Downloading is time consuming and bandwidth killing, but on the other hand old copies of Wikipedia, from when you have read them, contain the then current info. You can always go back to something stable if you need to reference something, as opposed to the current Wikipedia pages, which, by the time you need to look up something you read a while ago, they are long gone and edited out by other users. I do have an issue with how some of the pages get edited, and it's nice to hang on to a copy that I used to like. I can even foresee a future when the assault on wikipedia is so strong, that its current quality goes to complete crap, but then the goodies oldies come in handy. I do agree with your idea of diff or patch updates, incremental update snapshots, weekly, monthly, yearly.
Is hotmail nasty? I used to have a hotmail account right until MS bought them out. Then I just retained my yahoo email addresses. Now I'm paying for yahoo mail plus, but probably would have kept the service even if yahoo was bought out, simply because I'm worn out, worn out from being fed up with MS. Or maybe I'd move over to gmail completely. It depends who needs more help to maintain balance power, MS or Google. I always flock to the weaker side.
By the way it's clear why MS wants to buy Yahoo. According to http://www.alexa.com/site/ds/top_sites?ts_mode=lang&lang=en
Yahoo and Google are top. In a world where Silverlight is meant to replace Flash, if Yahoo and Google don't push silverlight but prefer flash instead, MS simply can't shove their will down everyone's throat, and can't make it a marketing success. It all depends on rankings. And people like me are so stuck in their ways, it's hard for MS to come up with anything "better" than Yahoo to make me switch to them and use their pages instead. It's bad enough that windows talks home any chance it gets. Even in maps, I prefer Google maps compared to whatever MS has, simply because Google isn't supplying my operating system, so there is a smaller chance of hidden easter egg-like backdoors being exploited by them.
There is a business aspect to this move. It will both cause people to upgrade to Vista, or purchase extra copies of XP while they are available, instead of waiting to make the purchasing decision when and if they are needed. The time value of money: make them spend their money today when they are not really sure they need something, instead of not spending it tomorrow when they are sure they don't need something. It can only help the bottom line. I don't see people driven to Linux while that's so difficult to use on the desktop (can't even surf the net right, flash 9 available? How about all the sites demanding internet explorer?), or even Apple, while that's so expensive.
Actually, MS has to make money. In the past they were so open about the Windows API, that it's practically impossible to close it now. Everything free can ride on it - gtk, gnu mingw, firefox, wxwidgets, and downright api compatible OS copies are available, like ReactOS, that do everything that Windows does, just like Linux was api compatible with unixes when it was created. Microsoft doesn't wanna be Linuxed out by a Reactos. What Microsoft needs is doing the same thing as XP as far as speed is concerned, just in an API incompatible way, at least until the world catches up, and then it's time to drop the whole thing again, like a hot potato, and get something new again. It's the only way to make money and keep the competition out, shift faster than they can keep up adapting to you. Now you have some borg that adapt to others in order to slowly assimilate them, but keep from being assimilated by creating a moving target out of themselves, shifting faster than others can adapt to them.
There is a million year old mating behavior hardcoded into male and female brains. It's not that women don't know what they want, it's that they are testing males. They can't help but follow their natural instincts in choosing a mate. As far as passing on genes go, females have to be more careful in their selection than men, so they send a lot of mixed signals, and wring and torture their prospective males emotionally just to make sure they won't quit. As far as a male is concerned, the most surviving male gene is one that goes around and inseminates as many females as possible, even if he has to lie to all of them. As far as a female is concerned, every choice she makes for a male that results in offsprings will have to be followed up by two decades of rearing, caring, nurturing and education. She has a lot more success in this if she has a male by her side, supporting her, dedicated to her. But she doesn't have forever to wait around, she has to hurry up more than males, the biological clock ticks faster, so such behavior of playing hard to get depends on the amount of attention she can summon around her, and ultimately all females have to find a way to be whores, and they have to do it in a hurry. They don't know what they want? They probably do the most efficient job ever of balancing the testing, playing hard to get, and dressing down, attracting attention, getting down to being a whore. There is a lot of lying on the male side, making females believe he'll be dedicated, just for her to find out that he is not even there in the morning, he jumps boat as soon as the act is done, ugh. This is somewhat hardcoded too, on the male side, but there is a subtle balance, because liers don't get a competitive advantage if their offsprings don't make it, while if there is no lying at all the female will think her sons won't be as "able" to get around a lot, so females love to be lied to, or at least they love males that get around a lot, and they try to lock them down for exclusively themselves. Good guys finish last, bad boys are such turnons, ehh? There is a reason why females can read men better than the other way around, they have to make the sophisticated decisions of 20 years investment in a single act of a moment, while a man goes by yeah, she's hot, sexy, I wanna boink her, they are less interested in whether "she'll stick around", or how her personality is. Actually that's not true. A lot of men go by personality, not just looks, and how sensitive, caring and even forgiving a female can be, basically how good a mother she makes. That's why bossy, tomboy, aggressive females have a harder time with men, than soft, sweet, sexy ladies, it's simply a hardcoded behavior too. There is a balance there too, care, softness vs. bossiness, you can't be a complete pushover as a mother, but being bossy is the easiest thing to be when you have complete reign over someone else, it's actually hard to have self control in it, so, while you can't be too soft, the balance is very much to the soft side, compared to the aggressive side for a female.
