Even predicting the recent past can be tricky.
on
5595 Days and Counting
·
· Score: 2, Informative
If we have nanotubes in quantity tomorrow, I'll be the first to cheer, because you can do many interesting things with them on a smaller scale than building space elevators. However, I'm not holding my breath.
Start cheering then. We have the means to manufacture them in quantity now. (Basically, continious process vapor deposition, with a few tricks to it).
The present problem is the old "how do they stick the teflon to the pan" problem; getting them to play nicely in composites, and finding ways to manufacture the composites.
I once got sent to an HR department, with a note to give them when it came time for the behavior test.
The note said "Don't give this candidate the behavior test. We are satisfied with his performance and have seen enough of his behaviors to know that we don't want to provoke any of them."
(As an independent consultant I'd already written a number of "Let's quantify the stupidity of corprorate policy X in simple, blunt terms that everyone can understand" reports.)
I was once in a situation much like this. They had gotten as far as a beta test with "dummied up data"--real HR files that had had sensitive fields replaced by random goo--before I hit upon the idea of demonstrating how easy it was to crack. I e-mailed my boss & my boss's boss copies of their (fake data) salery history & confidential performance reviews, along with an explanation of how to get anyone's data out of the system.
There followed a flurry while the specific hole was patched, and then I repeated the stunt on the new, "secured" system.
By round three or four of this the idea of using http for the web interface for HR data died a quiet death.
It sounds like what they are doing is identifying the shared patterns (rather than "factoring them out"); the idea being that, if a gene codes for something important it won't change much (important gene + random change = dead critter), while the junk regions will change at a steady rate. If you look at two creatures that are medium-distant relatives, the sequences they share are likely to be important genes.
This is exactly oposite the usual connotation of "factoring out," which implies that they are removing the common regions and looking for the differences.
-- MarkusQ
P.S. And, with regard to ZIP, they aren't looking for repetitions within a genome, but rather similar regions in corresponding portions of different genomes.
Unless you construct the preamble and tail in-place
Agreed. The point here was pseudo-code. The original poster said that they found they got better results if they wrote functional pseudo code. This makes sense: if you specify the results you want first, and then work out the most efficent way to get them, you are more likely to succeed than if you start by writing something that is efficent and just hope that it does what you want.
BTW, in most cases you don't need to use garbage collection to implement these sorts of functions (although it may be one of the most obvious implementations). Here, for example, you could do everything on a stack, making cleanup easy. On some systems this may even be faster than doing it in place (think dirty cache/locality of reference issues). Or you could flop back and forth between two buffers.
The point being that these are all how-to-do-it questions which should wait until after what-to-do is known.
I find that I end up writing much cleaner code when I prototype in functional pseudocode.
Could you give a simple example of what you mean?
I suspect he's talking about something like this (loosly paraphrased from an intro to Haskell, IIRC):
Problem: write a sort using the quicksort algorithm.
Imperative pseudocode:
The quicksort procedure takes an array, and an upper and lower bounds.
If the lower bound isn't less than the upper bound, it does nothing.
Pick a pivot value from an arbitrary point in the array between the upper bound and the lower bound.
Start two pointers walking through the array, one from each end.
Move the bottom pointer up till you find an item greater than or equal to the pivot value.
Move the upper pointer down until you find a value less than the pivot value or until it meets the lower pointer.
If the pointers haven't met, swap the elements they point to and repeat from step 5.
quicksort the portion of the array between the lower bound and the item before the point where the pointers met.
quicksort the portion of the array between the item after where the pointers met and the upper bound.
Functional pseudocode:
Sorting a empty list give you an empty list.
Sorting a list (call it L) with at least one item in it (call it x) gives you a list which is the concatination of a preamble, N occurrences of x, and a tail, where:
the preamble is a sorted list of everything in L less than x
N is the number of times x occurs in L.
the tail is a sorted list of everything in L greater than x
In this context DFA refers to "Deterministic Finite Automata". Among other things, they can be employed in highly efficient lexical analyzers. If you've got a decent book on compilers around, you can look up the details.
