What happens when the GPS unit doesn't get a lock on? Cell phones usually do not get a signal inside the big metal building where I work. I think it extremely unlikely that a GPS unit would be able to receive the five (?) different satellites needed for a lock. So does this scheme mean:
1) That locations like this are permanently locked out? There are going to be some extremely unhappy customers...
2) That if the GPS can't get a lock, it goes ahead and works anyhow? Aluminum foil will become a circumvention device...;-)
I have been dealing with maintaining Linux in a mixture of C and Assembler for over twelve years now. While it has been a lot of fun, every day dealing with these essentially dead languages has become harder and harder for me, and other programers seem to agree.
Thus, starting with Linux 3.0 (to be released hopefully by next summer), the kernel will be completely rewritten in the easy-to-use Visual Basic language. This will eliminate all issues involving buffer overruns, as well as streamlining porting of Windows programs to Linux, since Microsoft (who will now assume ownership of Linux) assure me that Windows is written entirely in VB as well.
Microsoft has also stated that they intend to incorporate Windows features, such as the RRS (Rapid Random Shutdown) in Windows 95, into Linux 3.0.
It might take an infinite number of monkeys to produce the internet, but apparently it only takes a finite number of pigeons to rank it... Or to fly over and make it rank.
And you are missing the issue: Is the conduct remedy adequate to stop MSFT from misusing it's monopoly?
"Gateway also faulted another provision of the new licensing agreement, which requires PC makers to pay a Windows royalty on every PC shipped, even if it didn't include Windows. To top it off, to qualify for market development funds [$10/PC], PC makers have to put a Microsoft OS on every PC. As a result, trying to sell non-Windows PCs, or even PCs without software, is a financial loser for computer makers."
That is, if you sell one Linux box, you not only have to pay Microsoft for software that _wasn't_ on it, you also get the price for all the MS licenses you did use raised by $10. MSFT is already committing more crimes. Seems to indicate that a little stiffer supervision is needed, eh?
IIRC, the 6502 took 1 clock to fetch a byte, and the next clock to do something with it. But they overlapped these so most of the time the program ran at 1 clock per byte of program or data. The Z80 could fetch a byte in 3 clocks, but the first byte of an instruction required 5 clocks -- 3 to get the byte, two to decode. And this wasn't overlapped. Averaging out the way this impacted instructions of various length, the Z80 had to be clocked 3 to 4 times as fast to match the 6502. OTOH, when the 6502 was available in 1 MHz only, the Z80 could clocked 8 MHz (twice as fast), if you wanted to pay the premiums for "fast" RAM and ROM... But the real test was in overall system performance as it seemed to the operator -- and the 1MHz 6502 Apple allegedly beat anything else in it's price range. If you needed real power you laid out much more for a fast Z80 system with all the trimmings, and the CPM OS.
Do they alwo use the waste heat from the electric plant to heat the buildings? Just how much electricity does this power plant produce? I know of Michigan colleges that heat their buildings from an electric plant, but there sure isn't enough extra heat to clear the sidewalks when it stays around 10 F for weeks...
Anyhow, note that the conductive concrete won't allow you to use the co-generated "free" heat, but requires the expensive electricity.
The straits connecting the Black Sea and the Mediterranean are the dividing line between Europe and Asia, and Turkey straddles those straits, with it's capital and largest city (Istanbul) on the European side. Geographically, Israel is in Asia, and Turkey is in both Europe and Asia. However, culturally and financially, Israel is definitely European, while Turkey is a mixture of a dozen nationalities, speaking a language from near Mongolia, of Islamic religion but with a culture that owes more to the Byzantine Greeks than the Arabs. And if they make it into the EU (there are some old national enmities they'll have to appease), they won't be the poorest country there.
Note that you can walk from Cairo to Athens, and the biggest river you'll have to cross is the Nile. Until the Suez Canal was built, you could walk from Africa to Asia and not even get your shoes damp. So how did certain points get picked to divide this landmass into three "continents"? It's easy to see the point of dividing Africa from Asia, but when you map the whole thing Europe is just a peninsula sticking out of western asia.
I think it mainly came from the world as viewed from Athens in the 5th Century BCE. Europe was their side of the Hellespont. Asia was the other side of the Hellespont, where those nasty Persians ruled, even though lots of Greeks lived in Anatolia too. (Anatolia is the big peninsula south of the Hellespont-Bosporus straits and the Black Sea.) They had legends about Jason traveling far into the Black Sea, but may not have know for sure that their _was_ a far end to it. I'm not sure if their ships could run down the Asian coast to Syria, Lebanon, and Palestine, or if other naval powers in that area blocked them. But their traders could strike due south and easily reach Egypt, in Africa.
