In the case of this exploit, a deep flaw in IDN and computer fonts means that character #1072 is rendered typographically as an "a". The irony is that this is one of the few cases in which a computer can readily tell the difference between "a" and #1072 and a person cannot. The only solution would be rules that prohibit isomorphic characters in typefaces or a in-browser warning system that analyses the potential for ambiguity and alerts the user.
The issue is that the strenght of the signal is not a linear function of distance.
I agree with you, but don't think it hinders the system as much as it might seem. The key is that most base stations have limited range -- if I can detect the signal at some normal strength level (e.g., not using a cantenna), I know I'm within a 50-100 meters of the transmitter.
What got me thinking about this was my experience on a recent business trip. Sitting in my room at a bed-n-breakfast, I was picking up 4 wireless networks - the network of the B&B, the one from an adjacent B&B, and two other networks in the residential neighborhood. I doubt there are many places in which all four networks are detectable. Conversely, that means that if I can detect all four networks, I must be in a particular fairly small error circle (say within 1/4 to 1/2 a city block) within the overlap of the four networks.
A simple WiFi navigation system would use binary detect/no-detect data to determine location. If one can detect stations A, B, and R, then one must be in a particular street, near a particular corner.
Yes, some transmitters might be detectable over a much greater distances. These signals are much less useful for navigation because detecting them does not mean one is in any particular location. But with enough data, the system can easily learn to ignore signals from longer-range WiFi transmitters.
Yes, some transmitters might be temporarily undetectable due to being off or in a magic voodoo WiFi dead-zone. But the statistical logic of the navigation system could be tied to what is detectable and not use any evidence about what isn't detectable.
A more sophisticated system could use signal strength, but it would need a much greater volume of data in order to compute statistically meaningful location estimates. Although the signal drop-off is not isotropic, time-invariant, it is not statistically arbitrary. One is unlikely to get a strong signal at 150 meters from a normal WiFi station. Statistical variations in signal strength with distance would determine the error region around any estimate of location.
As I said, it will never beat GPS. But it would be cool to calculate where you are from who you can hear.
Rather that use dead-trees, wardrivers could use the names and signal strengths to nearby stations to estimate their position and calculate movements toward the best accessible WiFi hotspots. Even if the base station is password protected, its broadcast name and existence can help war drivers figure out where they are on the war-driving map. And if each wardriver submits data on stations and signal strengths, statistical combination of all the data could help refine the map further.
It will never beat GPS, but it would be cool to create a city-wide navigation system that works on WiFi -- "just turn left when you get max signal from MAC 00 60 1D 1C B9 0D."
The recent/, story on media longevity highlights the growing problem of decomposing data-layers on current generation optical disk technology. This new disk, with its even higher density, would seem to be even more likely to suffer from longevity problems.
Perhaps the xxAA has nothing to worry about -- media buying customers will lose access to copied data through dye-decomposition sooner than through expiring DRM licenses.
I wonder when CAPTCHAs will be so hard that an increasing fraction of the human population fails them. Perhaps the true origin of SkyNet will be when some spammer's AI realizes that humans are superfluous in an age of totally automated click-throughs and e-commerce.
Thank you for the reply. I'm glad your work generalizes to longer search-term lists. Like so many other/. readers, I did not take the time to read your preprint before posting.
I've often wondered if one can use simple pair-wise distance estimates to reconstruct a polytope or distorted simplex for the set of items within a multidimensional space. In theory, an N-object system, with non-zero pairwise distances, requires (N-1) dimensions. But in practice, many real systems don't fill the space -- being M-dimensional (M less than N-1) and having only negligible (perhaps noise-induced) thickness in the other dimensions.
For semantic systems, the total number of semantic dimensions may be far less than the number semantic terms or tokens. A simple example dimensional flattening is the existence of synonyms -- the second word does not expand the space because it does not encode a new dimension of meaning. (Synonyms would also be negatively correlated in Google searches, but that's another issue). Also, the fact that each word can be defined in terms of other words suggests that the semantic nebula does not actually fill the space.
Accomplishing this would require a true distance metric. I notice that NGD does not satisfy the triangle inequality. Perhaps some minor transform or alternative formulation of NGD would yield a true metric.