In the context of formal languages, sure. (Note that that was the first interpretation on my list.) But in the context of Taylor series? Remember, the original poster was claiming a link between formal language theory and Calculus via a DFA <---> Taylor series mapping. How in the heck do you say that Deterministic Finite Automata are somehow the same as Taylor series? Structurally, I just can't see it. But I also can't see/find any other interpretation of DFA that makes this make sense. Thus my request that the original poster provide the offered details.
For example, DFA construction from a regular expression turns out to be a Taylor series expansion of the expression. (If anyone is curious, I can supply the details.)
Ok, I'll bite. This statement doesn't make any sense to me, and a google search turned up matches (with DFA in {"Deterministic Finite Automata", "Detrended Fluctuation Analysis", "Dynamic financial analysis", "Descrete functional analysis"...}), none of which seemed to be what you were refering to.
These two are actually the same book.
Are you certain? I only own POET, but I notice (from the provided Amazon links) that they have different ISBN numbers (0465067107 vs. 0465067093). Although they are both 272 pages long. Repackaging?
The No Electronic Theft law and the supposed "Internet Privacy Act" are two separate laws. Moreover, one was referenced in an article submission quoted from a reputable (subjective, I know) news source, and the other was an off-hand comment by one of the half-million or so Slashdot subscribers.
You are quite correct. My mistake.
-- MarkusQ
P.S. I'd blame it on an erratic sleep schedule, but I suspect that it isn't too far from the norm arround here. So I'll have to just plead careless reading on my part.
Man, I wish I had some of last week's mod points left. If you are correct (and it appears you are), this whole article/thread is a waste of time. It's almost a case of Slashdot trolling itself!
He wasn't joking. The code he posted was the very code the orginal poster was raving about. It's not like he pulled some obscure C out of left field to make his point; he simply posted the code 434 byte CSS descrambler that ebyrob brought up as an example of the power of C.
As soon as you try, the fiber won't be dark anymore, invalidating one of the conditions. You can have dark fiber, or fiber with TCP/IP (or just pretty lights for that matter) but not both at the same time.
Out of curiosity, why? I've heard a lot of off the cuff/gut feeling/emotional arguments but I've yet to find anyone who can offer a substantive reason for doubting that eutactic nanotech is feasible.
I do have a problem when they use copyright, a government enforced monopoly on the software, to extend that monopoly into another market, i.e. the ink cartridge replacement market.
I think this is the most succinct and lucid statement of the objection to Lexmark's conduct that I have seen. I'd mod you up if I had the points, but I don't so I'll just try to draw the attention of someone who does.
IIRC, there was an example in the Arabian Nights. Of course, this was long before computers but it still worked like magic.
-- MarkusQ
P.S. On a more serious note, I wouldn't be all that shocked to learn that 'undo' wasn't implemented on a computer before the 1970s. My main reason for thinking that is that it is relatively expensive, and really only useful in interactive environments. When you're editing your program by manually shuffling h-cards, there isn't much sense in having "undo".
MarkusQ:
They're just a shorthand for a sub-ring of the matricies
exp(pi*sqrt(163)) (aka Mr. 262537412640768744-epsilon):
Now that's bullshit if I ever read it. Any algebra that has a faithful matrix representation can be considered a sub-ring of a matrix algebra.
Agreed. But that isn't anything new. I wasn't disputing their claims, just their hype. What is it that's so wonderful here?
That, for any integer N>=1 you can find N orthagonal matricies to form the basis of a sub-ring?
That the sub-ring so formed will have the properties you'd expect of a ring?
That, knowing (choosing) the form of your basis, you can find computational shortcuts?
I guess it feels to me as if Bob Shmoe anounced the discovery of a new group of order 163, which he dubed "Shmo", and looking at it I saw it was {f^163=I}. I'd say "hey, that's just C-163. That wouldn't mean I was dismissing all modulo arithmetic.