The Egyptians made one contribution to this geography: they knew that there was a narrow neck of land just to their east (Suez), joining land masses too big for them to explore. (Possibly they circumnavigated Africa once, but never bothered with the interior except along the Nile.) So they located the Africa/Asia boundary at that narrow neck. This was also a convenient political division. Nothing in Africa was a threat to Egypt's power. But in Asia, other great powers continually rose and fell (Babylon, Assyria, Syria, Hittites, Persia), and one "Asian" group (the Hyksos -- probably Semites, akin to Hebrews, Arabs, and Phoenicians) even conquered and held Egypt for a century. (They should have paid a bit more attention to those quarrelsome and disunited Greeks, not to mention a little village in Italy called Roma, but hindsight is golden...)
Anyway, the 3 "continents" are based on historical accident as much as geography. By general ties of national descent, language, and customs, Israel is an outlier of Europe, and Turkey has both European and central Asian ties. The Arab lands now stretch from their original homeland (lower Mesopotamia and the adjoining deserts) all across north africa. "Middle East" is just a geographical designation for an area where arbitrarily defined arab nations continually clash with each other as well as the nearby non-arab tribes & nations (Iran, Turkey, Kurds, Armenians, Israel, Afghanistan). Egypt gets grouped in with the Middle East because, even though it's in africa and is defined by ancient natural boundaries, not by lines drawn on the map in a European capital, it often gets into Middle Eastern quarrels. (Meddling in "Asian" affairs is also an Egyptian tradition about 5,000 years old.)
The difference is that Harold's messages were _responses_ to messages from Goliath. If you send me something unsolicited (even items of some value by parcel post), I'm under no obligation to keep track of it. If I send you a message and you answer, I ought to read the answer...
It sounds like the concrete still has a pretty high resistance, so I wouldn't count on a reasonable thickness forming a good Faraday cage. With the conductive stuff costing At 2 or 3 times as much, you could likely get a better cage for less by just hanging metal mesh inside the forms and pouring regular concrete. (The mesh is part of your reinforcement, too.)
And the suggested use of electrically heated payment leaves me wondering where they plan to get free electricity.
But there is one good application for this. Electronics manufacturers need to control static throughout their facilities. Fixed objects are grounded by hooking up wires, but people walk around, circuit boards and parts are carried around on carts, etc., and the only way to ground these while in motion is through the floor. So we paid plenty for conductive tile, and some sort of conductive underlay. If we could have put a conductive layer in the concrete slab itself, it would have saved a bunch (even at 3x the price of regular concrete), and it would be more reliable and lower maintenance.
The article did mention the buried pipe method, but notes that it's expensive to put the pipes in to start with, and then when they split, corrode through, or plug up, you've got to jackhammer through the concrete to fix them. The point about repairs is good, but I do wonder how a grid of buried pipes could be more costly than paying two or three times as much for conductive concrete.
But the real reason heated pavement isn't used much is because electricity costs too much. You don't want to pay the power bill for trying to heat up the freaking outdoors! Conductive concrete won't change that. With buried pipes, you can also heat with a gas or oil furnace, which costs quite a lot less (90% efficient, compared to about 30% overall in the electric system), but it still costs too much in northern climates. And in southern climates where you'd only have to turn on the heat a few days a year, few people think snow is a big enough problem to add thousands to their initial construction cost.
What's private and not depends on the situation. Going into a bar, all they need to know is whether you are of legal age or not. So I would consider recording your name and address to be bogus...
What, the guy can look at the picture to see if it's really your ID, but can't read the birthdate to see if it's before this day in 1981? Even bouncers need _some_ brains.
Which is a pretty good illustration of round-off error in floating point. The first 9 digits are the same; many calculators only carry 8 digits, so they'd probably calculate it out as equal, even though anyone who knows basic number theory can see that it's NOT equal. On my Sharp EL-506A calculator (10 digits), it properly recognized the inequality, but the leading digit of 1922^12 - (1782^12 + 1841^12) was 3 instead of the correct 7. Thirtytwo bit floating point (type "float", on most computers) is either going to have too few bits in the mantissa to detect the difference, or too few bits in the exponent to do the calculation at all.
