The article did state that it was weighted towards the older albums. Mainly, he was trying to identify the groups that have sold platinums year after year. If you look at just the leading sellers in one year, in music, movies, or books -- most of the time, five years later it's "what were we thinking." If the band broke up about 30 years ago and two of them are dead, but their recordings are still selling, then it's likely there was something like real quality...
If I were to use a modified Linux kernel as the operating system for a VCR, plus an application program compiled with gnu C and using an LGPL'd library, then
1) I would have to send Linus a copy of the kernel source code.
2) I would have to provide the purchasers of the VCR some way to get the kernel source code.
3) The kernel would be GPL.
4) The application could be proprietary.
The main problem I can see in this (aside from execs who believe M$ FUD), is that if we made extensive hardware-related changes to the kernel, being able to download the source code would make it easier for our competitors to reverse-engineer the product. However, in most cases this could be avoided by putting the hardware dependent stuff in the proprietary modules, and doing kernel changes only to trim out the unused pieces. (e.g., no disk drive, so delete all that...)
You don't have to declare variables and their type before use. Sometimes that is good, sometimes it is bad. When I'm just playing around with a little code to see if something will work, I'd rather not have to go back to the top and insert a declaration every time I find I need a variable for a for loop or something. When I was first learning to program 30 years ago, as best I can remember FORTRAN didn't even _allow_ variable declarations except for arrays, and I think that worked out fairly well -- one less thing to worry about on the first few programs. The old unstructured BASIC was even better IMHO, someone with no idea at all of what they were doing could plunge in and start learning by writing programs that often worked. If Python will allow that approach to learning structured programming, even better. Many Computer Science professors will disagree about that ("they don't learn to plan ahead and map out their data structures"), but in my experience, no one can plan ahead _until_ they've written enough programs to get some idea of what they are doing.
But when I write a serious program that has to work right even under conditions I didn't quite anticipate, I certainly want to list all the variables ahead of time, and to have the compiler tell me when I misspell one or use the wrong type. So being able to choose whether or not definitions are mandatory is ideal. Sounds like Python does that.
pick any J. Random Licensing Agreement and you'll see words to the effect of "We promise nothing. We accept no responsibility. Don't be surprised if this software is utterly unfit for anything, including the task for which you are paying the license fee. Anything that happens is on your head. Now pay up, bitch."
You forgot: "You are not allowed to fix the defects in this program yourself. You may pay us extra for tech support, but we do not guarantee that even that will make it work, or that our tech support people will tell you when your problem comes from a known bug that affects many people. We may sue if you benchmark the program without our permission, because we don't want you publicizing the bugs. When we issue a bug fix, we may make you buy the program again to get it. We don't guarantee that the next version of the program (which you will pay for) will fix the known bugs."
This may be too complicated for pinhead executives to understand. Explain that it's like buying a used car "as-is" with the hood welded shut and a contract that does not allow you to fix it.
Virginia _did_ pass the UCITA, making it about the last place in the USA you would want to litigate a free software case. The Python license had a forum-selection clause specifying that litigation concerning it would be in VA; this is obviously a bad idea, but besides that the FSF doesn't allow any forum-selection clause in GPL's.
But there apparently were a couple of other issues which I don't understand -- there's a list of changes, but the only one explained in the posted letter from the FSF was eliminatiing the forum selection clause, and it's going to take more time than I am willing to put in even to find the original language.
If they can't use it in court, why are they collecting it? To post on the internet? Under J. Edgar Hoover, the FBI collected lots of data that was useless in court, but maybe could be used for politicking, blackmail, or just gave JEH a kick to read. Allegedly his position and the FBI budget were untouchable for 40 years because his secret files contained scandalous info on so many politicians. And in the 60's, he was always trying to get some newspaper to print his files on Martin Luther King, but they didn't consider the reverend's affairs newsworthy. If he'd had the internet so he could publish them anonymously himself...