Wow. That's really really far, a world apart. 3 minutes ain't too bad, but still, just having to wait and stare at the clock before someone can answer, for 3 long minutes, while thoughts fly across outer space, at the speed of light - that's very far.
The current timestamp authorities are not mirrored, they need to see the full text of your ip, and they function via email, and there is no money trail (which is not needed until there is massive spam.) Basically the powers that be can stop your email from getting through to any kind of timestamping authority, by reading the contents of the texts. Which basically means you just need to send them the md5sum as a standard practice, instead of the plain text document and the rest functions fine. I guess you already have the service I was trying to describe above in a highly convoluted way, that could be fine-tuned a bit, and made into standard practice for most notary publics, as an electronic duplicate practice to the signed paper document. It would drop the price of notarizing through the floor, but it could increase what gets notarzied and ease the burden on the court system over uncertainties of the past, trying to decide which witness is lying, a wonderful thing for society, but only if it becomes standard practice. PGP has been around for well over a decade, yet I still have to sign my first email with it, mostly because my emails are not that important. If I had a patent idea, that'd be different, I'd care about a timestamp/proof of being first/being the author.
That's a nice thing to have, but it does not prove anything more than a notary public would. Ideally there should be a technological way to timestamp/prove before which even kings or other people with power should be equal to the last homeless person on the street. Right now notary publics are corruptible because they are people, and can deny ever signing a document, claiming the signature is forged. Ideally there should be a completely free and mirrored/highly distributed way of getting and storing certificates, or "keys" that represent pretty much the same thing as a notary public signature. You wouldn't need to trust people, you'd only have to trust math, and openly accessible records of timestamps. Anybody who ever wanted a timestamp/certificate for any kind of "text" should be able to publicly disclose the md5sum-typof-thing of such "text" to these multiple mirrored servers, some timestamp certificate, etc.. lemme return to this because: You still need to somehow reference "time", which we already know it's relative, (it flows faster on a satellite than down here in this gravity well, even if it's nanosecond different), so you still need a "time" authorithy such as time.gov. They should be issueing encrypted timestamp replies to everyone supplying a public key with their request, and they should be encrypting the md5sum + public key together with their own key. Gazillions of such requests should be stored in multiple mirrored databases that are free to download and duplicate. Of course for a cost, so that it not's spammed, for the cost of keeping a record line in a database, such as 2 cents. The idea here is that you can create a "text", such as a patent document, (or any sequence of 0's and 1's), get the md5sum, get a timestamp certificate for that md5sum from a time authority, the key including the current time and the md5sum, encrypted duplicate with your public key and the time authority's public key. Then you submit this double encrypted md5sum key to a multiple mirrored notary public service, and pay a 2 cents fee. They have no clue what you're notarizing, other than you're doing it. Then you can go apply for a patent, and not fear your patent lawyer will steal your idea and claim he invented it first. Let's suppose someone disputes your claim that you were the first to invent something. Then all you need to do is go to court, and point to a publicly notarized md5sum in the many databases, databases that md5sum themselves hourly then from the hourly md5's compound a daily one, from the daily ones weekly, etc, to make sure nobody can go back and remove a single notary public item. This notary public even only you can decrypt and the time authority should be able to duplicate your effort as a separate excersize, but you alone decrypting it should be proof in case we have to mistrust time authorities. Then you can show that this md5sum is done on such and such text, such and such sequence of 0's and 1's. It would be very hard for your opponent to do a similar excersize on his claim, because it's very hard to create colliding md5sums, etc. You can still be held at gunpoint to provide your own secret part of the public key you used, together with the full document, etc, and then they can go to court and claim that they were the ones obtaining that timestamp. But if you got two, three such timestamps, and only release one of them at gunpoint, how can they tell there is more? They could shoot you, but can't be sure that someone else doesn't know about it? Ideally there should be a way to prove identity in court, such as getting your own social security number mingled in into the md5sum keys, and then even if they extort the information from you, the court sees the notary public line item notarized for a specific social security number, yours. Of course the social security database could be edited too, showing that you don't exist, you have to trust the government not to do that. Basically there is no such thing as security, but there are different degrees of security. In order to misprove your claim, you have to have
As long as we are talking about plastic scissors with rounded tips that are child safe, what about VB classic, or even shell scripts? C++, C#, Java don't really feel like plastic scissors, nor surgeon's knives.