The reason that estimating semantic dimensionality is useful is two fold. First, it says something about the cognitive complexity of humans and human systems. Second, it provides some insight into the required cognitive sophistication of autonomous learning systemd that need to interact "intelligently" with humans. How many words does a system need to truly understand to pass the Turing test?
Creating a full reconstruction, a more challenging task, would provide insight into the structure of human language and human language usage patterns. The dimensionality of clusters of words might provide insight into the complexity of subdomains of knowledge.
I wish you every success in creating better autonomous learning systems.
Although very clever, NGD (Normalized Google Distance) misses alll higher-order relationships and does not even distinguish between different categories of pairwise relationships. For example, NGD might assume that "Bush" & "Iraq" had the same relationship as "Slashdot" & "Geek" because the two word pairs co-occur with similar frequencies.
More interesting are analyses on n-Tuples (co-occurences and orderings of n-words at a time). Anyone who does ER (Entity-Relationship) diagrams for relational databases will appreciate that many relationships involve multiple entities that are decomposable into pairwise relationships.
Another limit is that Google is atrocious on its estimates of the number of hits. The actual number of hits is only fraction (about 60%?) of the estimated from my experience. This suggests that Google has a pairwise estimator built in that may be only partially empirical. If Google simply reports an estimated number of hits based on products of probabilities, then their is no information about the pair in the NGD. Obviously, these scientists have gotten useful results, but NGD may not be as good an estimate of the co-occurence of the words as the scientists assume.
When Microsoft reimburses users of lost time and aggravation with security problems, then it will have a legitimate point of competitive advantage against Linux. Current EULAs limit damages to the price of the software. But a truly secure company (secure in both its software and its confidence in the security of its systems) should be willing to reimburse users that adhere to basic security protocols but are attacked through faulty software. Obviously, users that get 0wned by using insecure passwords, deactivating security systems, or failing to apply patches in a timely manner would not be eligible. But a securely-configured user that is compromised due to holes in the software would be reimbursed some appropriate amount.
Unfortunately, nobody seems willing to take this sort of pledge because they know that most software is a house of cards, security-wise. Perhaps its time for software vendors to put their money where their mouth is when they make security claims.
I wonder if one could digitally create a series of independent virtual limited liability companies for the sole purpose of testing orphanhood of abandonware books and media. Each LLC would be endowed with $0.01 in assets and be the titular legal entity attached to the free distribution of one potentially orphaned book. Should the author/heirs/owners come forward, the LLC would die instantly but not affect any of the other LLC orphan-media-representing companies.
IANAL, but I wonder how the tools and sensibilities of digital systems (high scalability at extremely low costs) can be applied to legal and corporate structures. If one can create a new company as easily as one creates a new domain name, some interesting (and some not so pleasant) things might occur.
There is no surprise why Apple does not need to use Gillette's strategy. The difference has to do with intrinsic and perceived value. First, whereas a Gillette razor handle is useless without a Gillette razor blade, an iPod is immediately useful with ripped CDs, MP3s, etc. iPod preceded iTunes download service and people where willing to buy iPods without the download service. iPod and iTune do complement each other, but not in the same obligatory way of razor handles and razor blades. Second, Gillette had a problem of lowering the hurdle of adoption -- people refused to buy the razor handle at full price not knowing if the new shaving system would work for them. In contrast, Apple's reputation for "stuff that just works" meant that they had no such hurdle. Apple fanatics would buy iPods sight unseen, tell the world, and drive adoption without Apple needing to discount the initial price of the player.
If anything, Apple's strategy is the reverse (TFA points this out) -- making little or no money on music and enjoying handsome margins on the hardware.
Google could also roll out a thin client service in which you do everything within any browser window connected to Google. Google could host user accounts that go beyond email and search. A person could browse through the google browser, manage their googlefiles, run googleoffice, send gmail, buy stuff through froogle, etc. It would be a totally portable thin client service.
Game AI will never pass the Turing test (by much margin) because it would spoil the game for newbies. Too smart an opponent or an opponent that learns the game faster than the average player would be too depressing.
Commercial game players will carefully limit the smarts of the AI to make it fun without letting the AI win all the time. Its better to have AI opponents that permit a high kill ratio for the player than to have opponents that kill the player multiple times per opponent killed.