They claim the FBI has changed. I'm not convinced. At best, they still seem to prefer good publicity to good police work. At worst, they seem to tolerate and cover up their own incompetence (Ruby Ridge, Waco, the FBI labs). Let them secretly troll Congressional e-mail for blackmail material, and they could become completely untouchable...
Right and wrong. The Ninth Amendment: The enumeration in the Constitution, of certain rights, shall not be
construed to deny or disparage others retained by the people. It sounds like an invitation for the Supreme court to go discover new rights, but as far as I know they didn't until the 1950's in a case involving a Massachusetts law against contraceptives. The justices decided that this law violated some unenumerated right, which they called the "right to privacy", but which is more accurately described as "the right to control your own body." That is, they've applied it to contraceptives and abortion, but in regards to wiretapping and other snooping, their decisions have been based on the 4th Amendment (Unreasonable search and seizure), not the 9th.
The application of the 4th Amendment when only electrons are "seized" has always made some strict constructionists uncomfortable. So it is possible that with a few more justices who are unwilling to go beyond the actual words in the Constitution, the gov't would be free to wiretap at will, look through the walls of your house with infrared, and order fat people to go on a diet. So I would like to see two amendments specifically codifying a true right to privacy (no snooping without a warrant), and the right to control your own body (which would also invalidate drug laws and many of the FDA's powers).
The incandescent light bulb (with a filament of charred linen) was simultaneously invented by Edison's lab and by some British high-school teacher. They sued each other over patents, and each one won -- in his own country. However, maybe the real credit ought to go to whoever invented a vacuum pump good enough to keep the filament from burning up in residual oxygen. A frenchman (I think) had the basic idea 10 years earlier, but his vacuum wasn't good enough and the bulb died in minutes. The bulbs we use now have a tungsten filament (for a 10x improvement in life-expectancy), sometimes nitrogen gas fill instead of a vacuum, or even a bit of bromine and iodine to combine with evaporating tungsten atoms and re-deposit them on the filament ("halogen" bulbs); I remember that last invention was made by aircraft engineers trying to shrink the size of the wingtip running lights, but I have no idea who made the others.
There was also the electric arc lamp, which was around decades earlier than the incandescent bulb. But in the 19th century it resembled an arc-welder, and someone had to stay there and keep adjusting it as the electrodes burned down. Many decades later, it was tamed by being built into a bulb with automatic control circuits -- neon, mercury arc, and flourescent lights.
I never heard of that airplane in New Zealand. Are there more details on the web? It's possible that simple remoteness cheated the Kiwis out of proper recognition. But it's quite clear why the Wright's got the recognition out of all the Europeans and Americans who were riding wobbly powered kites up in the air at that time: they managed to crash the plane gently enough to be able to repair it and go up again the next day, and to keep it in one piece while they taught themselves how to _land_. Lilienthal (at least) was flying several times before the Wrights, but he always came down hard. He apparently assumed that when he got the airframe right, he could just automatically fly it, while the Wrights did everything possible to prepare themselves and then took baby steps. Even with all that preparation, the first flight ended in a tailspin -- but at 12 feet and maybe 20mph it didn't break the airplane much.
Er... How Tenochtitlan ever managed to import enough food to survive is hard to explain (unless the Aztec army marched slaves into the city as a food supply, but as far as I can recall that was about the only conceivable evil the Spanish didn't accuse the Aztecs of). The city probably was in the middle of extremely productive fields (they had water, and deserts usually are extremely productive when you first start irrigation), but a meat supply would have been darned difficult and the Aztecs don't seem likely to be vegetarians. Did runners wear themselves out bringing in bits of fresh meat for the king, while everyone else lived on corn, beans, and a little jerky?
One thing is that unlike European cities, Tenochtitlan did not have to be self-supporting economically. European cities traded manufactured and imported goods for the food they needed. The Aztec capital apparently depended on a harsh system of conquest, taxation, and slavery. There may have been more slaves carrying dried food in on their backs than there were people living in the city, plus vast numbers of tributary tribesmen (not quite slaves if they behaved themselves, but close enough) growing or catching and drying that food. Then in the city, a lot of slaves who probably got only corn and beans and eventually died of deficiency diseases, a number of nobles who visited for a few days then went back to the country where they could get fresh food, and the kings inner circle and personal guard who got the best of the local garden produce and occasionally a relatively fresh haunch of deer brought in by relays of runners. Maybe they imported dogs for the king's greatest feasts. In any case, the cost of supporting that city must have been a great burden upon the countryside, and the tributary tribes were ripe for revolt when Cortez showed up.