Of course, what is worse for engineers and programmers is that most calculations that _should_ come out to 0 or equality, don't if you do them in floating point. Numbers derived from real world measurements (including vote counts in the 2000 election) are usually only accurate to 2 or 3 figures, so exact calculations are pointless anyway, but if you forget that the numbers are fuzzy it's real easy to write "if (x==y)" and go nuts figuring out why the "then" branch is never taken. Or worse, to write "while (x!=y)", which will never terminate if the variables are floating point...
It's fairly easy to check Fermat's theorem to finite values of x, y, and n: Say, checking everything up to x, y 2000 and n 15 ought to run in a few hours. (Hint: you cannot use floating point -- so you've got to program multiplication and addition for _extremely_ long integers.)
But how in heck could a computer check this for n=4 and _all_ values of x, y, and z?
OTOH, as the transcript pointed out, you don't need to know how far Fermat's theorem has been tested to see that 1782^12 + 1841^12 = 1922^12 is wrong. Multiplying even numbers by even numbers always gives an even number. The equation is wrong, no arithmetic required. Multiplying odd numbers by odd numbers always gives an odd number. Add even to odd, and you get odd. Make it 1921^12, and we might need a forty-digit calculator to be absolutely _sure_ this wasn't the disproof of Fermat's last theorem...
Ontro's claim is that their heater is safer and cheaper than the various existing methods. It certainly isn't hard to be safer and cheaper than cordite!
Another post claimed that the Ontro process is to mix water with quicklime (CaO). IIRC, mixing pure quicklime with just the right amount of water releases enough heat to reach boiling temperatures. Don't do it in an open container, but in a sealed container at the bottom of the can, it ought to be quite safe. And there are few environmental issues; concentrated fresh slaked lime (what CaO + H2O forms) is alkaline enough to burn the skin, but in air it soon turns to CaC03, which you spread on your garden. It ought to be good for a landfill to have slaked lime slowly leaking out of the cans.
You don't want to eat CaO or slaked lime straight (way too alkaline), but diluted traces won't hurt you or anything else.
On exposure to air, quicklime and slaked lime absorb carbon dioxide and turn into calcium carbonate. AKA limestone. You spread it on your garden. Also, mortar for bricklaying is slaked lime + sand + (maybe) portland cement. I once spent a whole summer with my hands covered with mortar. Carrying those heavy, rough concrete blocks around hurt my skin, but the lime didn't.
The Pelamis device is substantially different from other wave systems I have seen. The usual trouble with wave power is that you put a lot of expensive equipment out in the way of the waves, and then a storm comes and the waves get too big and destroy it. Even Lake Michigan gets storms that will re-arrange boulders two yards in diameter; the waters off Scotland or Vancouver are far worse. But there is a lot of power there, so Pelamis designed for survivability in severe storms.
There's something weird about their website, so I cannot give you the URL to go straight to the how-it-works pdf. Navigate to Downloads, and open the bottom one: "'Water Power' magazine article". It's written by someone who never uses one short word where 4 long ones will do, so you might prefer my description:
The Pelamis generator is snake-shaped, made of many rigid steel cylinders jointed together, and floating on the surface. The head end is anchored and the snake swivels around it to keep the head into the waves. As the waves pass, it bends in the vertical plane, roughly following the shape of the waves. Each joint is attached to a hydraulic cylinder, so the bending pumps hydraulic fluid into an accumulator (pressure tank). Fluid from the accumulator runs a hydraulic motor to turn an alternator.
There are ways to tune the system response so it resonates with small waves to extract more power in relatively calm conditions, but as the waves get bigger it goes out of resonance so the energy extracted doesn't become more than the system is designed to handle. In a bad storm, it gives minimal opposition to the waves, so it doesn't get bashed like fixed installations. The weakest part is probably the anchor -- if that drags, the snake could get lost at sea or smashed into the rocks. This is roughly the same chance a ship at anchor runs, except that the snake is a much smaller cross section and so gets less drag, and also you can do things to secure the anchor like pouring concrete that ships don't do because they want the anchor back. OTOH, you don't anchor your ship out where the waves are biggest...
At an absolute minimum, they are stealing from the user's bandwidth -- because to go to Morpheus and then to Amazon (say) is certainly going to take more bytes transferred, and more time, than going directly.
1) Parent, nosy churchlady, or someone who couldn't pass the tests to become a postal inspector finds something on the web they don't like. They write a letter to the AG.