Not to mention low power. Microwatts, nanowatts. Unfortunately it's hard to play Pacman on it, let alone something like Unreal Tournament, but I've seen a paddle/ball video game implementation. If they could take that simplicity, and make it big and fast, where it can drive a screen better, but so simply that even a dumbo can program it, it would really flourish. It needs no harddrive - but you could connect one externally -, it needs no display - but you could connect one externally (my monitor has umm, how many pins? a 40 pin chip should have enough pins) - no keyboard, but you could connect one externally, etc. As far as pincount goes, some of the interfaces could be simplified to how USB2 simplified the how many pin parallel port, and if the feature set is complicated, provide and on-chip mechanism to do the translation (they have RS232 built in to newer chips, but you can do the bit banging by hand if you please, roll your own.) That's how it should be with all peripherals - the specs simple so even a dumbo can do the bit banging, and peripherals should be optional. None of this "drivers" mess in i386 that nobody knows what the heck is going on, and results in crashes so much. And once it reaches the speeds of a Commodore C64 or Z80, it's pretty usable and fun. Yes, I know the Z80 is there, you are welcome to learn it and play away with it, but it's too complex, the motherboard a Z80 runs on is not dumbo friendly. How about instead of Z80's, PIC's with USB stick memory, and some easy LCD display upsided from the 16x2 displays you see, with reduced pincount? You could have a whole range of chips, from superlow microwatts they have now, up to superfast 1GHz P3 level, and in all cases be able to have a standalone chip, with max 40 pins, or the option to connect harddrive, usb stick, display, keyboard, rs232, if you want, that is. Such a standard where things are simple and grandma can bitbang it would work wonders in industry/spaceshuttle/everything, because everyone could do it as opposed to a very specialized few.
Exactly. Moreover practice makes the master, but interesting practice at that, not boring. There is nothing like learning while having fun. Just look at little kittens.
The links are specially crafted by google to point to digg.com, a new slashdot contender, that points to a lot of videos. Videos/Flash are the ticket to undermining free and open webstandards, that's why they are so heavily pushed, but it'd be hard to and too obvious to push them on slashdot, because everyone would moan too hard here. Why can't there be free video standards everyone adheres to.. well, cuz...
I agree that win nt 3.5/4.0/5.0=approx. 5.1 being massive improvements in stability, but I've still seen them crash, even recently with such things as ASIO. Some things, such as audio drivers, simply have no time to wait around for the windows message system to let them speak up and get some action done. How do you balance stability/security with performance? Yes, win95 was relatively crashy, but very usable. If you tried running a win2k equivalent on a 486DX100 with 8 MB RAM, you may not have had a pleasurable performance - sometimes a crash is a smaller sacrifice than a constant drag and slowness. Note that power consumption was a lot less those days, didn't even need a fan sometimes, just a huge heatsink even on some P200 computers, and you got a lot of bang per buck as far as work per watt electricity consumed went, and it was good enough for such things as word processing, even spreadsheets, and definitely web browsing, but not good enough for multimedia big jpg's, high compression video codecs, etc. At least today the processor market is moving towards variable clockspeed, and low power, where they leave the 486's in the dust compared to idle watts, and processing watts, and easily deal with high performance demands by temporarily kicking up the clockspeed using lots of watts, but still, the OS market with the security paranoia, plus the dotnet "vision" for patentability, needing to run a lot of wasteful "checking" things in the background, hasn't caught up to this performance/watt issue.
This routine is good for precision. In the article he mentions some swede who had to get 48 significant digits for a thermo calculation for a doctorate student. Note that the 2nd iteration in the pdf is orders of magnitude more precise then the first, so all you need is enough iterations. You can always do a (do while current-previous(or a few previous moving average) epsilon), and build a loop, but that has to test a lot of conditionals ready for a branch(deep cpu pipeline predictors don't like branches) as opposed to just lining up 1 or 2 iterations in the code that makes it fly through the pipeline. True, when you do a lot of iterations, the speed gain from getting the first iteration so accurate doesn't count that much. It would save say 2 cycles out of a 20 cycle step, as opposed to making all the difference when there is only one iteration carried out, without a loop, as it is in the sample code given.
I see. So you say the unpredictable part is the assumption of floating point representation in sign, mantissa and exponent terms, which can vary from architecture to architecture.