Classifier Systems are a genetic algorithm analog for this type of streaming data/pattern analysis. With classifier systems a stream of incoming messages interacts with a constantly evolving population of classifier rules and an internally changing pool of working messages to create a stream of outputs. A reward/feedback loop drives adaption of the rule system to reinforce when it creates "good" outputs. The entire Classifier System concept is analogous to the mammalian immune system in the way that neural nets are analogous to brains and genetic algorithms are analogous to Darwinian evolution.
With a high enough stream processing speed (using StreamBase's methods), classifier systems might be useful for AI/adaptive learning scenarios.
What kind of responsibility do the game makers have to keeping a historical game accurate?
The same responsibility that book creators do.
But you raise a deeper point. What is the true purpose of learning history? Is it only to understand a set of facts about the past? Other than a flash-card game structure (rote learning wrapped in a game), history is ill-suited to gaming because history is fixed.
But what if the true purpose of learning history is it to prepare the student for making political decisions in the future. A game that teaches the consequences of political/governmental decisions may be more powerful than a historically-accurate docu-game. The student would be able to try alternative histories and learn the likely consequences (better or worse) of not sticking to history's script. A game, such as SimCity, could form the basis for some powerful lessons in civics and government.
Although books do transport one to another time and place, they are passive. A reader might imagine "what they would do" if they were in the character's pace, but they never get to try out that action. In a game, the player takes an active role: monitoring the situation, responding to the events of the game, and learning from their actions. The point is that if books have any built-in trial and error, it is a canned sequence that the reader has little involvement with.
It's a separate question of "what" people learn from games, especially violent fragfests, but that's a another topic.
I agree that open would be nice, but I can see several reasons why the camera's remain closed.
1. Competitive reasons: if documentation comes out before the camera (it would need to in order to make openness contribute to the camera's success), the competition gets to see/copy/out-market the new camera
2. Japanese: Most cameras are now designed and built by Japanese companies. Translation of the documentation, code, and specs into English (the open standard language) is a cost.
3. Proprietary chips: Companies such as Canon use custom hardware such as the Digic Processor making it hard to be open.
4. Support costs: Devoting engineers to supporting SDKs and openness initiatives would add cost.
The big problem is that camera electronics don't have the market persistence of other computer technologies. New generations of the electronics and software inside of cameras wholly supplant old ones. Its not like cameras run on decade's old protocols such as those found in networking and servers. The rapid design-sell-drop cycles leave little room for post-launch open source initiatives and the proprietary/competitive nature of the camera business leaves little room for design-phase involvement by non-employees.
As long as the camera supports a few simple open standard interface protocols (storage, USB, printing, etc.) then the internals can be (and will be) proprietary black boxes. At best, open source can help post-process images. Although a few geeks, such as the poster and myself, might complain at black-box cameras, we represent such a small sliver of the user base as to be easily ignored.
One of the wonders of software is that, through use of standards and abstraction, software will work on any kind of file. In layered networking models, for example, each layer is independent of the others to a great extent. The wiring, routers, transport protocols, and low-level network software don't know or care whether they are carrying bytes for a web page, a VOIP call, a Linux upgrade, or a bootlegged copy of a soon-to-be released "Major Motion Picture"(TM). DRM muddies the waters of abstraction and generalization by requiring specialization that is contextually sensitive to the type of data, the rights of the sender and the rights of the recipient. Just think about all the combinations of types of files, types of DRM restrictions, and combinations of senders and recipients (including fair use combinations where the sender and recipient are the same person, but different devices).
The point is that computers do not care about the subtleties of data types because there were designed, at their core, not to care. That lack of care is what enables the machines to do so much and be so extensible. Taking "reasonable care" is unreasonable.
Why does everyone always think that things must converge to some single future state? Regarding ad-blockers, I see three responses.
First, I'd wager that some sites will rearrange their content to be less pleasant to read with ad-blocking enabled or will create in-line text ads that are much harder to filter. Ad-hating people will stop visiting those sites, but the sites will still attract enough audience to survive. The number of free, ad-supported sites might decline, but will never go to zero.
Second, if anything, ad-blocking will further entrench the corporate subscription-only sites because it kills the natural migration path for small personal sites. Currently, a growing small site can recoup its bandwidth costs with ads. If that avenue is not open, then small sites must either sell-out to a big corporation or close up shop when the traffic gets too high.
Third, perhaps one solution is a bittorrent-like version of the WWW for small popular sites. Small sites that cannot afford to have a million or even a thousand daily viewers will submit their content to a bittorrent-like entity.