2) Nearly all the real kiddie porn will be gone within hours. So how is the AG going to collect evidence to go before the court and ask for an order to close it down?
3) Probably the AG has political ambitions, so he'll still try to find _something_ to block. Maybe purveyors of "barely legal" pictures. Maybe a URL that repeatedly gets complaints, even though there's nothing there when they look. Maybe Planned Parenthood sites; because these stay put, they'll probably log more complaints from the religious kooks than any actual porn site....
4) Compliant judge will sign the orders without actually looking at the "evidence".
5) Hundreds of lawsuits will be filed for violations of civil rights.
Folks, the 1st Amendment does not prohibit censorship by private parties of items passing through their servers. It does prohibit government censorship (with exceptions that I seem to be unable to find in the actual text)... By designating the sites to be blocked, the State of PA is putting itself right in the targets of every hungry lawyer that can find an innocent, or sleazy but legal, client on the block list.
333 hours/month = 11 hours a day (including Christmas), or 16 hours per working day. No good for a server, adequate for most desktops.
However, there are two big issues here:
1) IBM sure didn't do much to make this limitation obvious as they were selling the drives. In fact, it's not at all clear that they posted this limitation before the first GXP's were sold, and according to VIA they have not _yet_ posted this limitation in the 75GXP data sheets at all. And if "60", "75", and "120" mean GB, then where in hell did IBM expect them to go but servers?
2) AFAIK, it's a nonsensical limitation. Most drives last longer running 24/7 than being turned on and off daily. It sounds like just a ploy to blame their bad design or workmanship on the customers -- "You ran it too much". Or does IBM know the bearings have a predictable and rather short wearout, and figures that 333 hrs/month will get you to that point in some fixed period (end warranty?). If that was known, building them at all indicates gross misunderstanding (or not caring) of what customers expect from their hardware, and ever selling them without being clearly marked "NOT FOR SERVERS, LIMITED OPERATING TIME" is grossly irresponsible, at least.
However, anecdotally it doesn't sound like a predictable wearout exists -- many of the reported failures were premature by anyone's measure, unless they were operated in a time warp for about 72 hours a day. Which brings me back to the first theory -- put bogus and unreasonable limitations in your data sheet so you can blame the customers for failures, even though the running hours probably weren't why they failed...
huh? 'They all' being the metrics? Or the programmers?
If you don't know which programmers are producing, eventually most of those remaining WILL be useless. And in the present litigious environment, when it's time to fire the deadwood, it does help if you have numbers rather than a subjective evaluation -- even though if you are clueful, your subjective evaluation is probably more accurate than any metric...
Amen. We once had one of those "indispensable" progammers here. He went on vacation and never returned, just called back to say he'd found a better job. (And then the guy he talked to forgot to tell anyone else, but that's another story...) So a few months later, one constant used in just one place in a C program he wrote had to be changed. It took 2 days to find the spot in the dozens of.c and.h files, 3 seconds to change it, 1 day to figure out how to compile (he didn't archive the Make file, either), and an hour to test it.
Proper design documents should have clued me in that I'd find this feature handled in this module, and cut that down to 2 or 3 hours all total. But there were no design docs, and the only comments were something like/* Author: Ash Ole, 1991 */
KLOC in any form is a useless metric. Duh!!! The problem is that the boss _has_ to have some metric, and so will seize on a bad one if no good one is available. And neither Charles' post, nor yours, offer anything useful. "Time until feature complete" is not useful per se, because features come in all different sizes. Figure out a way to score features by complexity, and you've got a starting point.
Charles' post does have a "born yesterday" quality to it. Some smart people have been struggling with these issues for 50 years. There are various scoring systems, for instance function points. (Just search Google.) They are rather subjective, but if used honestly (not warping the scores to make your estimate fit a budget or time target), they are useful in estimating software cost...
But now you have to take into account the laws of management behavior. Function points appear subjective, and take thinking to evaluate. Also, FP is a rather large measure (1 per man per week is about average), and because an FP is just part of a completed feature it's pretty hard to tell how many FP's were completed on the weekly reports. (And upper management MUST have those weekly reports to see if the project is still on schedule, and whether they are going to have to change all the marketing material from "XYZ 2002" to "XYZ 2003".) LOC might be meaningless, but everyone counts them the same, and if you're a micromanager you can count them every time someone checks in a piece of code and track "progress" on a daily, maybe even hourly, basis. That's if you believe that the project will actually require 800 KLOC exactly, and therefore when 600 KLOC have been checked in, you are 3/4 done. Never mind that the last 100KLOC are going to take longer than the first 700KLOC, and then the project still won't be done, those hard numbers were very comforting right up until the moment reality crashed in...