Yes, but most biological matter is easily degraded by UV radiation, plus infection - bacteria would love to feast on your new solar cells.
As a sidenote, to quote the artilce: "Dr Stephen Curry Opens in new window, a structural biologist from Imperial College London's Division of Cell and Molecular Biology who participated in the research explains: "This work has shown that it is possible to manipulate molecules and proteins that occur naturally in the human body by changing one small detail of their make-up, such as the type of metal at the heart of a porphyrin molecule, as we did in this study.
Naturally occur in a human? I was hoping they'd be talking about cow-derived materials, unless they are interested in genetically engineering photosynthetic human beings? Maybe one of our great great grand children will be engineered enough to be vacuum resistant and fully photosynthetic, then he can fly around in outer space while living off of sunshine.
I think the original constant might have been derived with simple Monte Carlo methods, and the method itself is probably from some book from long time ago, from the 50's when vacuum tube computers first came onto the scene and such speed gains by optimizing binary code to the limit by phd mathematicians was at the forefront of research. To quote the pdf article you cite, "However the analysis should be thoroughly tested on a real machine, since hardware often does not play well with theory." That's nonsense - the machine is predictable, therefore plays well with theory, it's you who get overly engulfed and lost in the derivations by following too many threads of thought and losing track of your assumptions and approximations, it's the theory that's got the problem, and the machine is just an easy and lazy way to double check your thoughts, but you could technically do the same calculations the machine does with a paper and pencil, it would just take longer. I think this article is a bait, to see who will point out subtle issues in the report that might have been directly planted, and such people who dare figure out wat's wrong will instantly win a prize of a dozen shitflies settling down onto them from the NSA/FBI/CIA/MSFT/Dept. of Homeland security, and instantly win a "slavejob-dreamjob" at a fancy place or more like a "fancy prison" of none of their choice, to continue developing their talent and put it to good work for Da Man.
The day the Iraqi war started, someone asked for my opinion. I said, pssh, dumbasses, how are you gonna tell civilians from combatants? He said, just kill'em all. Well, that's your answer. It's not that you can't win, but the cost of winning is too high. And I'd like to think it's humanity holding these jerks back, and not the fact that if they did kill'em all, then that'd be the end of the winners too as far as the rest of the world is concerned, because they can't kill'em all the rest of the world, at least not just yet.
Let me note that I'm less trustful of search engine results than peer contributed wikipedia text additions, as far as trust goes, especially on touchy subjects. This will last as long as there are very few very centralized search engines in the world. I used to use Alltheweb for a while to double check things from other search engines, until yahoo scooped it up in a merger and killed it. And very recently google seems to increasingly point me to articles that require paid access. How come google has paid access? They are supposed to crawl the web like any user could, and read what any user could. I used to come to google with the assumption that a link they gave me had cached text and was accessible to me just like to their crawler, but now it's turning into a marketing front for paid only content. Soon the majority of results might either direct you to paid content only, and crappy free content, without showing you the quality free content, because there is a very serious financial incentive in play, paid for by their paid information sellers giving special access to google crawlers being in bed with the search engine. Where there is a buck to be made by screwing you over how can you have trust? Invisible hand? I don't mind seeing google find matches in journal abstracts that I myself can read freely, and then buy the hidden content, but a lot of their matches show fragments of text from inside the paid content, not the freely available abstract. I think that would be the purpose of an abstract. How can you trust search engines anymore? Wikipedia is not that trustworthy, but is anything else in this world?
Offtopic, but why does your http://everything2.com/index.pl?node_id=735553 everything2 page list last seen 1.4 years ago. One is inclined to think you've been "eliminated." If you're still on the web it'd be nice to put a comment there on "I've been mostly active on slashdot, etc.. not coming here much anymore." Just a suggestion.
Unless the US wants a quick and dirty win by nuking and leveling the rest of the world to the ground, the production capacity of the rest of the world is a significant military asset. You could have made this same argument in say mid 1930's about the Germans having superior military reserves, and making political decisions based on them, but it wasn't until the US production and economy was put in full gear to military service with serious devotion that such statements lost their meanings very quickly. If Japan for instance seriously devoted themselves to producing weapons, within two years you might have better military technology present in the world than anything currently available. But for now, they instead devote their attention to building better cars, or such things as beating speed records in train transportation.
It's not that simple. Those who don't like it don't like it not because they want to know about something, but because they are control freaks, they'd like to sell the same "knowledge" to you, or collect a membership fee, but it's hard when there is a free alternative, free as in both freedom and beer.