In short, technology and trends will mean that there will always be some number of big for-pay news sites (e.g., WSJ); medium-sized ad-supported sites (e.g.,/.), and small, free personal news sites (blogs).
horrible aerodynamic drag on paddle-wheel tires
on
Reinventing the Wheel
·
· Score: 1, Interesting
That picture with those paddle spokes in the tire makes me wonder about aerodynamic drag (and noise too). The top of a rotating tire has an air speed that is twice the vehicle speed. Those paddles should do wonders at turing gasoline into stirred the air. I can only hope that they will enclose the spokes in a smooth sidewall.
On the otherhand, if they angled the spokes properly, the tire would suck the air out from under the car and create a wonderful ground suction effect from improved cornering (but even worse fuel efficiency).
A friend in high school had an HP-45 and his dad had an HP-65 programmable calculator. This got me interested in computers. I bought an HP-25 at the end of sophomore year, learned to program (49 steps of assembly language memory!) and then moved upward.
Our high school also had a "math lab" with a paper teletype with 110 baud dial-up time-share access to the local university's computers (a Dec 10 and a CDC 6600/6400). A bunch of us proto-geeks spent before-school and after-school time mucking about on this. Using the TTY and a bit of help from another friend, I learned BASIC. When I outgrew BASIC, I latched on to APL which was a seriously cool language. Some of my friends had computers (if a 6502 with 16k of RAM, dual cassette tape drives, can be considered a computer).
I think "learning computers" means something very very different these days. In the early days, everything was programming and nobody used canned applications. If you wanted a computer to anything, you wrote the program yourself. Moreover, the operating systems of that day (especially for hobbyist systems) were extremely simple. One could understand what everything did both in software and in hardware.
These days, "learning computers" means more learning to use 3rd-party applications and learning to manage the OS. I'd wager that,/.ers excepted, most people don't ever learn to program these days. In the olden days programming was all you could do.
I just upgraded from Office 98 to Office 2004. What a complete waste of money. Aside from OS X code and antialiased fonts, the new version is less stable, slower, crankier, and festooned even more Microsoft User Interface Atrocities than ever. Six years and 3 versions later, Office has failed to fix most (any?) of the annoyances from the 1998 version. I guess near-100% market share means the company does not have to do anything to charge money for its double-speak "upgrades".
The challenge for IT security is that computer science loves to use abstractions, encapsulation, APIs, libraries and what not that let the programmer ignore the details of the internal complexity of systems. The problem is that it leads one to assume that these systems behave in some idealized fashion (the logical, black-box model of the system). In reality, the systems don't always follow the assumed logical model or the ignored internals create side-effects that are unforeseen by the original programmer, but exploited by malicious actors.
For example, assumptions about metadata and syntax give rise to buffer overflow or malformed string exploits. In trusting that an input string will be its stated length or follow the official syntax, the programmer adheres to the logical model of the system but creates a vulnerability. Similarly, physical power consumption artifacts can let a cracker guess the state or internal activities of a smartcard encryption chip. The original programmer is unaware that the code creates these artifacts since most coding paradigms ignore issues such as the exact execution time of subroutines, power consumption of CPU instructions, etc.
Becoming security conscious means unlearning all the tricks that let a programmer ignore the complexity inside a system. It means understanding the real behavior of all the internals, all the side-effects, and all the system properties that might be observable or influenceable by a malicious party. That makes programming for security very different and very much harder that standard programming.
To mangle a metaphor, security means that one must peel the onion to ensure that it does not have contain an open door in its core.
Something tells me that we already have budding multitaskers in our midst. Unfortunately, they are medicated into monotaskers because everyone thinks they have Attention Deficient Disorder ro some form of hyperactivity problem. Given that all the authorities are aging monotaskers, it no surprise that they can't deal with people who can't help multitasking because they have been raised in a task-rich, info-rich, stimulation-rich environment.
Do multitaskers have more children?
on
Life Interrupted
·
· Score: 1
If humanity is to "evolve" toward multitasking, then multitaskers must have more children than non-multitaskers. Evolution is not learned, it is breed in to the species. Given the "lack of sex" jokes on/. and the likelihood that/.ers are the closest humanity has to multitaskers, I'd wager that multitasking does not raise fitness levels.
Slashdot Mirror
This brings up the amusing problem of character recognition by human and non-human intelligences. Douglas Hofstadter discusses this issue in on seeing A's and seeing As.