Humankind: the only species capable, and dumb enough to wipe out... themselves. Not true. Bacteria will wipe themselves out quite well in the right conditions. You fill a dish with rich bacteria food, drop in a few bacteria, keep it warm and wait. Within days (or maybe hours) there will be millions of dead bacteria, poisoned by their own wastes. What _is_ disheartening is that, as smart as one human is, a million humans together are usually no smarter than bacteria...
It's not like ancient man had the tools nessecery to kill off a sufficient quantity of any animal as to drive it to extinction. Yes they did, at least for herd herbivores (horses, camels, mammoths). You follow the herd until it is near a cliff, then the whole tribe comes out screaming and waving torches or whatever, and scare the whole herd over the cliff. Everyone eats until they puke for a few days, then when the remaining 99% of the meat smells too bad to eat you go looking for the next herd. Of course, you do have to live off of roots and bugs for a week or two until the next opportunity for a herd kill arises, but on the average you are well fed.
I think the bison ("buffalo") avoided going with the rest of the large North American herbivores because they mostly live where there aren't enough roots, bugs, and field mice to keep a man alive until the next bison herd comes along. Indians living on the fringes of the great plains where there is barely sufficient sustenance only bagged one or two buffalo herds a year before Europeans brought back the horse, and they did it by driving them over a cliff.
The large predators that also died when man moved into North America are more of a mystery, but it is not inexplicable. The large animals they depended on for most of their meat became scarce. Strange new animals were around and looked like easy prey teetering on their hind legs, but few of those who tried eating them lived to learn better. And in the end, "trophy hunting" may have become a factor; grandpa keeps showing off his necklace with dozens of saber-tooth tiger fangs, and you really want some, but he won't share so you go hunt down the last saber-tooth in the area...
The basic difference is: the early americans ate all the horses. Someplace near present-day Ukrainia, some of my ancestors learned to ride them instead, and the horse-riding peoples spread over Eurasia in waves of conquest (Celts, Hellenes, Huns, Goths, Normans, Mongols... Similarly, tameable breeds of camels, cows, sheep, goats, and donkeys were domesticated in Eurasia or Africa, but eaten in North America. The result 13,000 years later was that the europeans easily overwhelmed the native americans and took their land. Horses and a wider variety of farmable plants helped, but maybe the biggest difference was that with no tame meat animal except the dog, Indian towns had to stay small enough to allow farms nearby to grow the corn needed and the men to _walk_ to their hunting grounds, while much larger European cities were fed by cattle driven into town and grain hauled in by horse-drawn wagons. City life gave Europeans metal, gunpowder, government, large armies, measles, and smallpox, and all the Indians had to counter with was skill in the woods.
So when we drive another hundred species to extinction today, just what possible uses might we have missed?
FORTRAN does fit pure number-crunching applications (science and engineering) fairly well. The syntax and (since 1977) program structure are a pretty good match to specifications written in math, and the compiled executables are _fast_. There are thousands of libraries for mathematical and statistical applications. Finally, it's easier to write a compiler that will automatically divide up the work among multiple CPU's in FORTRAN than in most procedural languages. (This is primarily due to a lack of those features that make C/C++ so flexible but also allow all sorts of non-obvious side effects.)
So if you are going to write programs that will tie up the FPU for hours, days, weeks, or even months, and don't need a user interface or care much what the printout looks like as long as the numbers are accurate, then FORTRAN is the right tool (assuming the job is too big for math worksheet tools like MATLAB). If you need a decent user interface, or have to produce nice looking reports for managers, then using FORTRAN is like building a car body with a hammer and anvil. I did have to do a short hardware-driver type of program in FORTRAN back in engineering school (it was the only compiler the obsolete minicomputer in that lab had), but if there'd been an _assembler_ available it would have been easier. I knew of companies that did all their programming in FORTRAN through the 60's and 70's, from corporate accounting to machine control programs, but it sure wasn't a smart choice.