-- MarkusQ
P.S. If you think I'm missing something here, I'd be glad to know what. I think I read through more of their web pages than most posters, but I did it with a 13-month old on my lap, so I may well have bounced past some key point.
consisting of Microsoft Netmeeting (which almost everyone has on their machines),
Who told you that?
I would say that most people don't have net meeting. But even if they did, is it realy what you want? I'd think even just e-mail, or some sort of wiki would be much better.
IIRC, that theme occurs several places, sometimes with stutters or other ornaments; I don't recall off hand what part of the score I copied it from though.
Number systems with more than 2 coordinates are treated like matrices, because no one is able to find rules of muliplication and addition for say a 5-coordinate system that makes the numbers associative, commutative and distributive.
I read everything in the left two columns of links. And there wasn't more than vague promises of N+ solving signal processing problems and basic descriptions of real number field theory. But nothing saying how their N+ numbers are associative, commutative and distributive. So this is just bullshit.
I had the same reaction, but after digging a little deeper on their site (hurah Google) I did turn up a explanation, and guess what? They're just a shorthand for a sub-ring of the matricies.
The formal question you pose is reasonable. However, I must take issue with your first point. {...don't reinvent the wheel...] It's always worthwhile to use the right tool for the right job.
The main point you make is quite reasonable; however, I think you have missed mine. Before you can determine "the right tool for the job" you must first know what the job is. From what he's described, I'm not sure that either package is the right tool.
Why?
Imagine the question was:
I deliver newspapers in a rural community. I need to keep good maps of my route in case I get sick and my brother need to do it for me. I also need to keep track of how much people owe me. Which package would you recommend...
I'm a choreographer and I'm working on tracking our troup's performances, as well as planning/developing new dances. Which package would be best...
And so forth...
The point is, unless we know what he's actually trying to do, it's hard to answer his question. Is he collecting statistics from trial groups? Is he attempting to model bonding site affinity/sensitivity/selectivity or something from 3d molecular models? Is he trying to reverse engineer some chemestry from blood samples? Is he trying to count the spots on a hundred and one petri dishes? Is he trying to judge the statistical significance of a micro-array test? What, in short, is the problem?
Realistically, why would you need either? You can't be getting your data out to more than a few digits; even a single/short float would be overkill. So, for that matter, is machine learning...
I guess I'll rephrase this as a question: just what are you hoping that the math package will do for you? I'm not asking for buzzwords, just a plain-talk statement of the problem you are trying to solve.
Slashdot Mirror
The present problem is the old "how do they stick the teflon to the pan" problem; getting them to play nicely in composites, and finding ways to manufacture the composites.
-- MarkusQ
I once got sent to an HR department, with a note to give them when it came time for the behavior test.
The note said "Don't give this candidate the behavior test. We are satisfied with his performance and have seen enough of his behaviors to know that we don't want to provoke any of them."
(As an independent consultant I'd already written a number of "Let's quantify the stupidity of corprorate policy X in simple, blunt terms that everyone can understand" reports.)
-- MarkusQ
P.S. I got the job.
I was once in a situation much like this. They had gotten as far as a beta test with "dummied up data"--real HR files that had had sensitive fields replaced by random goo--before I hit upon the idea of demonstrating how easy it was to crack. I e-mailed my boss & my boss's boss copies of their (fake data) salery history & confidential performance reviews, along with an explanation of how to get anyone's data out of the system.
There followed a flurry while the specific hole was patched, and then I repeated the stunt on the new, "secured" system.
By round three or four of this the idea of using http for the web interface for HR data died a quiet death.
-- MarkusQ
It sounds like what they are doing is identifying the shared patterns (rather than "factoring them out"); the idea being that, if a gene codes for something important it won't change much (important gene + random change = dead critter), while the junk regions will change at a steady rate. If you look at two creatures that are medium-distant relatives, the sequences they share are likely to be important genes.