Slashdot Mirror
What happens when the GPS unit doesn't get a lock on? Cell phones usually do not get a signal inside the big metal building where I work. I think it extremely unlikely that a GPS unit would be able to receive the five (?) different satellites needed for a lock. So does this scheme mean:
;-)
1) That locations like this are permanently locked out? There are going to be some extremely unhappy customers...
2) That if the GPS can't get a lock, it goes ahead and works anyhow? Aluminum foil will become a circumvention device...
on what basis do you claim that her conclusions about Clipper are "stupid"?
Thinking that we'll trust the government not to misuse crypto keys is very, very stupid.
It might take an infinite number of monkeys to produce the internet, but apparently it only takes a finite number of pigeons to rank it...
Or to fly over and make it rank.
I think the idea is to ensure that any abductor would not be able to remove it.
Bolt cutters.
And you are missing the issue: Is the conduct remedy adequate to stop MSFT from misusing it's monopoly?
"Gateway also faulted another provision of the new licensing agreement, which requires PC makers to pay a Windows royalty on every PC shipped, even if it didn't include Windows. To top it off, to qualify for market development funds [$10/PC], PC makers have to put a Microsoft OS on every PC. As a result, trying to sell non-Windows PCs, or even PCs without software, is a financial loser for computer makers."
That is, if you sell one Linux box, you not only have to pay Microsoft for software that _wasn't_ on it, you also get the price for all the MS licenses you did use raised by $10. MSFT is already committing more crimes. Seems to indicate that a little stiffer supervision is needed, eh?
Gee, the 4004 must be _really_ fast then. (4-bit micro...)
IIRC, the 6502 took 1 clock to fetch a byte, and the next clock to do something with it. But they overlapped these so most of the time the program ran at 1 clock per byte of program or data. The Z80 could fetch a byte in 3 clocks, but the first byte of an instruction required 5 clocks -- 3 to get the byte, two to decode. And this wasn't overlapped. Averaging out the way this impacted instructions of various length, the Z80 had to be clocked 3 to 4 times as fast to match the 6502. OTOH, when the 6502 was available in 1 MHz only, the Z80 could clocked 8 MHz (twice as fast), if you wanted to pay the premiums for "fast" RAM and ROM... But the real test was in overall system performance as it seemed to the operator -- and the 1MHz 6502 Apple allegedly beat anything else in it's price range. If you needed real power you laid out much more for a fast Z80 system with all the trimmings, and the CPM OS.
Do they alwo use the waste heat from the electric plant to heat the buildings? Just how much electricity does this power plant produce? I know of Michigan colleges that heat their buildings from an electric plant, but there sure isn't enough extra heat to clear the sidewalks when it stays around 10 F for weeks...
Anyhow, note that the conductive concrete won't allow you to use the co-generated "free" heat, but requires the expensive electricity.
The straits connecting the Black Sea and the Mediterranean are the dividing line between Europe and Asia, and Turkey straddles those straits, with it's capital and largest city (Istanbul) on the European side. Geographically, Israel is in Asia, and Turkey is in both Europe and Asia. However, culturally and financially, Israel is definitely European, while Turkey is a mixture of a dozen nationalities, speaking a language from near Mongolia, of Islamic religion but with a culture that owes more to the Byzantine Greeks than the Arabs. And if they make it into the EU (there are some old national enmities they'll have to appease), they won't be the poorest country there.
Note that you can walk from Cairo to Athens, and the biggest river you'll have to cross is the Nile. Until the Suez Canal was built, you could walk from Africa to Asia and not even get your shoes damp. So how did certain points get picked to divide this landmass into three "continents"? It's easy to see the point of dividing Africa from Asia, but when you map the whole thing Europe is just a peninsula sticking out of western asia.
I think it mainly came from the world as viewed from Athens in the 5th Century BCE. Europe was their side of the Hellespont. Asia was the other side of the Hellespont, where those nasty Persians ruled, even though lots of Greeks lived in Anatolia too. (Anatolia is the big peninsula south of the Hellespont-Bosporus straits and the Black Sea.) They had legends about Jason traveling far into the Black Sea, but may not have know for sure that their _was_ a far end to it. I'm not sure if their ships could run down the Asian coast to Syria, Lebanon, and Palestine, or if other naval powers in that area blocked them. But their traders could strike due south and easily reach Egypt, in Africa.