I haven't looked at Python hard enough yet, but at first glance it does look like a pretty good candidate for the first language. One of the big questions here is just what should be taught in the first course: how computers work (C or even assembly), or how to program in a clean and easy language (Python or some other very high level OO language). I'd get them programming first, then do the low-level stuff in the second term, then cover the languages you'll need on the job the second and subsequent years (C++/Java and database programming). C++ is just too darned hard for the first programming course. Java is better, but I think (I haven't used it) there are still too many traps for novices.
By the way, I've been wondering if my grandson (a very smart and hyper 3rd grader) could get started on programming this summer. 10 years ago it would have been BASIC, but I wouldn't inflict Visual Basic on any beginner. Python maybe? Or does someone have other suggestions.
The most important part of this article may be the photo. If they can actually make a tank that big that holds together under 10,000 psi and make it light enough to lift, that is quite an accomplishment.
Not that I'd be particularly happy to see tanks like that traveling down the road. Even filled with a non-reactive gas such as nitrogen, a 10,000 psi tank is a bomb waiting for a hard bump to set it off. Filled with hydrogen, which makes an explosive mixture with air... 8-(
I wonder if they plan to include a small turbo-generator to recover all the energy used to compress the gas.
If you get the electricity from solar panels and store the excess daytime power from them as hydrogen from seawater, then you may get some real savings. I say "may", not "will", because making those solar cells has its own cost in energy and pollution, deploying them and the electrolysis plant has significant environmental impacts, etc. But without spending a few months trying to find all the figures, it does seem reasonable that putting solar cells on all the southward facing roof-tops in the sunny part of the USA would in the long run have a positive environmental impact as compared to burning fossil fuels at an ever-increasing rate.
However, this plan is extremely expensive. At present (I suspect artificially low) US electricity rates, a solar cell probably won't pay back it's purchase cost in it's expected life of 30 years. Maybe higher volume production of solar cells or better technology will cut the price, but they've been saying that ever since 1973... Besides requiring millions of individuals to make large and probably losing investment in solar cells, the power companies have to make some hardware investments (not big ones, but they budget for everything to last 20+ years so replacing one-way meters, etc., is going to go slowly) and (far more difficult) change the way they do business: instead of buying power at a few big plants and selling it to thousands of consumers, they have to both buy and sell from consumers, and sell to a few big electrolysis plants. (For a couple of decades they'd also be running some old fossil fuel plants every night -- otherwise they'd lose too much of the money they put into building them.)
Do you truly have a cable network that can deliver VOD to everyone at once, or do you have VOD that works as long as just a few people use it? If you really have a network capable of 1 channel per user, you are mighty lucky to be served by one of the few forward-looking companies. Probably you won't know until the system does overload -- unless you can get a few hundred neighbors to cooperate in a test and bring it down NOW. I've heard too many stories of cable executives enthusiastically pushing cable modems while being utterly clueless about the network segmentation ("Huh?") that is needed to make those work for more than a few people. Chances are the same thing will happen with VOD.
VOD and PVR are not the same thing, no matter how much nCube tries. Yep. Well, you could deliver PVR from a server over a VOD-capable network, but you can also crack nuts by setting off a hand grenade near them. That doesn't make the hand grenade a nutcracker... 8-)
"Worst non-technical article about a technical subject", or maybe "Most errors in one page".
The worst error of all: for this to work at all like Tivo, the cable company would have to dedicate one channel to each subscriber. That means cable loops with less than 100 subscribers on each, which will usually require running fiber further into neighborhoods and installing more fiber-to-cable units. Also the fiber-to-cable units have to be upgraded to select the channels instead of just dumping everything they get to the cable, the fiber bandwidth has to be increased to carry thousands of subscriber channels, and the central office needs lots of high-powered servers.