Slashdot Mirror
Downloading is time consuming and bandwidth killing, but on the other hand old copies of Wikipedia, from when you have read them, contain the then current info. You can always go back to something stable if you need to reference something, as opposed to the current Wikipedia pages, which, by the time you need to look up something you read a while ago, they are long gone and edited out by other users. I do have an issue with how some of the pages get edited, and it's nice to hang on to a copy that I used to like. I can even foresee a future when the assault on wikipedia is so strong, that its current quality goes to complete crap, but then the goodies oldies come in handy. I do agree with your idea of diff or patch updates, incremental update snapshots, weekly, monthly, yearly.
Is hotmail nasty? I used to have a hotmail account right until MS bought them out. Then I just retained my yahoo email addresses. Now I'm paying for yahoo mail plus, but probably would have kept the service even if yahoo was bought out, simply because I'm worn out, worn out from being fed up with MS. Or maybe I'd move over to gmail completely. It depends who needs more help to maintain balance power, MS or Google. I always flock to the weaker side. By the way it's clear why MS wants to buy Yahoo. According to http://www.alexa.com/site/ds/top_sites?ts_mode=lang&lang=en Yahoo and Google are top. In a world where Silverlight is meant to replace Flash, if Yahoo and Google don't push silverlight but prefer flash instead, MS simply can't shove their will down everyone's throat, and can't make it a marketing success. It all depends on rankings. And people like me are so stuck in their ways, it's hard for MS to come up with anything "better" than Yahoo to make me switch to them and use their pages instead. It's bad enough that windows talks home any chance it gets. Even in maps, I prefer Google maps compared to whatever MS has, simply because Google isn't supplying my operating system, so there is a smaller chance of hidden easter egg-like backdoors being exploited by them.
There is a business aspect to this move. It will both cause people to upgrade to Vista, or purchase extra copies of XP while they are available, instead of waiting to make the purchasing decision when and if they are needed. The time value of money: make them spend their money today when they are not really sure they need something, instead of not spending it tomorrow when they are sure they don't need something. It can only help the bottom line. I don't see people driven to Linux while that's so difficult to use on the desktop (can't even surf the net right, flash 9 available? How about all the sites demanding internet explorer?), or even Apple, while that's so expensive.
Actually, MS has to make money. In the past they were so open about the Windows API, that it's practically impossible to close it now. Everything free can ride on it - gtk, gnu mingw, firefox, wxwidgets, and downright api compatible OS copies are available, like ReactOS, that do everything that Windows does, just like Linux was api compatible with unixes when it was created. Microsoft doesn't wanna be Linuxed out by a Reactos. What Microsoft needs is doing the same thing as XP as far as speed is concerned, just in an API incompatible way, at least until the world catches up, and then it's time to drop the whole thing again, like a hot potato, and get something new again. It's the only way to make money and keep the competition out, shift faster than they can keep up adapting to you. Now you have some borg that adapt to others in order to slowly assimilate them, but keep from being assimilated by creating a moving target out of themselves, shifting faster than others can adapt to them.
There is a million year old mating behavior hardcoded into male and female brains. It's not that women don't know what they want, it's that they are testing males. They can't help but follow their natural instincts in choosing a mate. As far as passing on genes go, females have to be more careful in their selection than men, so they send a lot of mixed signals, and wring and torture their prospective males emotionally just to make sure they won't quit. As far as a male is concerned, the most surviving male gene is one that goes around and inseminates as many females as possible, even if he has to lie to all of them. As far as a female is concerned, every choice she makes for a male that results in offsprings will have to be followed up by two decades of rearing, caring, nurturing and education. She has a lot more success in this if she has a male by her side, supporting her, dedicated to her. But she doesn't have forever to wait around, she has to hurry up more than males, the biological clock ticks faster, so such behavior of playing hard to get depends on the amount of attention she can summon around her, and ultimately all females have to find a way to be whores, and they have to do it in a hurry. They don't know what they want? They probably do the most efficient job ever of balancing the testing, playing hard to get, and dressing down, attracting attention, getting down to being a whore. There is a lot of lying on the male side, making females believe he'll be dedicated, just for her to find out that he is not even there in the morning, he jumps boat as soon as the act is done, ugh. This is somewhat hardcoded too, on the male side, but there is a subtle balance, because liers don't get a competitive advantage if their offsprings don't make it, while if there is no lying at all the female will think her sons won't be as "able" to get around a lot, so females love to be lied to, or at least they love males that get around a lot, and they try to lock them down for exclusively themselves. Good guys finish last, bad boys are such turnons, ehh? There is a reason why females can read men better than the other way around, they have to make the sophisticated decisions of 20 years investment in a single act of a moment, while a man goes by yeah, she's hot, sexy, I wanna boink her, they are less interested in whether "she'll stick around", or how her personality is. Actually that's not true. A lot of men go by personality, not just looks, and how sensitive, caring and even forgiving a female can be, basically how good a mother she makes. That's why bossy, tomboy, aggressive females have a harder time with men, than soft, sweet, sexy ladies, it's simply a hardcoded behavior too. There is a balance there too, care, softness vs. bossiness, you can't be a complete pushover as a mother, but being bossy is the easiest thing to be when you have complete reign over someone else, it's actually hard to have self control in it, so, while you can't be too soft, the balance is very much to the soft side, compared to the aggressive side for a female.
