But it makes me sick to think where cell phones could be if only they'd focus on the phone instead of trying to make it a GameBoy, too.
How much evolution has there been in ordinary phones since Bell? Not that much.
There just is not much room for improving a device designed just to transfer sound. I think the phone-aspect of a mobile phone has been pretty stable for ten years now.
Stand On Zanzibar is a treasure-trove for future technology. While some parts of it are somewhat outdated (eg. there are only few computers), the rest describes inventions that I have not met elsewhere in literature. And it's not limited to technology either; the environment of "Zanzibar" is easily one of the most intriguing I've ever met in a sci-fi novel.
In 1968 Brunner presented Viagra (called 'stiffeners'), and in 1975 he foresaw the Internet (The Shockwave Rider). He seems like a sure bet for the reading list with Stand on Zanzibar, perhaps with The Shockwave Rider too.
All three points are more or less incorrect. Quantum computing does not necessarily improve every aspect of classical computing. The main difference is that classical computers are dead-on deterministic, while quantum computers are more of a probabilistic sort.
Claim #1: No encryption.
There is a quantum algorithm for factoring large numbers (Shor's algorithm). It will break RSA... but so what? There are elliptic curves and countless other methods of cryptography still available. Quantum computing might or might not be able to break these.
And then there is quantum encryption,...
Claim #2: Improved compression
The concept of compression is a complex issue of information/communicaton theory. It applies to information, not to computing. Quantum computing is basically just computing; you might find be able to compute the compressed file faster, but no computational method can squeeze information beyond the theoretic limit.
The pi scheme mentioned here is totally unusable: it takes as much bits to represent the index as it takes to represent the data itself.
Claim #3: Faster computing
This too is slightly incorrect. There are some things that are faster to implement with quantum computing. Some things, like adding two to two are suitable for classical computers. Even the Shor's algorithm uses classical FFT at one step. I don't think it's certain that a quantum computer can solve NP-complete problems faster than classical computers.
Besides, for some things *the* optimal solution simply can't be expressed in any computer, quantum or classical. (Think for example the equation x^2 = 2. The answer can't be represented numerically in a computer; however, it can be approximated to as many digits as is necessary.)
Quantum computing does not necessarily imply massive parallel computing. For some applications something like this happens, but it's not the same as having n different computations running simultaniously; more like the the same computation running over with some variation.
Last of all, we don't get full certainty. The Shor's algorithm (IIRC) can give us an answer with very high probability, but that probability isn't 100%. (99.999% perhaps)
Re:Badass compression algorithm?
on
Share The Pi!
·
· Score: 1
Mathematics will have its say in the matter. Using some probability we find that the length of tge expected value of the index is the the same as the the length of the value of data itself.
I'll use base 10 in the following, but it works in other bases too; simply use the base you want instead of 10, and remember to also to change the base of the logarithm.
Let's assume that pi is normal and we want to find number N in pi, be it warez or metallica mp3z or even something useful. Now look at some index i; there's one in ten possibility that the digit of pi at that index equals the first digit of your number.
The next digits have to be equal too, and that also happens with the probability of 1/10. This goes on until you've checked all the digits in N.
To get the probability that the index i indeed is the start of a sequence you need to multiply all those individual probabilities, and you'll get probability p = (1/10)^L, that is, one tenth to the power of L, where L is the length of N in digits. (It so happens that L is roughly lg N, lg being base 10 logarithm)
Now the probability p is constant for all indices of pi; what you basically have is a Bernoulli test for all indices. (Imagine flipping a coin at each index with such coin that it will land on one side with probability p and on the other side rest of the time.)
Now, in k trials (ie. you test k indices) the expected value of successes is k*p; we want only one success, so that allows us to calculate the number of trials we most likely need.
k = 1/p = 10^L. Because L = lg N (roughly), that simplifies to k = 10^lg N = N. That is, you most likely need to test N indices to find a data N; to represent that information you'll have to use the same number of bits. Sucks, eh?
The moral of the story: do not fight mathematics, because mathematics always wins.