I do expect all those hardware upgrades to happen in about 10 years, but it's not going to happen just for timeshifting -- it will happen because the hardware is necessary for (1) good high-speed internet service to homes, and (2) to enable the cable companies to sell video rentals. It's going to take a long time to work out the details (mainly how the servers and the content providers split up the money), but on-line video rentals are going to be _big_ someday.
As far as scheduled programming goes, everything the article claimed as a reason for consumers to buy the service looks to me like a reason to avoid it: give me targeted advertisements embedded in the program or with fast-forward locked-out, and I'll spend a lot more time reading books!
Finally, "I can scale it, depending on how popular the service is. It's all under my roof. And I don't have to send a truck out to every customer who wants it. It presents great efficiencies for cable operators." WTF? No one had to send out a truck to put in my VCR or my DVD player. No one has to send out a truck to install Tivo boxes with the hard drive. But if they do implement the proposed scheme, they'll have to send out lots of trucks to do the network upgrades.
No, a DAC step is the change you get when you flip the last bit in the position code -- like 100010001000 to 100010001001. This is a small fraction of a track (because on modern drives you have to compensate for thermal expansion and wear or you won't find the track at all). So when the feedback loop flips the last bit, the head is going to jump a small fraction of a track, and this will leave unerased areas on the fringes.
Stationary erase heads: those would work for a complete wipe, but moving erase heads would also allow for selective erase of one file at a time, making this feature useful through the working life of the machine rather than just when it was moved to a different role. Either way, you would have to give the drive the capability of low-level formatting itself after the wipe, I forgot that. I'd guess that the military would prefer the selective wipe capability, because in the situation where the "instant full erase" was really needed, they'd probably prefer just to package a thermite grenade with the drive -- you might not have power to run an erase cycle but yank that pin when you've got to bug out, and the disk (and much else) is a puddle of melted glass and metal 5 seconds later... And finally, yes I did mean what you call a "random write treatment a few times." There is a mil-spec for that, something like write all 0's, write all 1's, write 010101...., write 101010..., and repeat X times. That is, the erase head would have to be capable of writing 1's and 0's both, and doing this harder and wider than the write head.
The reason businesses would store their data at a service bureau is that they don't want the hassles of managing the server themselves. People who truly know what they are doing are both rare and hard for a layman to distinguish from all those that just think they know what they are doing, so contracting the whole thing out to a company with apparent expertise is quite appealing. And with a big enough service bureau, you do get a certain assurance that somewhere in the organization there must be someone that is really competent, and if you yell loud enough about problems you experience they are going to assign that competent person to straighten things out.
How about a service bureau that administers your network and servers at your site instead of dragging your data off to their site?
Actually, most banks in 1800 were reasonably reliable, because most people lived in small towns and their banker was someone they knew from childhood and had good reason to trust. Also, in a small town you rip off one customer and soon you'll have no customers. But there were enough fools who couldn't figure out who to trust to make a few crooked bankers enormously rich, even if they had to leave town immediately afterwards, and there were many more bankers that were honest but stupid and lost their customers' money along with their own... But now, you can walk into a bank and hand your money to a total stranger, without even checking the financial condition of the bank, and be pretty confident about getting your money back. But the reason for that isn't that the _banks_ found solutions, it is that the legislatures imposed very heavy regulations on the banks. There are strict limits to how much risk the banks can take. Their books are kept in very specific ways. Inspectors from both the state and the federal government (if the bank is federally insured) will come in and check the books. People go to jail if the regulations aren't followed...
It's pretty hard to imagine software thriving in an atmosphere like that. Legislators understand money pretty well, but are clueless about technical issues. Even if they understood, they couldn't change the laws fast enough to keep up with technology that changes significantly even while the legislature is in session, but banks still do pretty much the same thing they did 200 years ago -- they just use computers and phone lines to do it faster.
If you've paid money, that changes the story entirely, and you'll be able to get action immediately. Most of these companies ensure that you've somehow agreed to a contract that completely covers their rear and leave you no right to sue for anything. A smart businessman would insist on changes in that contract, but if your account is too small to be worthwhile bringing in lawyers, it's probably going to be "take it or leave it."