The Egyptians made one contribution to this geography: they knew that there was a narrow neck of land just to their east (Suez), joining land masses too big for them to explore. (Possibly they circumnavigated Africa once, but never bothered with the interior except along the Nile.) So they located the Africa/Asia boundary at that narrow neck. This was also a convenient political division. Nothing in Africa was a threat to Egypt's power. But in Asia, other great powers continually rose and fell (Babylon, Assyria, Syria, Hittites, Persia), and one "Asian" group (the Hyksos -- probably Semites, akin to Hebrews, Arabs, and Phoenicians) even conquered and held Egypt for a century. (They should have paid a bit more attention to those quarrelsome and disunited Greeks, not to mention a little village in Italy called Roma, but hindsight is golden...)
Anyway, the 3 "continents" are based on historical accident as much as geography. By general ties of national descent, language, and customs, Israel is an outlier of Europe, and Turkey has both European and central Asian ties. The Arab lands now stretch from their original homeland (lower Mesopotamia and the adjoining deserts) all across north africa. "Middle East" is just a geographical designation for an area where arbitrarily defined arab nations continually clash with each other as well as the nearby non-arab tribes & nations (Iran, Turkey, Kurds, Armenians, Israel, Afghanistan). Egypt gets grouped in with the Middle East because, even though it's in africa and is defined by ancient natural boundaries, not by lines drawn on the map in a European capital, it often gets into Middle Eastern quarrels. (Meddling in "Asian" affairs is also an Egyptian tradition about 5,000 years old.)
The difference is that Harold's messages were _responses_ to messages from Goliath. If you send me something unsolicited (even items of some value by parcel post), I'm under no obligation to keep track of it. If I send you a message and you answer, I ought to read the answer...
It sounds like the concrete still has a pretty high resistance, so I wouldn't count on a reasonable thickness forming a good Faraday cage. With the conductive stuff costing At 2 or 3 times as much, you could likely get a better cage for less by just hanging metal mesh inside the forms and pouring regular concrete. (The mesh is part of your reinforcement, too.)
And the suggested use of electrically heated payment leaves me wondering where they plan to get free electricity.
But there is one good application for this. Electronics manufacturers need to control static throughout their facilities. Fixed objects are grounded by hooking up wires, but people walk around, circuit boards and parts are carried around on carts, etc., and the only way to ground these while in motion is through the floor. So we paid plenty for conductive tile, and some sort of conductive underlay. If we could have put a conductive layer in the concrete slab itself, it would have saved a bunch (even at 3x the price of regular concrete), and it would be more reliable and lower maintenance.
The article did mention the buried pipe method, but notes that it's expensive to put the pipes in to start with, and then when they split, corrode through, or plug up, you've got to jackhammer through the concrete to fix them. The point about repairs is good, but I do wonder how a grid of buried pipes could be more costly than paying two or three times as much for conductive concrete.
But the real reason heated pavement isn't used much is because electricity costs too much. You don't want to pay the power bill for trying to heat up the freaking outdoors! Conductive concrete won't change that. With buried pipes, you can also heat with a gas or oil furnace, which costs quite a lot less (90% efficient, compared to about 30% overall in the electric system), but it still costs too much in northern climates. And in southern climates where you'd only have to turn on the heat a few days a year, few people think snow is a big enough problem to add thousands to their initial construction cost.
What's private and not depends on the situation. Going into a bar, all they need to know is whether you are of legal age or not. So I would consider recording your name and address to be bogus...
What, the guy can look at the picture to see if it's really your ID, but can't read the birthdate to see if it's before this day in 1981? Even bouncers need _some_ brains.
Which is a pretty good illustration of round-off error in floating point. The first 9 digits are the same; many calculators only carry 8 digits, so they'd probably calculate it out as equal, even though anyone who knows basic number theory can see that it's NOT equal. On my Sharp EL-506A calculator (10 digits), it properly recognized the inequality, but the leading digit of 1922^12 - (1782^12 + 1841^12) was 3 instead of the correct 7. Thirtytwo bit floating point (type "float", on most computers) is either going to have too few bits in the mantissa to detect the difference, or too few bits in the exponent to do the calculation at all.