Wow. That's really really far, a world apart. 3 minutes ain't too bad, but still, just having to wait and stare at the clock before someone can answer, for 3 long minutes, while thoughts fly across outer space, at the speed of light - that's very far.
The current timestamp authorities are not mirrored, they need to see the full text of your ip, and they function via email, and there is no money trail (which is not needed until there is massive spam.) Basically the powers that be can stop your email from getting through to any kind of timestamping authority, by reading the contents of the texts. Which basically means you just need to send them the md5sum as a standard practice, instead of the plain text document and the rest functions fine. I guess you already have the service I was trying to describe above in a highly convoluted way, that could be fine-tuned a bit, and made into standard practice for most notary publics, as an electronic duplicate practice to the signed paper document. It would drop the price of notarizing through the floor, but it could increase what gets notarzied and ease the burden on the court system over uncertainties of the past, trying to decide which witness is lying, a wonderful thing for society, but only if it becomes standard practice. PGP has been around for well over a decade, yet I still have to sign my first email with it, mostly because my emails are not that important. If I had a patent idea, that'd be different, I'd care about a timestamp/proof of being first/being the author.
That's a nice thing to have, but it does not prove anything more than a notary public would. Ideally there should be a technological way to timestamp/prove before which even kings or other people with power should be equal to the last homeless person on the street. Right now notary publics are corruptible because they are people, and can deny ever signing a document, claiming the signature is forged. Ideally there should be a completely free and mirrored/highly distributed way of getting and storing certificates, or "keys" that represent pretty much the same thing as a notary public signature. You wouldn't need to trust people, you'd only have to trust math, and openly accessible records of timestamps. Anybody who ever wanted a timestamp/certificate for any kind of "text" should be able to publicly disclose the md5sum-typof-thing of such "text" to these multiple mirrored servers, some timestamp certificate, etc.. lemme return to this because: You still need to somehow reference "time", which we already know it's relative, (it flows faster on a satellite than down here in this gravity well, even if it's nanosecond different), so you still need a "time" authorithy such as time.gov. They should be issueing encrypted timestamp replies to everyone supplying a public key with their request, and they should be encrypting the md5sum + public key together with their own key. Gazillions of such requests should be stored in multiple mirrored databases that are free to download and duplicate. Of course for a cost, so that it not's spammed, for the cost of keeping a record line in a database, such as 2 cents. The idea here is that you can create a "text", such as a patent document, (or any sequence of 0's and 1's), get the md5sum, get a timestamp certificate for that md5sum from a time authority, the key including the current time and the md5sum, encrypted duplicate with your public key and the time authority's public key. Then you submit this double encrypted md5sum key to a multiple mirrored notary public service, and pay a 2 cents fee. They have no clue what you're notarizing, other than you're doing it. Then you can go apply for a patent, and not fear your patent lawyer will steal your idea and claim he invented it first. Let's suppose someone disputes your claim that you were the first to invent something. Then all you need to do is go to court, and point to a publicly notarized md5sum in the many databases, databases that md5sum themselves hourly then from the hourly md5's compound a daily one, from the daily ones weekly, etc, to make sure nobody can go back and remove a single notary public item. This notary public even only you can decrypt and the time authority should be able to duplicate your effort as a separate excersize, but you alone decrypting it should be proof in case we have to mistrust time authorities. Then you can show that this md5sum is done on such and such text, such and such sequence of 0's and 1's. It would be very hard for your opponent to do a similar excersize on his claim, because it's very hard to create colliding md5sums, etc. You can still be held at gunpoint to provide your own secret part of the public key you used, together with the full document, etc, and then they can go to court and claim that they were the ones obtaining that timestamp. But if you got two, three such timestamps, and only release one of them at gunpoint, how can they tell there is more? They could shoot you, but can't be sure that someone else doesn't know about it? Ideally there should be a way to prove identity in court, such as getting your own social security number mingled in into the md5sum keys, and then even if they extort the information from you, the court sees the notary public line item notarized for a specific social security number, yours. Of course the social security database could be edited too, showing that you don't exist, you have to trust the government not to do that. Basically there is no such thing as security, but there are different degrees of security. In order to misprove your claim, you have to have
As long as we are talking about plastic scissors with rounded tips that are child safe, what about VB classic, or even shell scripts? C++, C#, Java don't really feel like plastic scissors, nor surgeon's knives.