In the case of this exploit, a deep flaw in IDN and computer fonts means that character #1072 is rendered typographically as an "a". The irony is that this is one of the few cases in which a computer can readily tell the difference between "a" and #1072 and a person cannot. The only solution would be rules that prohibit isomorphic characters in typefaces or a in-browser warning system that analyses the potential for ambiguity and alerts the user.
The issue is that the strenght of the signal is not a linear function of distance.
I agree with you, but don't think it hinders the system as much as it might seem. The key is that most base stations have limited range -- if I can detect the signal at some normal strength level (e.g., not using a cantenna), I know I'm within a 50-100 meters of the transmitter.
What got me thinking about this was my experience on a recent business trip. Sitting in my room at a bed-n-breakfast, I was picking up 4 wireless networks - the network of the B&B, the one from an adjacent B&B, and two other networks in the residential neighborhood. I doubt there are many places in which all four networks are detectable. Conversely, that means that if I can detect all four networks, I must be in a particular fairly small error circle (say within 1/4 to 1/2 a city block) within the overlap of the four networks.
A simple WiFi navigation system would use binary detect/no-detect data to determine location. If one can detect stations A, B, and R, then one must be in a particular street, near a particular corner.
Yes, some transmitters might be detectable over a much greater distances. These signals are much less useful for navigation because detecting them does not mean one is in any particular location. But with enough data, the system can easily learn to ignore signals from longer-range WiFi transmitters.
Yes, some transmitters might be temporarily undetectable due to being off or in a magic voodoo WiFi dead-zone. But the statistical logic of the navigation system could be tied to what is detectable and not use any evidence about what isn't detectable.
A more sophisticated system could use signal strength, but it would need a much greater volume of data in order to compute statistically meaningful location estimates. Although the signal drop-off is not isotropic, time-invariant, it is not statistically arbitrary. One is unlikely to get a strong signal at 150 meters from a normal WiFi station. Statistical variations in signal strength with distance would determine the error region around any estimate of location.
As I said, it will never beat GPS. But it would be cool to calculate where you are from who you can hear.
Rather that use dead-trees, wardrivers could use the names and signal strengths to nearby stations to estimate their position and calculate movements toward the best accessible WiFi hotspots. Even if the base station is password protected, its broadcast name and existence can help war drivers figure out where they are on the war-driving map. And if each wardriver submits data on stations and signal strengths, statistical combination of all the data could help refine the map further.
It will never beat GPS, but it would be cool to create a city-wide navigation system that works on WiFi -- "just turn left when you get max signal from MAC 00 60 1D 1C B9 0D."
The recent /, story on media longevity highlights the growing problem of decomposing data-layers on current generation optical disk technology. This new disk, with its even higher density, would seem to be even more likely to suffer from longevity problems.
Perhaps the xxAA has nothing to worry about -- media buying customers will lose access to copied data through dye-decomposition sooner than through expiring DRM licenses.
I wonder when CAPTCHAs will be so hard that an increasing fraction of the human population fails them. Perhaps the true origin of SkyNet will be when some spammer's AI realizes that humans are superfluous in an age of totally automated click-throughs and e-commerce.
Thank you for the reply. I'm glad your work generalizes to longer search-term lists. Like so many other /. readers, I did not take the time to read your preprint before posting.
I've often wondered if one can use simple pair-wise distance estimates to reconstruct a polytope or distorted simplex for the set of items within a multidimensional space. In theory, an N-object system, with non-zero pairwise distances, requires (N-1) dimensions. But in practice, many real systems don't fill the space -- being M-dimensional (M less than N-1) and having only negligible (perhaps noise-induced) thickness in the other dimensions.
For semantic systems, the total number of semantic dimensions may be far less than the number semantic terms or tokens. A simple example dimensional flattening is the existence of synonyms -- the second word does not expand the space because it does not encode a new dimension of meaning. (Synonyms would also be negatively correlated in Google searches, but that's another issue). Also, the fact that each word can be defined in terms of other words suggests that the semantic nebula does not actually fill the space.
Accomplishing this would require a true distance metric. I notice that NGD does not satisfy the triangle inequality. Perhaps some minor transform or alternative formulation of NGD would yield a true metric.