Pi will not, unfortunately, give you an arithmetical method of producing random digits.
If you pick digits from pi's decimal expansion with some deterministic method, say, every third digit, the sequence will be the same each and every time you run it. What you do get from pi are non-repeating pseudorandom numbers: you can eg. pick every nth digit where n is your seed (cf. usual (pseudo)random number generators)
To get truly random numbers from pi, you need pick the digits randomly... for which you of course need a random number generator...
A recent genetic study has showed that the people living now in Finland are related to other Europeans, including the Swedish. The language, however, is Fenno-Ugric, which lead many people to think that Finns are from somewhere Russia.
The original inhabitants were, IIRC, pushed away by the movement from the South. I assume that the new inhabitants took their language as their own, thus creating this genetic-linguistic-discrepancy.
I remember my history teacher telling me about the mention Tacitus has in his Annals; it probably referred to the aboriginals.
What I think would be an interesting bonus would be the possibility to subscribe to some appointment servers, and request new 'appointments'. I'm not just talking about meetings:
* Want to tape all the episodes of B5, but women's ice-hockey games postpone the episodes? Subscribe to the appropriate server, and get the timetable, with any possible changes in the timetable.
* Students could get times for their lectures and tests etc. directly to their own calendar.
* etc.
(I'm not thinking about a corporate calendar, but rather the Emacs (or Gnus) of all calendar-software; one you could use at home, at work, in your palm, offline, online,... This might explain the lack of business oriented examples:)
Another feature I would find useful is profiles; cf. Nokia phones or Debian's/etc/inferfaces. I'd like to be able to have different views according to which mode I'm in: work, home, holiday, etc.
It would need to have some sort of a filtering capability, as well as some prioritizing options for appointments. Example:
At work, I'd see all the work-related appointments, and all the others with priority 1. At home, I'd see all appointments relating to freetime, all public appointments on my gf's calendar, all 1 priority TV shows.
There's a whole lot more ideas, but none of them is IMHO as imporant as the ones listed above.
Unfortunetely, you gain a lot more brownie points in scietific cicles for publishing anything on dead tree than publishing it online. Even if neither version are peer-reviewed, the dead-tree version counts for more. It doesn't matter that the online version is of higher quality or have more readers.
It is also more accepted to quote from dead-tree sources than from online sources, which is a further incitament for authors to publish on dead tree.
About quoting; Assume that you are reading a paper which then references eg. Mathworld. At the moment, you're crapped. There's no way anyone can verify your claims. Paper has not only stability, but there is also very large framework for locating the information.
You need to do more than just put your findings on a.tex-file on your homepage. That's almost equivalent to having the only hardcopy on your own bookshelf. There are on-line mathematical archives, though, which provide electronic math articles, and I'd love to see that become the standard (in addition to journals and conferences, naturally.)
Online texts are useful in some situations, but extremely annoying in others. Personally I prefer my math as a book when the issue is more than one or two pages. Printed.pdfs and.pses just won't cut it. And the prestige is something; even I, a 21-year old CS student (ie. an information highway roadwarrior), would be excited about getting my stuff in a Book. Any old fool can throw stuff to ~/public_html/...
Despite my luddite feelings, I think Mathworld was one of the most best sites on the whole web. I had thought of sending some entries there, since I ran into "Become a Contributor" -buttons once too often:) Now... we'll see. I hope and believe the site will come back on-line, but until then I'm doing math the hard way.
Oh my. This crashes now and then, but oh my oh my. It's so fast. It's so pretty. It's so configurable.
The Multiple Document Interface isn't my bag, and the bookmark UI was a bit furry at start.
But I'm buying this senorita. This is IMHO the best graphical web browser on Linux, although IE 5(.5) still has the overall lead.
Some of the more frequent crashes occur when I try to choose a certain option in preferences. Annoying, but I can live without it. Some pages are lethal to the beta, but I'm hoping all the crashes are ironed out in the final version.
I'm not afraid of the radiation that goes through my head. It's the part that gets absorbed that gives me the willies.