Not to mention low power. Microwatts, nanowatts. Unfortunately it's hard to play Pacman on it, let alone something like Unreal Tournament, but I've seen a paddle/ball video game implementation. If they could take that simplicity, and make it big and fast, where it can drive a screen better, but so simply that even a dumbo can program it, it would really flourish. It needs no harddrive - but you could connect one externally -, it needs no display - but you could connect one externally (my monitor has umm, how many pins? a 40 pin chip should have enough pins) - no keyboard, but you could connect one externally, etc. As far as pincount goes, some of the interfaces could be simplified to how USB2 simplified the how many pin parallel port, and if the feature set is complicated, provide and on-chip mechanism to do the translation (they have RS232 built in to newer chips, but you can do the bit banging by hand if you please, roll your own.) That's how it should be with all peripherals - the specs simple so even a dumbo can do the bit banging, and peripherals should be optional. None of this "drivers" mess in i386 that nobody knows what the heck is going on, and results in crashes so much. And once it reaches the speeds of a Commodore C64 or Z80, it's pretty usable and fun. Yes, I know the Z80 is there, you are welcome to learn it and play away with it, but it's too complex, the motherboard a Z80 runs on is not dumbo friendly. How about instead of Z80's, PIC's with USB stick memory, and some easy LCD display upsided from the 16x2 displays you see, with reduced pincount? You could have a whole range of chips, from superlow microwatts they have now, up to superfast 1GHz P3 level, and in all cases be able to have a standalone chip, with max 40 pins, or the option to connect harddrive, usb stick, display, keyboard, rs232, if you want, that is. Such a standard where things are simple and grandma can bitbang it would work wonders in industry/spaceshuttle/everything, because everyone could do it as opposed to a very specialized few.
"Then it occured to me that the customer simply wanted the impossible done."
Or maybe the customer just wanted a lower price, and they found you. Happens a lot.
How to explain a joke... it always makes it funnier.
Exactly. Moreover practice makes the master, but interesting practice at that, not boring. There is nothing like learning while having fun. Just look at little kittens.
What's with all these Vista news? I guess even Slashdot has become a marketing outlet for Da Man. Who gets to choose these stories?
The links are specially crafted by google to point to digg.com, a new slashdot contender, that points to a lot of videos. Videos/Flash are the ticket to undermining free and open webstandards, that's why they are so heavily pushed, but it'd be hard to and too obvious to push them on slashdot, because everyone would moan too hard here. Why can't there be free video standards everyone adheres to.. well, cuz...
I agree that win nt 3.5/4.0/5.0=approx. 5.1 being massive improvements in stability, but I've still seen them crash, even recently with such things as ASIO. Some things, such as audio drivers, simply have no time to wait around for the windows message system to let them speak up and get some action done. How do you balance stability/security with performance? Yes, win95 was relatively crashy, but very usable. If you tried running a win2k equivalent on a 486DX100 with 8 MB RAM, you may not have had a pleasurable performance - sometimes a crash is a smaller sacrifice than a constant drag and slowness. Note that power consumption was a lot less those days, didn't even need a fan sometimes, just a huge heatsink even on some P200 computers, and you got a lot of bang per buck as far as work per watt electricity consumed went, and it was good enough for such things as word processing, even spreadsheets, and definitely web browsing, but not good enough for multimedia big jpg's, high compression video codecs, etc. At least today the processor market is moving towards variable clockspeed, and low power, where they leave the 486's in the dust compared to idle watts, and processing watts, and easily deal with high performance demands by temporarily kicking up the clockspeed using lots of watts, but still, the OS market with the security paranoia, plus the dotnet "vision" for patentability, needing to run a lot of wasteful "checking" things in the background, hasn't caught up to this performance/watt issue.
This routine is good for precision. In the article he mentions some swede who had to get 48 significant digits for a thermo calculation for a doctorate student. Note that the 2nd iteration in the pdf is orders of magnitude more precise then the first, so all you need is enough iterations. You can always do a (do while current-previous(or a few previous moving average) epsilon), and build a loop, but that has to test a lot of conditionals ready for a branch(deep cpu pipeline predictors don't like branches) as opposed to just lining up 1 or 2 iterations in the code that makes it fly through the pipeline. True, when you do a lot of iterations, the speed gain from getting the first iteration so accurate doesn't count that much. It would save say 2 cycles out of a 20 cycle step, as opposed to making all the difference when there is only one iteration carried out, without a loop, as it is in the sample code given.
