighter jet because they've read a chapter on Bernoulli's principle. On a purely pragmatic basis, plutonium and uranium have some really awful mechanical properties, and most 1st year physics students don't even know how to run a metal lathe, let alone know how to machine metal that's prone to catching fire on contact with air. N. Korea may have a nuke, but that's because they have the engineering know-how and industrial base to build things like fighter jets, not because building nukes is so easy.
What possible use is it to be able to sequence that many individuals of a single species in a short time? The only practical value worth that kind of money is genetic profiling; sort of the Nazi eugenics approach to social purification, but on steroids, with none of that messy subjective stuff like whose brow sticks out the furthest, or nose is largest, or jaw is most rugged.
If you have a technology that can sequence that many individuals in that short a time, then you are in spitting distance of making genomics a clinical tool. Think: "You have disease X. We generally treat disease X with drug Y, but it turns out that you have genotype A, and drug Y is ineffective in people with genotype A, and in fact will kill 10% of them. We're going to give you drug Z instead." Or, "You have genotype B. People with genotype B who are exposed to benzene are 100,000 times more likely to develop cancer the people with genotype A. Stay away from benzene!" Some specific tests like this are done now, but usually only after a problem has developed. The prize is motivated by wanting to have a cheap multipurpose screen.
o, asides from the coolness factor, what are the real advantages of decoding your genes?
It's the ability to quickly spot genetic variations that's important. For example, it may turn out that a small genetic variation partially determines the effectiveness of chemotherapy for a particular type of cancer. Say a particular chemotherapy shrinks the tumor in 60% of people who have an 'A' at position 12342245 on chromosome 1, but is completely ineffective in people who have a 'T' at that position. If you were in the unfortunately position of having to decide whether or not you were going to do chemotherapy, you would probably want to know whether you were an 'A' or a 'T'.
My work is my work, and if I choose to reuse it in a similar situation, this not only demonstrates that I understood the assignment, but that I recognized that I had already done the assignment. It is a mark of intelligence to recognize this.
You are so not getting the point of classwork. After it's graded, the final product of your work is usually irrelevant. The point is to make you go through the exercise of creating the final product, thereby gaining insight and practice into the underlying principles. As much as most students want to deny it, repetition is an important tool in learning. If you were weight training at a gym, you wouldn't object: "But I lifted those weights yesterday! It's a fools errand to lift them again today.". As the parent poster pointed out, in any given assignment the instructor may be fine with your taking advantage of your previous work, you just have to be honest that you are doing so.
All right, so you're a theorist. I'm an experimentalist. I can accept that it may violate the laws of physics as we know them. That would only mean that we need to rethink a couple of things. (Long term that's a given anyway.) All I want to see is a working model verified by a third party.
But look at what Shawyer has actually provides: he says he has a working model, but what he publishes is theory about why his device should work, not a detailed protocol of his experiments. At this point it's his theory against "establishment" theory, and the establishment theory is the one that's got all the published experimental evidence.
The New Scientist article makes much of the fact that major corporations are looking at his work, presumably to give us confidence that competent people are reviewing his work. I am put in mind though of several major electronics firms that embarassed themselves investing in a plan to send video over plain old phone lines. See for example: VisionTek
What do you guys think of a RAID setup as an alternative to explicitly backing up files for a home user?
Not much. You probably won't loose your data due to a single hard disk crash, but if your computer really gets clobbered (fire, flood, theft) you are hosed. Of course, it depends on how attached you are to the contents of your hard disk. I've got decades of correspondence and development work, as well as years of pictures. I'd be really sad if I lost those, so I backup every week to an Iomega Rev drive, and rotate a copy every month to my office at work.
On a related note, ever notice how all the research and development facilities in the United States are in the West and the South (the land of the troglodytes), away from the vaunted Ivy-League colleges of the East coast, the land of the enlightened?