The reason that estimating semantic dimensionality is useful is two fold. First, it says something about the cognitive complexity of humans and human systems. Second, it provides some insight into the required cognitive sophistication of autonomous learning systemd that need to interact "intelligently" with humans. How many words does a system need to truly understand to pass the Turing test?
Creating a full reconstruction, a more challenging task, would provide insight into the structure of human language and human language usage patterns. The dimensionality of clusters of words might provide insight into the complexity of subdomains of knowledge.
I wish you every success in creating better autonomous learning systems.
Although very clever, NGD (Normalized Google Distance) misses alll higher-order relationships and does not even distinguish between different categories of pairwise relationships. For example, NGD might assume that "Bush" & "Iraq" had the same relationship as "Slashdot" & "Geek" because the two word pairs co-occur with similar frequencies.
More interesting are analyses on n-Tuples (co-occurences and orderings of n-words at a time). Anyone who does ER (Entity-Relationship) diagrams for relational databases will appreciate that many relationships involve multiple entities that are decomposable into pairwise relationships.
Another limit is that Google is atrocious on its estimates of the number of hits. The actual number of hits is only fraction (about 60%?) of the estimated from my experience. This suggests that Google has a pairwise estimator built in that may be only partially empirical. If Google simply reports an estimated number of hits based on products of probabilities, then their is no information about the pair in the NGD. Obviously, these scientists have gotten useful results, but NGD may not be as good an estimate of the co-occurence of the words as the scientists assume.
When Microsoft reimburses users of lost time and aggravation with security problems, then it will have a legitimate point of competitive advantage against Linux. Current EULAs limit damages to the price of the software. But a truly secure company (secure in both its software and its confidence in the security of its systems) should be willing to reimburse users that adhere to basic security protocols but are attacked through faulty software. Obviously, users that get 0wned by using insecure passwords, deactivating security systems, or failing to apply patches in a timely manner would not be eligible. But a securely-configured user that is compromised due to holes in the software would be reimbursed some appropriate amount.
Unfortunately, nobody seems willing to take this sort of pledge because they know that most software is a house of cards, security-wise. Perhaps its time for software vendors to put their money where their mouth is when they make security claims.
I wonder if one could digitally create a series of independent virtual limited liability companies for the sole purpose of testing orphanhood of abandonware books and media. Each LLC would be endowed with $0.01 in assets and be the titular legal entity attached to the free distribution of one potentially orphaned book. Should the author/heirs/owners come forward, the LLC would die instantly but not affect any of the other LLC orphan-media-representing companies.
IANAL, but I wonder how the tools and sensibilities of digital systems (high scalability at extremely low costs) can be applied to legal and corporate structures. If one can create a new company as easily as one creates a new domain name, some interesting (and some not so pleasant) things might occur.
There is no surprise why Apple does not need to use Gillette's strategy. The difference has to do with intrinsic and perceived value. First, whereas a Gillette razor handle is useless without a Gillette razor blade, an iPod is immediately useful with ripped CDs, MP3s, etc. iPod preceded iTunes download service and people where willing to buy iPods without the download service. iPod and iTune do complement each other, but not in the same obligatory way of razor handles and razor blades. Second, Gillette had a problem of lowering the hurdle of adoption -- people refused to buy the razor handle at full price not knowing if the new shaving system would work for them. In contrast, Apple's reputation for "stuff that just works" meant that they had no such hurdle. Apple fanatics would buy iPods sight unseen, tell the world, and drive adoption without Apple needing to discount the initial price of the player.
If anything, Apple's strategy is the reverse (TFA points this out) -- making little or no money on music and enjoying handsome margins on the hardware.
Google could also roll out a thin client service in which you do everything within any browser window connected to Google. Google could host user accounts that go beyond email and search. A person could browse through the google browser, manage their googlefiles, run googleoffice, send gmail, buy stuff through froogle, etc. It would be a totally portable thin client service.
Game AI will never pass the Turing test (by much margin) because it would spoil the game for newbies. Too smart an opponent or an opponent that learns the game faster than the average player would be too depressing.
Commercial game players will carefully limit the smarts of the AI to make it fun without letting the AI win all the time. Its better to have AI opponents that permit a high kill ratio for the player than to have opponents that kill the player multiple times per opponent killed.