What stops you from getting a hands-free set? Even if the radiation would be harmless at close range, I don't think anyone will get brain cancer from a phone kept in your pocket. (There are other organs which might get bit by the photons though...)
I'm really afraid of Bluetooth devices. A light MP3 or MD player with wireless BT-earphones would be an amazing combination; it's the stuff dreams are made of, something I could consider Adopting Early.
But at least for now, having a working brain outweighs all trendiness factors...
I wouldn't say the quality of mp3's recorded on an MD has better sound quality. MD uses ATRAC compression, which isn't quite as sophisticated as MP3. The newest versions of ATRAC are competetive with mp3 in terms of quality, though, and I don't know about the quality of portable MP3 players.
But recording mp3's on a MD isn't the ideal. First you compress the original with MP3, and then rape it a bit more with ATRAC. If your CD player has optical digital output, it's the simplest and most quality preserving way.
IIRC, Win2000 release wasn't announced here either. I don't think anyone complained, since MS knows marketing in and out. You couldn't miss it.
We've heard about NASDAQ dropping like a dead bird even here in Finland. I can't imagine anyone in the States missing the news. And if so happens, how about the next-door-software-engineer who vanished without a trace after buying a gun, or seen people in IRC moaning after their millions went down the drain.
You couldn't have missed this. This isn't news anymore.
Regardless of where the industry goes, the physical booklet with lyrics will eventually make way for artist's web sites, which can hold infinitely more information than the little booklet.
I just can't see this happening. There are bands who publish most (or more) of the information available on the booklet also on the Net. However, the only time I use those resources is when I do not own the CD yet.
Web-only lyrics (and other resources) have downsides. First of all, although I listen to music while using my computer, those two aren't so connected. In fact, I try to avoid all distractions (incl. my computing machine) when listening to my new CD's for the first time time.
Somehow it's clumsier too. Having the small booklet on your lap while you're concentrating on the music feels much nicer than staring at the bulky screen... call me a neo-luddite, but I like paper.
Slashdot Mirror
How much evolution has there been in ordinary phones since Bell? Not that much.
There just is not much room for improving a device designed just to transfer sound. I think the phone-aspect of a mobile phone has been pretty stable for ten years now.
While this might not be what you mean, check out http://hardware.mp3.com/hardware/individual/3859.h tml
... it's just a tad big :)
There's LCD, large HD,
Stand On Zanzibar is a treasure-trove for future technology. While some parts of it are somewhat outdated (eg. there are only few computers), the rest describes inventions that I have not met elsewhere in literature. And it's not limited to technology either; the environment of "Zanzibar" is easily one of the most intriguing I've ever met in a sci-fi novel.
In 1968 Brunner presented Viagra (called 'stiffeners'), and in 1975 he foresaw the Internet (The Shockwave Rider). He seems like a sure bet for the reading list with Stand on Zanzibar, perhaps with The Shockwave Rider too.
All three points are more or less incorrect. Quantum computing does not necessarily improve every aspect of classical computing. The main difference is that classical computers are dead-on deterministic, while quantum computers are more of a probabilistic sort.
...
Claim #1: No encryption.
There is a quantum algorithm for factoring large numbers (Shor's algorithm). It will break RSA... but so what? There are elliptic curves and countless other methods of cryptography still available. Quantum computing might or might not be able to break these.
And then there is quantum encryption,
Claim #2: Improved compression
The concept of compression is a complex issue of information/communicaton theory. It applies to information, not to computing. Quantum computing is basically just computing; you might find be able to compute the compressed file faster, but no computational method can squeeze information beyond the theoretic limit.
The pi scheme mentioned here is totally unusable: it takes as much bits to represent the index as it takes to represent the data itself.
Claim #3: Faster computing
This too is slightly incorrect. There are some things that are faster to implement with quantum computing. Some things, like adding two to two are suitable for classical computers. Even the Shor's algorithm uses classical FFT at one step. I don't think it's certain that a quantum computer can solve NP-complete problems faster than classical computers.
Besides, for some things *the* optimal solution simply can't be expressed in any computer, quantum or classical. (Think for example the equation x^2 = 2. The answer can't be represented numerically in a computer; however, it can be approximated to as many digits as is necessary.)