No, I can't say that I have. In fact your statement is demonstrably false. Try looking up the following research laboratories on a map: Courant Institute, MIT Lincoln Laboratory, Cold Spring Harbor, The Jackson Laboratory, Brookhaven National Laboratory, Argonne National Laboratory, Fermi National Accelerator Laboratory, Bell Labs, IBM Research Headquarters, Kodak Research Laboratories, Westinghouse Research. Those are just the ones I'm familiar with off the top of my head, and I live in the West. Bell Labs, IBM, Kodak, and Westinghouse are among the largest and most famous R&D labs in the world. Do you often make blanket assertions about fields you know nothing about?
Re:School can help, in more ways than the obvious.
on
How Old is Too Old?
·
· Score: 1
you are overly concerned with Big O Notation before your code is even functional
Ye'gads! You'd better be thinking about the Big O of your code before it is functional, because if you don't, you stand a very good chance of having to throw a lot of that functional code away.
You have a point in that most programming jobs don't need the stuff you learn in a computer science program. Most programming jobs are all about scooping data out of a database, displaying it on a screen, providing a UI so the user can make changes to the data, and then schlepping the data back to the database. Nothing wrong with that, and you don't usually need to know about asymptotic complexity, or finite state automata in those jobs. What you are missing though, is that there are programming jobs that do make use of that whacky esoteric stuff. If you want to work on physics engines, database servers, web servers, operating systems, compilers, ai, or bioinformatics, then a computer science degree is pretty useful.
Those equations seems different than the description of a turing machine
Turing was a very talented mathematician, and worked in several areas besides the theory of computation and cryptograhpy. The equations discussed in the article (reaction-diffusion equations) are partial differential equations that model chemical reactions. They don't have anything to do with Turing machines, other then the fact that you could use a Turing machine to solve them numerically.
One of the major contributing factors to the S&L failures was defaults on commercial paper (junk bonds)
My mistake here. Junk bonds are not commercial paper. None the less, I think my point stands. Companies do default on their commerical paper. See for example this Moody's report.
They NEVER go down in value - they are usually priced on a par value of $10/unit, and you get your $10/unit back (unless there's been a complete financial meltdown), plus whatever interest has accrued.
It's worth remembering that "complete financial meltdowns" occured at least twice in the US in the last century: the bank failures of the Great Depression, and the Savings & Loan failures of the 1980s. One of the major contributing factors to the S&L failures was defaults on commercial paper (junk bonds). Most folks were saved from principal losses on S&L accounts only by federal insurance. Money market mutual funds are not FDIC insured.
The original poster should understand that banks and funds can declare bankruptcy and walk away from whatever they owe him, but he cannot walk away from his student loan. Student loans are not cleared by bankruptcy. Barring death, disability, or the collapse of the federal government he will have to repay his student loans.
If you are patient and have good timing,you can usually walk out of such a deal with a 50-100% return on your intial investment.
I have seen folks get rich doing this. I have a couple of friends who lost everything. That's investing for you. Remember though that is advice for a student, who is using his student loans for capital. Patience and timing don't enter in to it for him. Six months after he leaves school he has got to start making interest payments at the very least, no matter what the state of the real estate market. There certainly an element of savvy in "good timing", but there is also a huge amount of luck.
I'd also like to remind everybody that bankruptcy will not clear your student loans!
A money market's rate of increase can go down with the stock market, but you won't lose your principal. Basically it protects you from inflation.
You certainly can loose your principal. Money market funds are sending your money out somewhere, and if it doesn't come back, you're screwed. Weren't you around for the S&L bust during the 1980s? Millions of folks would have lost their principal in that fiasco if it wasn't for federal insurance bailing them out. Almost all money market ads have big bold print on their advertising saying Not FDIC Insured. I'm not saying this is likely, but since his principal is borrowed money that he absolutely will have to pay back, the risk should be considered.