This is exactly oposite the usual connotation of "factoring out," which implies that they are removing the common regions and looking for the differences.
-- MarkusQ
P.S. And, with regard to ZIP, they aren't looking for repetitions within a genome, but rather similar regions in corresponding portions of different genomes.
Unless you construct the preamble and tail in-place
Agreed. The point here was pseudo-code. The original poster said that they found they got better results if they wrote functional pseudo code. This makes sense: if you specify the results you want first, and then work out the most efficent way to get them, you are more likely to succeed than if you start by writing something that is efficent and just hope that it does what you want.
BTW, in most cases you don't need to use garbage collection to implement these sorts of functions (although it may be one of the most obvious implementations). Here, for example, you could do everything on a stack, making cleanup easy. On some systems this may even be faster than doing it in place (think dirty cache/locality of reference issues). Or you could flop back and forth between two buffers.
The point being that these are all how-to-do-it questions which should wait until after what-to-do is known.
-- MarkusQ
There were rumours that they were being jailed in someplace called "Chroot" which I think is in Lebanon. I never heard if there was confirmation.
-- MarkusQ
I find that I end up writing much cleaner code when I prototype in functional pseudocode.
Could you give a simple example of what you mean?
I suspect he's talking about something like this (loosly paraphrased from an intro to Haskell, IIRC):
-- MarkusQIn this context DFA refers to "Deterministic Finite Automata". Among other things, they can be employed in highly efficient lexical analyzers. If you've got a decent book on compilers around, you can look up the details.
In the context of formal languages, sure. (Note that that was the first interpretation on my list.) But in the context of Taylor series? Remember, the original poster was claiming a link between formal language theory and Calculus via a DFA <---> Taylor series mapping. How in the heck do you say that Deterministic Finite Automata are somehow the same as Taylor series? Structurally, I just can't see it. But I also can't see/find any other interpretation of DFA that makes this make sense. Thus my request that the original poster provide the offered details.
-- MarkusQ
For example, DFA construction from a regular expression turns out to be a Taylor series expansion of the expression. (If anyone is curious, I can supply the details.)
Ok, I'll bite. This statement doesn't make any sense to me, and a google search turned up matches (with DFA in {"Deterministic Finite Automata", "Detrended Fluctuation Analysis", "Dynamic financial analysis", "Descrete functional analysis"...}), none of which seemed to be what you were refering to.
-- MarkusQ
These two are actually the same book. Are you certain? I only own POET, but I notice (from the provided Amazon links) that they have different ISBN numbers (0465067107 vs. 0465067093). Although they are both 272 pages long. Repackaging?
-- MarkusQ
-- MarkusQ
P.S. I'd blame it on an erratic sleep schedule, but I suspect that it isn't too far from the norm arround here. So I'll have to just plead careless reading on my part.
Man, I wish I had some of last week's mod points left. If you are correct (and it appears you are), this whole article/thread is a waste of time. It's almost a case of Slashdot trolling itself!
-- MarkusQ
-- MarkusQ
I know you're joking but it's missing the point.
He wasn't joking. The code he posted was the very code the orginal poster was raving about. It's not like he pulled some obscure C out of left field to make his point; he simply posted the code 434 byte CSS descrambler that ebyrob brought up as an example of the power of C.
-- MarkusQ
Goal: TCP/IP over dark fiber.
So far as I can see, you can't do it.
As soon as you try, the fiber won't be dark anymore, invalidating one of the conditions. You can have dark fiber, or fiber with TCP/IP (or just pretty lights for that matter) but not both at the same time.
-- MarkusQ
Out of curiosity, why? I've heard a lot of off the cuff/gut feeling/emotional arguments but I've yet to find anyone who can offer a substantive reason for doubting that eutactic nanotech is feasible.
-- MarkusQ
I do have a problem when they use copyright, a government enforced monopoly on the software, to extend that monopoly into another market, i.e. the ink cartridge replacement market.