Quantum computing does not necessarily imply massive parallel computing. For some applications something like this happens, but it's not the same as having n different computations running simultaniously; more like the the same computation running over with some variation.
Last of all, we don't get full certainty. The Shor's algorithm (IIRC) can give us an answer with very high probability, but that probability isn't 100%. (99.999% perhaps)
Mathematics will have its say in the matter. Using some probability we find that the length of tge expected value of the index is the the same as the the length of the value of data itself.
I'll use base 10 in the following, but it works in other bases too; simply use the base you want instead of 10, and remember to also to change the base of the logarithm.
Let's assume that pi is normal and we want to find number N in pi, be it warez or metallica mp3z or even something useful. Now look at some index i; there's one in ten possibility that the digit of pi at that index equals the first digit of your number.
The next digits have to be equal too, and that also happens with the probability of 1/10. This goes on until you've checked all the digits in N.
To get the probability that the index i indeed is the start of a sequence you need to multiply all those individual probabilities, and you'll get probability p = (1/10)^L, that is, one tenth to the power of L, where L is the length of N in digits. (It so happens that L is roughly lg N, lg being base 10 logarithm)
Now the probability p is constant for all indices of pi; what you basically have is a Bernoulli test for all indices. (Imagine flipping a coin at each index with such coin that it will land on one side with probability p and on the other side rest of the time.)
Now, in k trials (ie. you test k indices) the expected value of successes is k*p; we want only one success, so that allows us to calculate the number of trials we most likely need.
k = 1/p = 10^L. Because L = lg N (roughly), that simplifies to k = 10^lg N = N. That is, you most likely need to test N indices to find a data N; to represent that information you'll have to use the same number of bits. Sucks, eh?
The moral of the story: do not fight mathematics, because mathematics always wins.
Pi will not, unfortunately, give you an arithmetical method of producing random digits.
If you pick digits from pi's decimal expansion with some deterministic method, say, every third digit, the sequence will be the same each and every time you run it. What you do get from pi are non-repeating pseudorandom numbers: you can eg. pick every nth digit where n is your seed (cf. usual (pseudo)random number generators)
To get truly random numbers from pi, you need pick the digits randomly... for which you of course need a random number generator...
The original inhabitants were, IIRC, pushed away by the movement from the South. I assume that the new inhabitants took their language as their own, thus creating this genetic-linguistic-discrepancy.
I remember my history teacher telling me about the mention Tacitus has in his Annals; it probably referred to the aboriginals.
What I think would be an interesting bonus would be the possibility to subscribe to some appointment servers, and request new 'appointments'. I'm not just talking about meetings:
... This might explain the lack of business oriented examples :)
/etc/inferfaces. I'd like to be able to have different views according to which mode I'm in: work, home, holiday, etc.
* Want to tape all the episodes of B5, but women's ice-hockey games postpone the episodes? Subscribe to the appropriate server, and get the timetable, with any possible changes in the timetable.
* Students could get times for their lectures and tests etc. directly to their own calendar.
* etc.
(I'm not thinking about a corporate calendar, but rather the Emacs (or Gnus) of all calendar-software; one you could use at home, at work, in your palm, offline, online,
Another feature I would find useful is profiles; cf. Nokia phones or Debian's
It would need to have some sort of a filtering capability, as well as some prioritizing options for appointments. Example:
At work, I'd see all the work-related appointments, and all the others with priority 1. At home, I'd see all appointments relating to freetime, all public appointments on my gf's calendar, all 1 priority TV shows.
There's a whole lot more ideas, but none of them is IMHO as imporant as the ones listed above.
Unfortunetely, you gain a lot more brownie points in scietific cicles for publishing anything on dead tree than publishing it online. Even if neither version are peer-reviewed, the dead-tree version counts for more. It doesn't matter that the online version is of higher quality or have more readers.
It is also more accepted to quote from dead-tree sources than from online sources, which is a further incitament for authors to publish on dead tree.