The essay in question was looking at a sample of peer reviewed research. They were not claiming that there are no dissenting publications, there certainly are. They were simply making the point that the consensus for anthropogenic warming is very strong. As in any other dispute that revolves around interpretations of statistical data, consensus is by no means a guarantee, but it is frequently the smart way to bet. The original essay (which was not peer reviewed, but an invited essay) was in Science 3 December 2004: Vol. 306. no. 5702, p. 1686
No, msvcrt.dll is the dynamically linked version of the C runtime library. If you used dynamic linking it has to be present at run time. If you use static linking it isn't used at compile time or at runtime.
Here is the deal: many languages, including C and Visual Basic, depend on a library of commonly used subroutines. Some of the subroutines you are probably aware of, say malloc or strncpy. Others work behind the scenes to set up the heap and such. You can link to this library statically, in which case the code for those subroutines is pulled into your executable. Or, you can link to the library dynamically, in which case only placeholders go into your executable. Then, when your program is run, it loads the dynamic version of the library into memory and fixes up the stubs. The advantage of this is that all dynamically linked program can share a single copy of the library in memory. The disadvantage is that your program depends on the user having a copy of the dynamic library.
All of the major C compilers provide both static and dynamic versions of the C runtime. In the GNU world the For some reason Microsoft only provides a dynamic version of the VB runtime. The VB runtime also contains the code for the p-code interpreter but this isn't called when native-code generation is turned on.
This was true for the older versions of VB 4 and earlier, but VB 5 can generate native code. Yes, you still need the run time library, just as most applications writen in C need a runtime library like libgc or msvcrt. The fly in the ointment is that Microsoft doesn't provide a static version of the VB runtime library.
Since you have a degree in education, perhaps you can offer me some insight into a problem that has long troubled me. It seems to me that educators are too prone to jumping on band wagons and fighting over them with a tenacity that puts "vi vs emacs" flamewars to shame. For example, I've seen school districts just about disintegrate in the debate over whole language vs phonetics. As a consumer of education it seemed to me that both schools had their points, and that children would benefit from being exposed to both methods. That seemed unacceptable to many educators though who seemed to demand that educators hew to one method or the other.
I was educated under the "new math" paradigm in the sixties, and I many others absolutely loved it. However, other kids didn't do at all well, and so "new math" was consigned to the rubbish heap. Kids obviously differ in their learning styles, why do we have to have a monolithic approach to instructional design?
I'm not clear whether this is a matter of helping out an aquaintance or of going the extra mile for a loyal customer. Either way, I think the answer is simply to tell them that prior commitments and obligations prevent you from helping them out. If it is a close friend, you can give them as much detail as they'll stand to listen to. If it is a business relationship, I'd avoid giving them Too Much Information, all they really want to know is whether you can help or not.
Only in the sense that riding in a airplane is aeronautical engineering. To be utterly pedantic, I suppose it is necessary to understand union, intersection, and cartesian product in writing queries, but surely you covered those in the first 10 minutes of your discrete math course! I can only observe that I've got an M.S. in math and at one point I was working with a very smart gentleman who learned SQL at a trade school. He didn't know discrete math from the indiscrete charm of the bourgoise, but when I wanted help with a tricky query he was the guy to talk to. If you could give me an example of a theorem from discrete math that actually helped you frame a query, I'd be delighted to hear about it.
A UI will definately contain "Graph Theory"
An example would be helpful here too, because I have no idea what you are talking about. Do you understand about that graph theory is not about computer graphics? Roughly, a graph is a set, and a symetric relationship defined on that set. If you haven't run into it yet in discrete math class, you soon will. Trees are a subcategory of graphs, and a variety of data structures that model trees are teriffically useful in software. The developer of the GUI internals may very well have used graphs and trees to store the elements of the UI, but the typical workaday programmer adding a list box to a web page is no more using graph theory then I'm using fluid mechanics when I'm flying coach across the atlantic.