Of course, what is worse for engineers and programmers is that most calculations that _should_ come out to 0 or equality, don't if you do them in floating point. Numbers derived from real world measurements (including vote counts in the 2000 election) are usually only accurate to 2 or 3 figures, so exact calculations are pointless anyway, but if you forget that the numbers are fuzzy it's real easy to write "if (x==y)" and go nuts figuring out why the "then" branch is never taken. Or worse, to write "while (x!=y)", which will never terminate if the variables are floating point...
It's fairly easy to check Fermat's theorem to finite values of x, y, and n: Say, checking everything up to x, y 2000 and n 15 ought to run in a few hours. (Hint: you cannot use floating point -- so you've got to program multiplication and addition for _extremely_ long integers.)
But how in heck could a computer check this for n=4 and _all_ values of x, y, and z?
OTOH, as the transcript pointed out, you don't need to know how far Fermat's theorem has been tested to see that 1782^12 + 1841^12 = 1922^12 is wrong. Multiplying even numbers by even numbers always gives an even number. The equation is wrong, no arithmetic required. Multiplying odd numbers by odd numbers always gives an odd number. Add even to odd, and you get odd. Make it 1921^12, and we might need a forty-digit calculator to be absolutely _sure_ this wasn't the disproof of Fermat's last theorem...
Ontro's claim is that their heater is safer and cheaper than the various existing methods. It certainly isn't hard to be safer and cheaper than cordite!
Another post claimed that the Ontro process is to mix water with quicklime (CaO). IIRC, mixing pure quicklime with just the right amount of water releases enough heat to reach boiling temperatures. Don't do it in an open container, but in a sealed container at the bottom of the can, it ought to be quite safe. And there are few environmental issues; concentrated fresh slaked lime (what CaO + H2O forms) is alkaline enough to burn the skin, but in air it soon turns to CaC03, which you spread on your garden. It ought to be good for a landfill to have slaked lime slowly leaking out of the cans.
You don't want to eat CaO or slaked lime straight (way too alkaline), but diluted traces won't hurt you or anything else.
On exposure to air, quicklime and slaked lime absorb carbon dioxide and turn into calcium carbonate. AKA limestone. You spread it on your garden. Also, mortar for bricklaying is slaked lime + sand + (maybe) portland cement. I once spent a whole summer with my hands covered with mortar. Carrying those heavy, rough concrete blocks around hurt my skin, but the lime didn't.
The Pelamis device is substantially different from other wave systems I have seen. The usual trouble with wave power is that you put a lot of expensive equipment out in the way of the waves, and then a storm comes and the waves get too big and destroy it. Even Lake Michigan gets storms that will re-arrange boulders two yards in diameter; the waters off Scotland or Vancouver are far worse. But there is a lot of power there, so Pelamis designed for survivability in severe storms.
There's something weird about their website, so I cannot give you the URL to go straight to the how-it-works pdf. Navigate to Downloads, and open the bottom one: "'Water Power' magazine article". It's written by someone who never uses one short word where 4 long ones will do, so you might prefer my description:
The Pelamis generator is snake-shaped, made of many rigid steel cylinders jointed together, and floating on the surface. The head end is anchored and the snake swivels around it to keep the head into the waves. As the waves pass, it bends in the vertical plane, roughly following the shape of the waves. Each joint is attached to a hydraulic cylinder, so the bending pumps hydraulic fluid into an accumulator (pressure tank). Fluid from the accumulator runs a hydraulic motor to turn an alternator.
There are ways to tune the system response so it resonates with small waves to extract more power in relatively calm conditions, but as the waves get bigger it goes out of resonance so the energy extracted doesn't become more than the system is designed to handle. In a bad storm, it gives minimal opposition to the waves, so it doesn't get bashed like fixed installations. The weakest part is probably the anchor -- if that drags, the snake could get lost at sea or smashed into the rocks. This is roughly the same chance a ship at anchor runs, except that the snake is a much smaller cross section and so gets less drag, and also you can do things to secure the anchor like pouring concrete that ships don't do because they want the anchor back. OTOH, you don't anchor your ship out where the waves are biggest...
At an absolute minimum, they are stealing from the user's bandwidth -- because to go to Morpheus and then to Amazon (say) is certainly going to take more bytes transferred, and more time, than going directly.
Let's see, how is this going to play out:
1) Parent, nosy churchlady, or someone who couldn't pass the tests to become a postal inspector finds something on the web they don't like. They write a letter to the AG.