Classifier Systems are a genetic algorithm analog for this type of streaming data/pattern analysis. With classifier systems a stream of incoming messages interacts with a constantly evolving population of classifier rules and an internally changing pool of working messages to create a stream of outputs. A reward/feedback loop drives adaption of the rule system to reinforce when it creates "good" outputs. The entire Classifier System concept is analogous to the mammalian immune system in the way that neural nets are analogous to brains and genetic algorithms are analogous to Darwinian evolution.
With a high enough stream processing speed (using StreamBase's methods), classifier systems might be useful for AI/adaptive learning scenarios.
But wouldn't this be extremely driven by the game developer's views on both history and human interaction?
Absolutely true! Are textbooks any different? Was Christopher Columbus a hero or a villain? It depends on which textbook you read.
What kind of responsibility do the game makers have to keeping a historical game accurate?
The same responsibility that book creators do.
But you raise a deeper point. What is the true purpose of learning history? Is it only to understand a set of facts about the past? Other than a flash-card game structure (rote learning wrapped in a game), history is ill-suited to gaming because history is fixed.
But what if the true purpose of learning history is it to prepare the student for making political decisions in the future. A game that teaches the consequences of political/governmental decisions may be more powerful than a historically-accurate docu-game. The student would be able to try alternative histories and learn the likely consequences (better or worse) of not sticking to history's script. A game, such as SimCity, could form the basis for some powerful lessons in civics and government.
Although books do transport one to another time and place, they are passive. A reader might imagine "what they would do" if they were in the character's pace, but they never get to try out that action. In a game, the player takes an active role: monitoring the situation, responding to the events of the game, and learning from their actions. The point is that if books have any built-in trial and error, it is a canned sequence that the reader has little involvement with.
It's a separate question of "what" people learn from games, especially violent fragfests, but that's a another topic.
I agree that open would be nice, but I can see several reasons why the camera's remain closed.
1. Competitive reasons: if documentation comes out before the camera (it would need to in order to make openness contribute to the camera's success), the competition gets to see/copy/out-market the new camera
2. Japanese: Most cameras are now designed and built by Japanese companies. Translation of the documentation, code, and specs into English (the open standard language) is a cost.
3. Proprietary chips: Companies such as Canon use custom hardware such as the Digic Processor making it hard to be open.
4. Support costs: Devoting engineers to supporting SDKs and openness initiatives would add cost.
The big problem is that camera electronics don't have the market persistence of other computer technologies. New generations of the electronics and software inside of cameras wholly supplant old ones. Its not like cameras run on decade's old protocols such as those found in networking and servers. The rapid design-sell-drop cycles leave little room for post-launch open source initiatives and the proprietary/competitive nature of the camera business leaves little room for design-phase involvement by non-employees.
As long as the camera supports a few simple open standard interface protocols (storage, USB, printing, etc.) then the internals can be (and will be) proprietary black boxes. At best, open source can help post-process images. Although a few geeks, such as the poster and myself, might complain at black-box cameras, we represent such a small sliver of the user base as to be easily ignored.
One of the wonders of software is that, through use of standards and abstraction, software will work on any kind of file. In layered networking models, for example, each layer is independent of the others to a great extent. The wiring, routers, transport protocols, and low-level network software don't know or care whether they are carrying bytes for a web page, a VOIP call, a Linux upgrade, or a bootlegged copy of a soon-to-be released "Major Motion Picture"(TM). DRM muddies the waters of abstraction and generalization by requiring specialization that is contextually sensitive to the type of data, the rights of the sender and the rights of the recipient. Just think about all the combinations of types of files, types of DRM restrictions, and combinations of senders and recipients (including fair use combinations where the sender and recipient are the same person, but different devices).
The point is that computers do not care about the subtleties of data types because there were designed, at their core, not to care. That lack of care is what enables the machines to do so much and be so extensible. Taking "reasonable care" is unreasonable.
Why does everyone always think that things must converge to some single future state? Regarding ad-blockers, I see three responses.
/.), and small, free personal news sites (blogs).
First, I'd wager that some sites will rearrange their content to be less pleasant to read with ad-blocking enabled or will create in-line text ads that are much harder to filter. Ad-hating people will stop visiting those sites, but the sites will still attract enough audience to survive. The number of free, ad-supported sites might decline, but will never go to zero.
Second, if anything, ad-blocking will further entrench the corporate subscription-only sites because it kills the natural migration path for small personal sites. Currently, a growing small site can recoup its bandwidth costs with ads. If that avenue is not open, then small sites must either sell-out to a big corporation or close up shop when the traffic gets too high.