Slashdot Mirror
The definition of the kilogram was originally made in terms of a particular volume of water, but was later changed to the weight of a particluar ingot of irridium-platinum.
For the purposes of creating the reference sequence they essentially ignored them. In the public human genome project the DNA from a handful of individuals was used. The Celera project used mostly the DNA of one individual, Craig Venter, the head of Celera. This does make the reference sequence arbitrary, but so was the block of platinum that was used to define the kilogram. The idea is that you measure differences from the standard.
The rule of thumb is that the sequence of any two individualss differ in about 1 base in 1000. This ignores complications like that fact that women have of two copies of the X chromosome and men have 1 X and 1 Y chromosome, and that whole sections of sequence can sometimes get shifted from one chromosome to another. As the other responder pointed out the variations are a major focus of research, particularly Single Nucleotide Polymorphisms (SNPs) where 90% of the population have an 'A' in a particulary position and 10% have a 'G'.
So, as you have repeatedly been asked, which criteria does OS X fail to meet, that makes it 'NOT UNIX'?
Yes, but since I mentioned that I've been working with UNIX systems of various flavors since 1976, a reasonable person might give my perception some consideration. I stand open to correction, but you haven't provided any further information.
Let me try to make my question more clear: Is there some feature set, some API, that OS X doesn't implement that makes it 'NOT UNIX'. Or perhaps you are refering to the fact that it is not 'UNIX(TM)'? If so, do you also correct people when they refer to "Kleenex(TM)" or "Xerox(TM)"? If Leopard Server is certified by the Open Group will it become UNIX with the stroke of a pen?
OK, I'll bite: in what sense is it not UNIX? I've been working with various flavors of UNIX since v7. OS X cetainly feels like UNIX to me. I switched to a iBook as my laptop when I moved from a Windows shop to a Linux shop where I do 'C' and Python development. I move my source back and forth between Linux and OS X daily. My understanding is that the threading model on OS X sucks, so it makes a poor server platform, buy what makes it 'NOT UNIX'?
You have been watching too many movies and taking them way too seriously. This is like saying that a 1st year physics student could build a fighter jet because they've read a chapter on Bernoulli's principle. On a purely pragmatic basis, plutonium and uranium have some really awful mechanical properties, and most 1st year physics students don't even know how to run a metal lathe, let alone know how to machine metal that's prone to catching fire on contact with air. N. Korea may have a nuke, but that's because they have the engineering know-how and industrial base to build things like fighter jets, not because building nukes is so easy.
If you have a technology that can sequence that many individuals in that short a time, then you are in spitting distance of making genomics a clinical tool. Think: "You have disease X. We generally treat disease X with drug Y, but it turns out that you have genotype A, and drug Y is ineffective in people with genotype A, and in fact will kill 10% of them. We're going to give you drug Z instead." Or, "You have genotype B. People with genotype B who are exposed to benzene are 100,000 times more likely to develop cancer the people with genotype A. Stay away from benzene!" Some specific tests like this are done now, but usually only after a problem has developed. The prize is motivated by wanting to have a cheap multipurpose screen.
It's the ability to quickly spot genetic variations that's important. For example, it may turn out that a small genetic variation partially determines the effectiveness of chemotherapy for a particular type of cancer. Say a particular chemotherapy shrinks the tumor in 60% of people who have an 'A' at position 12342245 on chromosome 1, but is completely ineffective in people who have a 'T' at that position. If you were in the unfortunately position of having to decide whether or not you were going to do chemotherapy, you would probably want to know whether you were an 'A' or a 'T'.
You are so not getting the point of classwork. After it's graded, the final product of your work is usually irrelevant. The point is to make you go through the exercise of creating the final product, thereby gaining insight and practice into the underlying principles. As much as most students want to deny it, repetition is an important tool in learning. If you were weight training at a gym, you wouldn't object: "But I lifted those weights yesterday! It's a fools errand to lift them again today.". As the parent poster pointed out, in any given assignment the instructor may be fine with your taking advantage of your previous work, you just have to be honest that you are doing so.