2) Nearly all the real kiddie porn will be gone within hours. So how is the AG going to collect evidence to go before the court and ask for an order to close it down?
3) Probably the AG has political ambitions, so he'll still try to find _something_ to block. Maybe purveyors of "barely legal" pictures. Maybe a URL that repeatedly gets complaints, even though there's nothing there when they look. Maybe Planned Parenthood sites; because these stay put, they'll probably log more complaints from the religious kooks than any actual porn site....
4) Compliant judge will sign the orders without actually looking at the "evidence".
5) Hundreds of lawsuits will be filed for violations of civil rights.
Folks, the 1st Amendment does not prohibit censorship by private parties of items passing through their servers. It does prohibit government censorship (with exceptions that I seem to be unable to find in the actual text)... By designating the sites to be blocked, the State of PA is putting itself right in the targets of every hungry lawyer that can find an innocent, or sleazy but legal, client on the block list.
333 hours/month = 11 hours a day (including Christmas), or 16 hours per working day. No good for a server, adequate for most desktops.
However, there are two big issues here:
1) IBM sure didn't do much to make this limitation obvious as they were selling the drives. In fact, it's not at all clear that they posted this limitation before the first GXP's were sold, and according to VIA they have not _yet_ posted this limitation in the 75GXP data sheets at all. And if "60", "75", and "120" mean GB, then where in hell did IBM expect them to go but servers?
2) AFAIK, it's a nonsensical limitation. Most drives last longer running 24/7 than being turned on and off daily. It sounds like just a ploy to blame their bad design or workmanship on the customers -- "You ran it too much". Or does IBM know the bearings have a predictable and rather short wearout, and figures that 333 hrs/month will get you to that point in some fixed period (end warranty?). If that was known, building them at all indicates gross misunderstanding (or not caring) of what customers expect from their hardware, and ever selling them without being clearly marked "NOT FOR SERVERS, LIMITED OPERATING TIME" is grossly irresponsible, at least.
However, anecdotally it doesn't sound like a predictable wearout exists -- many of the reported failures were premature by anyone's measure, unless they were operated in a time warp for about 72 hours a day. Which brings me back to the first theory -- put bogus and unreasonable limitations in your data sheet so you can blame the customers for failures, even though the running hours probably weren't why they failed...
Amen. We once had one of those "indispensable" progammers here. He went on vacation and never returned, just called back to say he'd found a better job. (And then the guy he talked to forgot to tell anyone else, but that's another story...) So a few months later, one constant used in just one place in a C program he wrote had to be changed. It took 2 days to find the spot in the dozens of .c and .h files, 3 seconds to change it, 1 day to figure out how to compile (he didn't archive the Make file, either), and an hour to test it.
/* Author: Ash Ole, 1991 */
Proper design documents should have clued me in that I'd find this feature handled in this module, and cut that down to 2 or 3 hours all total. But there were no design docs, and the only comments were something like
KLOC in any form is a useless metric. Duh!!! The problem is that the boss _has_ to have some metric, and so will seize on a bad one if no good one is available. And neither Charles' post, nor yours, offer anything useful. "Time until feature complete" is not useful per se, because features come in all different sizes. Figure out a way to score features by complexity, and you've got a starting point.
Charles' post does have a "born yesterday" quality to it. Some smart people have been struggling with these issues for 50 years. There are various scoring systems, for instance function points. (Just search Google.) They are rather subjective, but if used honestly (not warping the scores to make your estimate fit a budget or time target), they are useful in estimating software cost...
But now you have to take into account the laws of management behavior. Function points appear subjective, and take thinking to evaluate. Also, FP is a rather large measure (1 per man per week is about average), and because an FP is just part of a completed feature it's pretty hard to tell how many FP's were completed on the weekly reports. (And upper management MUST have those weekly reports to see if the project is still on schedule, and whether they are going to have to change all the marketing material from "XYZ 2002" to "XYZ 2003".) LOC might be meaningless, but everyone counts them the same, and if you're a micromanager you can count them every time someone checks in a piece of code and track "progress" on a daily, maybe even hourly, basis. That's if you believe that the project will actually require 800 KLOC exactly, and therefore when 600 KLOC have been checked in, you are 3/4 done. Never mind that the last 100KLOC are going to take longer than the first 700KLOC, and then the project still won't be done, those hard numbers were very comforting right up until the moment reality crashed in...