About quoting; Assume that you are reading a paper which then references eg. Mathworld. At the moment, you're crapped. There's no way anyone can verify your claims. Paper has not only stability, but there is also very large framework for locating the information.
You need to do more than just put your findings on a .tex-file on your homepage. That's almost equivalent to having the only hardcopy on your own bookshelf. There are on-line mathematical archives, though, which provide electronic math articles, and I'd love to see that become the standard (in addition to journals and conferences, naturally.)
Online texts are useful in some situations, but extremely annoying in others. Personally I prefer my math as a book when the issue is more than one or two pages. Printed .pdfs and .pses just won't cut it. And the prestige is something; even I, a 21-year old CS student (ie. an information highway roadwarrior), would be excited about getting my stuff in a Book. Any old fool can throw stuff to ~/public_html/...
Despite my luddite feelings, I think Mathworld was one of the most best sites on the whole web. I had thought of sending some entries there, since I ran into "Become a Contributor" -buttons once too often :) Now... we'll see. I hope and believe the site will come back on-line, but until then I'm doing math the hard way.
Oh my. This crashes now and then, but oh my oh my. It's so fast. It's so pretty. It's so configurable.
The Multiple Document Interface isn't my bag, and the bookmark UI was a bit furry at start.
But I'm buying this senorita. This is IMHO the best graphical web browser on Linux, although IE 5(.5) still has the overall lead.
Some of the more frequent crashes occur when I try to choose a certain option in preferences. Annoying, but I can live without it. Some pages are lethal to the beta, but I'm hoping all the crashes are ironed out in the final version.
I'm not afraid of the radiation that goes through my head. It's the part that gets absorbed that gives me the willies.
What stops you from getting a hands-free set? Even if the radiation would be harmless at close range, I don't think anyone will get brain cancer from a phone kept in your pocket. (There are other organs which might get bit by the photons though...)
I'm really afraid of Bluetooth devices. A light MP3 or MD player with wireless BT-earphones would be an amazing combination; it's the stuff dreams are made of, something I could consider Adopting Early.
But at least for now, having a working brain outweighs all trendiness factors...
Now I hate you, Jon.
For most of us non-US -people, the 4th of July doesn't mean a thing. It's something we like to watch from distance and chuckle.
But this is too much. Saying a French upholds American values, and is therefore a hero of the day almost literally makes me sick.
The final words, "Happy independence day, Jose" were the last straw. France has it's own independence day, for Pete's sake!
I'm not even French, but the über-American attitude gives me the creeps.
I wouldn't say the quality of mp3's recorded on an MD has better sound quality. MD uses ATRAC compression, which isn't quite as sophisticated as MP3. The newest versions of ATRAC are competetive with mp3 in terms of quality, though, and I don't know about the quality of portable MP3 players.
But recording mp3's on a MD isn't the ideal. First you compress the original with MP3, and then rape it a bit more with ATRAC. If your CD player has optical digital output, it's the simplest and most quality preserving way.
IIRC, Win2000 release wasn't announced here either. I don't think anyone complained, since MS knows marketing in and out. You couldn't miss it.
We've heard about NASDAQ dropping like a dead bird even here in Finland. I can't imagine anyone in the States missing the news. And if so happens, how about the next-door-software-engineer who vanished without a trace after buying a gun, or seen people in IRC moaning after their millions went down the drain.
You couldn't have missed this. This isn't news anymore.
Regardless of where the industry goes, the physical booklet with lyrics will eventually make way for artist's web sites, which can hold infinitely more information than the little booklet.
I just can't see this happening. There are bands who publish most (or more) of the information available on the booklet also on the Net. However, the only time I use those resources is when I do not own the CD yet.
Web-only lyrics (and other resources) have downsides. First of all, although I listen to music while using my computer, those two aren't so connected. In fact, I try to avoid all distractions (incl. my computing machine) when listening to my new CD's for the first time time.
Somehow it's clumsier too. Having the small booklet on your lap while you're concentrating on the music feels much nicer than staring at the bulky screen... call me a neo-luddite, but I like paper.