But look at what Shawyer has actually provides: he says he has a working model, but what he publishes is theory about why his device should work, not a detailed protocol of his experiments. At this point it's his theory against "establishment" theory, and the establishment theory is the one that's got all the published experimental evidence.
The New Scientist article makes much of the fact that major corporations are looking at his work, presumably to give us confidence that competent people are reviewing his work. I am put in mind though of several major electronics firms that embarassed themselves investing in a plan to send video over plain old phone lines. See for example: VisionTek
Not much. You probably won't loose your data due to a single hard disk crash, but if your computer really gets clobbered (fire, flood, theft) you are hosed. Of course, it depends on how attached you are to the contents of your hard disk. I've got decades of correspondence and development work, as well as years of pictures. I'd be really sad if I lost those, so I backup every week to an Iomega Rev drive, and rotate a copy every month to my office at work.
No, I can't say that I have. In fact your statement is demonstrably false. Try looking up the following research laboratories on a map: Courant Institute, MIT Lincoln Laboratory, Cold Spring Harbor, The Jackson Laboratory, Brookhaven National Laboratory, Argonne National Laboratory, Fermi National Accelerator Laboratory, Bell Labs, IBM Research Headquarters, Kodak Research Laboratories, Westinghouse Research. Those are just the ones I'm familiar with off the top of my head, and I live in the West. Bell Labs, IBM, Kodak, and Westinghouse are among the largest and most famous R&D labs in the world. Do you often make blanket assertions about fields you know nothing about?
Ye'gads! You'd better be thinking about the Big O of your code before it is functional, because if you don't, you stand a very good chance of having to throw a lot of that functional code away.
You have a point in that most programming jobs don't need the stuff you learn in a computer science program. Most programming jobs are all about scooping data out of a database, displaying it on a screen, providing a UI so the user can make changes to the data, and then schlepping the data back to the database. Nothing wrong with that, and you don't usually need to know about asymptotic complexity, or finite state automata in those jobs. What you are missing though, is that there are programming jobs that do make use of that whacky esoteric stuff. If you want to work on physics engines, database servers, web servers, operating systems, compilers, ai, or bioinformatics, then a computer science degree is pretty useful.
Turing was a very talented mathematician, and worked in several areas besides the theory of computation and cryptograhpy. The equations discussed in the article (reaction-diffusion equations) are partial differential equations that model chemical reactions. They don't have anything to do with Turing machines, other then the fact that you could use a Turing machine to solve them numerically.
My mistake here. Junk bonds are not commercial paper. None the less, I think my point stands. Companies do default on their commerical paper. See for example this
Moody's report.
It's worth remembering that "complete financial meltdowns" occured at least twice in the US in the last century: the bank failures of the Great Depression, and the Savings & Loan failures of the 1980s. One of the major contributing factors to the S&L failures was defaults on commercial paper (junk bonds). Most folks were saved from principal losses on S&L accounts only by federal insurance. Money market mutual funds are not FDIC insured.
The original poster should understand that banks and funds can declare bankruptcy and walk away from whatever they owe him, but he cannot walk away from his student loan. Student loans are not cleared by bankruptcy. Barring death, disability, or the collapse of the federal government he will have to repay his student loans.
I have seen folks get rich doing this. I have a couple of friends who lost everything. That's investing for you. Remember though that is advice for a student, who is using his student loans for capital. Patience and timing don't enter in to it for him. Six months after he leaves school he has got to start making interest payments at the very least, no matter what the state of the real estate market. There certainly an element of savvy in "good timing", but there is also a huge amount of luck.
I'd also like to remind everybody that bankruptcy will not clear your student loans!