Third, perhaps one solution is a bittorrent-like version of the WWW for small popular sites. Small sites that cannot afford to have a million or even a thousand daily viewers will submit their content to a bittorrent-like entity.
In short, technology and trends will mean that there will always be some number of big for-pay news sites (e.g., WSJ); medium-sized ad-supported sites (e.g.,
That picture with those paddle spokes in the tire makes me wonder about aerodynamic drag (and noise too). The top of a rotating tire has an air speed that is twice the vehicle speed. Those paddles should do wonders at turing gasoline into stirred the air. I can only hope that they will enclose the spokes in a smooth sidewall.
On the otherhand, if they angled the spokes properly, the tire would suck the air out from under the car and create a wonderful ground suction effect from improved cornering (but even worse fuel efficiency).
A friend in high school had an HP-45 and his dad had an HP-65 programmable calculator. This got me interested in computers. I bought an HP-25 at the end of sophomore year, learned to program (49 steps of assembly language memory!) and then moved upward.
/.ers excepted, most people don't ever learn to program these days. In the olden days programming was all you could do.
Our high school also had a "math lab" with a paper teletype with 110 baud dial-up time-share access to the local university's computers (a Dec 10 and a CDC 6600/6400). A bunch of us proto-geeks spent before-school and after-school time mucking about on this. Using the TTY and a bit of help from another friend, I learned BASIC. When I outgrew BASIC, I latched on to APL which was a seriously cool language. Some of my friends had computers (if a 6502 with 16k of RAM, dual cassette tape drives, can be considered a computer).
I think "learning computers" means something very very different these days. In the early days, everything was programming and nobody used canned applications. If you wanted a computer to anything, you wrote the program yourself. Moreover, the operating systems of that day (especially for hobbyist systems) were extremely simple. One could understand what everything did both in software and in hardware.
These days, "learning computers" means more learning to use 3rd-party applications and learning to manage the OS. I'd wager that,
I just upgraded from Office 98 to Office 2004. What a complete waste of money. Aside from OS X code and antialiased fonts, the new version is less stable, slower, crankier, and festooned even more Microsoft User Interface Atrocities than ever. Six years and 3 versions later, Office has failed to fix most (any?) of the annoyances from the 1998 version. I guess near-100% market share means the company does not have to do anything to charge money for its double-speak "upgrades".
Sorry for the rant.
The challenge for IT security is that computer science loves to use abstractions, encapsulation, APIs, libraries and what not that let the programmer ignore the details of the internal complexity of systems. The problem is that it leads one to assume that these systems behave in some idealized fashion (the logical, black-box model of the system). In reality, the systems don't always follow the assumed logical model or the ignored internals create side-effects that are unforeseen by the original programmer, but exploited by malicious actors.
For example, assumptions about metadata and syntax give rise to buffer overflow or malformed string exploits. In trusting that an input string will be its stated length or follow the official syntax, the programmer adheres to the logical model of the system but creates a vulnerability. Similarly, physical power consumption artifacts can let a cracker guess the state or internal activities of a smartcard encryption chip. The original programmer is unaware that the code creates these artifacts since most coding paradigms ignore issues such as the exact execution time of subroutines, power consumption of CPU instructions, etc.
Becoming security conscious means unlearning all the tricks that let a programmer ignore the complexity inside a system. It means understanding the real behavior of all the internals, all the side-effects, and all the system properties that might be observable or influenceable by a malicious party. That makes programming for security very different and very much harder that standard programming.
To mangle a metaphor, security means that one must peel the onion to ensure that it does not have contain an open door in its core.
Something tells me that we already have budding multitaskers in our midst. Unfortunately, they are medicated into monotaskers because everyone thinks they have Attention Deficient Disorder ro some form of hyperactivity problem. Given that all the authorities are aging monotaskers, it no surprise that they can't deal with people who can't help multitasking because they have been raised in a task-rich, info-rich, stimulation-rich environment.
If humanity is to "evolve" toward multitasking, then multitaskers must have more children than non-multitaskers. Evolution is not learned, it is breed in to the species. Given the "lack of sex" jokes on /. and the likelihood that /.ers are the closest humanity has to multitaskers, I'd wager that multitasking does not raise fitness levels.