I think this is the most succinct and lucid statement of the objection to Lexmark's conduct that I have seen. I'd mod you up if I had the points, but I don't so I'll just try to draw the attention of someone who does.
-- MarkusQ
IIRC, there was an example in the Arabian Nights. Of course, this was long before computers but it still worked like magic.
-- MarkusQ
P.S. On a more serious note, I wouldn't be all that shocked to learn that 'undo' wasn't implemented on a computer before the 1970s. My main reason for thinking that is that it is relatively expensive, and really only useful in interactive environments. When you're editing your program by manually shuffling h-cards, there isn't much sense in having "undo".
In case the troll moderation has confused the original poster (on the assumption that you're actually reading the thread you started):
I am not a troll. I mean my question seriously. What are you trying to do?
Without knowing that, it's hard to answer your question.
-- MarkusQ
MarkusQ: They're just a shorthand for a sub-ring of the matricies
exp(pi*sqrt(163)) (aka Mr. 262537412640768744-epsilon): Now that's bullshit if I ever read it. Any algebra that has a faithful matrix representation can be considered a sub-ring of a matrix algebra.
Agreed. But that isn't anything new. I wasn't disputing their claims, just their hype. What is it that's so wonderful here?
- That, for any integer N>=1 you can find N orthagonal matricies to form the basis of a sub-ring?
- That the sub-ring so formed will have the properties you'd expect of a ring?
- That, knowing (choosing) the form of your basis, you can find computational shortcuts?
I guess it feels to me as if Bob Shmoe anounced the discovery of a new group of order 163, which he dubed "Shmo", and looking at it I saw it was {f^163=I}. I'd say "hey, that's just C-163. That wouldn't mean I was dismissing all modulo arithmetic.-- MarkusQ
P.S. If you think I'm missing something here, I'd be glad to know what. I think I read through more of their web pages than most posters, but I did it with a 13-month old on my lap, so I may well have bounced past some key point.
consisting of Microsoft Netmeeting (which almost everyone has on their machines),
Who told you that?
I would say that most people don't have net meeting. But even if they did, is it realy what you want? I'd think even just e-mail, or some sort of wiki would be much better.
-- MarkusQ
-- MarkusQ
Number systems with more than 2 coordinates are treated like matrices, because no one is able to find rules of muliplication and addition for say a 5-coordinate system that makes the numbers associative, commutative and distributive.
I read everything in the left two columns of links. And there wasn't more than vague promises of N+ solving signal processing problems and basic descriptions of real number field theory. But nothing saying how their N+ numbers are associative, commutative and distributive. So this is just bullshit.
I had the same reaction, but after digging a little deeper on their site (hurah Google) I did turn up a explanation, and guess what? They're just a shorthand for a sub-ring of the matricies.
-- MarkusQ
The formal question you pose is reasonable. However, I must take issue with your first point. {...don't reinvent the wheel...] It's always worthwhile to use the right tool for the right job.
The main point you make is quite reasonable; however, I think you have missed mine. Before you can determine "the right tool for the job" you must first know what the job is. From what he's described, I'm not sure that either package is the right tool.
Why?
Imagine the question was:
The point is, unless we know what he's actually trying to do, it's hard to answer his question. Is he collecting statistics from trial groups? Is he attempting to model bonding site affinity/sensitivity/selectivity or something from 3d molecular models? Is he trying to reverse engineer some chemestry from blood samples? Is he trying to count the spots on a hundred and one petri dishes? Is he trying to judge the statistical significance of a micro-array test? What, in short, is the problem?-- MarkusQ
Realistically, why would you need either? You can't be getting your data out to more than a few digits; even a single/short float would be overkill. So, for that matter, is machine learning...
I guess I'll rephrase this as a question: just what are you hoping that the math package will do for you? I'm not asking for buzzwords, just a plain-talk statement of the problem you are trying to solve.
-- MarkusQ