The essay in question was looking at a sample of peer reviewed research. They were not claiming that there are no dissenting publications, there certainly are. They were simply making the point that the consensus for anthropogenic warming is very strong. As in any other dispute that revolves around interpretations of statistical data, consensus is by no means a guarantee, but it is frequently the smart way to bet. The original essay (which was not peer reviewed, but an invited essay) was in Science 3 December 2004: Vol. 306. no. 5702, p. 1686
Sorry, that should have been:
In the GNU world the static C runtime library is libc.a and the dynamic (or shared) version is libc.so.
No, msvcrt.dll is the dynamically linked version of the C runtime library. If you used dynamic linking it has to be present at run time. If you use static linking it isn't used at compile time or at runtime.
Here is the deal: many languages, including C and Visual Basic, depend on a library of commonly used subroutines. Some of the subroutines you are probably aware of, say malloc or strncpy. Others work behind the scenes to set up the heap and such. You can link to this library statically, in which case the code for those subroutines is pulled into your executable. Or, you can link to the library dynamically, in which case only placeholders go into your executable. Then, when your program is run, it loads the dynamic version of the library into memory and fixes up the stubs. The advantage of this is that all dynamically linked program can share a single copy of the library in memory. The disadvantage is that your program depends on the user having a copy of the dynamic library.
All of the major C compilers provide both static and dynamic versions of the C runtime. In the GNU world the For some reason Microsoft only provides a dynamic version of the VB runtime. The VB runtime also contains the code for the p-code interpreter but this isn't called when native-code generation is turned on.
This was true for the older versions of VB 4 and earlier, but VB 5 can generate native code. Yes, you still need the run time library, just as most applications writen in C need a runtime library like libgc or msvcrt. The fly in the ointment is that Microsoft doesn't provide a static version of the VB runtime library.
Since you have a degree in education, perhaps you can offer me some insight into a problem that has long troubled me. It seems to me that educators are too prone to jumping on band wagons and fighting over them with a tenacity that puts "vi vs emacs" flamewars to shame. For example, I've seen school districts just about disintegrate in the debate over whole language vs phonetics. As a consumer of education it seemed to me that both schools had their points, and that children would benefit from being exposed to both methods. That seemed unacceptable to many educators though who seemed to demand that educators hew to one method or the other.
I was educated under the "new math" paradigm in the sixties, and I many others absolutely loved it. However, other kids didn't do at all well, and so "new math" was consigned to the rubbish heap. Kids obviously differ in their learning styles, why do we have to have a monolithic approach to instructional design?
I'm not clear whether this is a matter of helping out an aquaintance or of going the extra mile for a loyal customer. Either way, I think the answer is simply to tell them that prior commitments and obligations prevent you from helping them out. If it is a close friend, you can give them as much detail as they'll stand to listen to. If it is a business relationship, I'd avoid giving them Too Much Information, all they really want to know is whether you can help or not.
Only in the sense that riding in a airplane is aeronautical engineering. To be utterly pedantic, I suppose it is necessary to understand union, intersection, and cartesian product in writing queries, but surely you covered those in the first 10 minutes of your discrete math course! I can only observe that I've got an M.S. in math and at one point I was working with a very smart gentleman who learned SQL at a trade school. He didn't know discrete math from the indiscrete charm of the bourgoise, but when I wanted help with a tricky query he was the guy to talk to. If you could give me an example of a theorem from discrete math that actually helped you frame a query, I'd be delighted to hear about it.
An example would be helpful here too, because I have no idea what you are talking about. Do you understand about that graph theory is not about computer graphics? Roughly, a graph is a set, and a symetric relationship defined on that set. If you haven't run into it yet in discrete math class, you soon will. Trees are a subcategory of graphs, and a variety of data structures that model trees are teriffically useful in software. The developer of the GUI internals may very well have used graphs and trees to store the elements of the UI, but the typical workaday programmer adding a list box to a web page is no more using graph theory then I'm using fluid mechanics when I'm flying coach across the atlantic.