I see. So you say the unpredictable part is the assumption of floating point representation in sign, mantissa and exponent terms, which can vary from architecture to architecture.
Yes, but most biological matter is easily degraded by UV radiation, plus infection - bacteria would love to feast on your new solar cells.
As a sidenote, to quote the artilce: "Dr Stephen Curry Opens in new window, a structural biologist from Imperial College London's Division of Cell and Molecular Biology who participated in the research explains: "This work has shown that it is possible to manipulate molecules and proteins that occur naturally in the human body by changing one small detail of their make-up, such as the type of metal at the heart of a porphyrin molecule, as we did in this study.
Naturally occur in a human? I was hoping they'd be talking about cow-derived materials, unless they are interested in genetically engineering photosynthetic human beings? Maybe one of our great great grand children will be engineered enough to be vacuum resistant and fully photosynthetic, then he can fly around in outer space while living off of sunshine.
I think the original constant might have been derived with simple Monte Carlo methods, and the method itself is probably from some book from long time ago, from the 50's when vacuum tube computers first came onto the scene and such speed gains by optimizing binary code to the limit by phd mathematicians was at the forefront of research. To quote the pdf article you cite, "However the analysis should be thoroughly tested on a real machine, since hardware often does not play well with theory." That's nonsense - the machine is predictable, therefore plays well with theory, it's you who get overly engulfed and lost in the derivations by following too many threads of thought and losing track of your assumptions and approximations, it's the theory that's got the problem, and the machine is just an easy and lazy way to double check your thoughts, but you could technically do the same calculations the machine does with a paper and pencil, it would just take longer. I think this article is a bait, to see who will point out subtle issues in the report that might have been directly planted, and such people who dare figure out wat's wrong will instantly win a prize of a dozen shitflies settling down onto them from the NSA/FBI/CIA/MSFT/Dept. of Homeland security, and instantly win a "slavejob-dreamjob" at a fancy place or more like a "fancy prison" of none of their choice, to continue developing their talent and put it to good work for Da Man.
The day the Iraqi war started, someone asked for my opinion. I said, pssh, dumbasses, how are you gonna tell civilians from combatants? He said, just kill'em all. Well, that's your answer. It's not that you can't win, but the cost of winning is too high. And I'd like to think it's humanity holding these jerks back, and not the fact that if they did kill'em all, then that'd be the end of the winners too as far as the rest of the world is concerned, because they can't kill'em all the rest of the world, at least not just yet.
Let me note that I'm less trustful of search engine results than peer contributed wikipedia text additions, as far as trust goes, especially on touchy subjects. This will last as long as there are very few very centralized search engines in the world. I used to use Alltheweb for a while to double check things from other search engines, until yahoo scooped it up in a merger and killed it. And very recently google seems to increasingly point me to articles that require paid access. How come google has paid access? They are supposed to crawl the web like any user could, and read what any user could. I used to come to google with the assumption that a link they gave me had cached text and was accessible to me just like to their crawler, but now it's turning into a marketing front for paid only content. Soon the majority of results might either direct you to paid content only, and crappy free content, without showing you the quality free content, because there is a very serious financial incentive in play, paid for by their paid information sellers giving special access to google crawlers being in bed with the search engine. Where there is a buck to be made by screwing you over how can you have trust? Invisible hand? I don't mind seeing google find matches in journal abstracts that I myself can read freely, and then buy the hidden content, but a lot of their matches show fragments of text from inside the paid content, not the freely available abstract. I think that would be the purpose of an abstract. How can you trust search engines anymore? Wikipedia is not that trustworthy, but is anything else in this world?
Offtopic, but why does your http://everything2.com/index.pl?node_id=735553 everything2 page list last seen 1.4 years ago. One is inclined to think you've been "eliminated." If you're still on the web it'd be nice to put a comment there on "I've been mostly active on slashdot, etc.. not coming here much anymore." Just a suggestion.
Unless the US wants a quick and dirty win by nuking and leveling the rest of the world to the ground, the production capacity of the rest of the world is a significant military asset. You could have made this same argument in say mid 1930's about the Germans having superior military reserves, and making political decisions based on them, but it wasn't until the US production and economy was put in full gear to military service with serious devotion that such statements lost their meanings very quickly. If Japan for instance seriously devoted themselves to producing weapons, within two years you might have better military technology present in the world than anything currently available. But for now, they instead devote their attention to building better cars, or such things as beating speed records in train transportation.
It's not that simple. Those who don't like it don't like it not because they want to know about something, but because they are control freaks, they'd like to sell the same "knowledge" to you, or collect a membership fee, but it's hard when there is a free alternative, free as in both freedom and beer.