Iomega didn't get away clean because this concerned hardware sold to consumers. Judges understand hardware (at least that it is a tangible item, which is sold, not licensed). And when hardware is sold to consumers, the manufacturer cannot disavow the "implied warranty of merchantability", including "incidental and consequential damages." (Note that when a business buys the hardware, it is presumed that they are capable of understanding and agreeing to the warranty, which in most cases severely limits the recourses available.) I forget which is "incidental" and which is "consequential", but in the Iomega case one would have been the damage from loss of data, the other the time you'd spend trying to get the problem fixed. [Begin off-topic rant] However, IMO Iomega got off extremely lightly: they paid the lawyers millions, and gave the consumers coupons good for a few $ off on Iomega products. The coupons were worthless unless you spent _more_ money on Iomega, and why would anyone burned by the click of death do that? It's like a tobacco company settling a lung-cancer suit with 25% off on a truckload of cigarettes -- wait, I know people dumb enough to fall for that!
But for software and other computer-related intangibles, the industry has so far managed to hoodwink the courts into treating what definitely looks like a "sale" as a sort of indefinite rental instead. You try to sue Microsoft because Windows crashed on your home computer, you spent 10 hours re-installing it, and you lost the pictures of your wedding, and they will point you at that EULA which says (1) you don't own your copy of Windows, you just "licensed" it, and (2) they aren't responsible for incidental and consequential damages or anything else beyond supplying a disk with the object code properly written on it. That the object code is obviously defective doesn't matter -- until some judge decides that if it looks like a sale, it is a sale, and the UCC's implied warranty does apply. Once again note that this only applies to consumers, not businesses. [End off-topic rant]
Anyway, with an ASP, you get a service agreement, which is a contract between you and the ASP, and among other things it says what rights you have when something goes wrong and the ASP can't deliver the agreed-upon service. If you let the ASP's lawyers write this, you probably don't even get your money back. If your lawyer writes it, then they'll owe you plenty. Of course, if the service fails because the power and comm lines were shut off when they ran out of money to pay their bills, they might owe you plenty but you aren't going to get it! So you need to send in both a lawyer and a geek who knows how to specify backups that will enable you to re-build the services you need even if the entire ASP is unavailable. Then you need to keep that geek around to verify that the backups are really being made, along with negotiating other issues that come up during implementation. Of course, you might start wondering why you pay someone who's capable of doing the job himself just to check up on whether someone else is doing it properly...
Bootable Linux floppy would be fine, if the kernel plus the erase program would fit. With DOS, the kernel is tiny (msdos.sys, io.sys, command.com) so I _know_ it would fit.
"Stepping motors" was indeed a mis-statement, sorry. I'm not sure what the author of the article meant by "microstepper", but I believe what any modern disk drive has for head positioning is a tiny electromagnet ("audio coil") driven by a DAC, with DAC steps fine enough that you can use servo data to center the head on the track. The audio coil has a continuous response curve, but the DAC has a limited number of bits and thus a finite step size. So the head is still going to wander a little (by 1/2 a DAC step at least), so there will be data left unerased along the edges here and there.
Maybe you missed this: the article is not concerned with data recovery using the drive as is, but with taking it apart and using a very fine and very sensitive magnetic probe, which would be able to search the fringes for data, besides searching the center track in case the overwrites left ghosts of the original data. If you can get direct control of the DAC, you could control the head position to ensure the entire surface got erased, even between the tracks. I don't know if IDE and SCSI interfaces allow this; the article seemed to assume you can, but that "microstepper" blooper makes me wonder if the author understands the difference between MFM and IDE.
Incidentally, there may be a business opportunity here. Make a disk drive in which a somewhat larger and stronger erase head rides in front of the read/write head. Build in commands to use this to erase a sector or the whole drive, using that big erase head to sweep everything, and following the mil-spec for erasing classified data. Sell it to DOD for $300 each -- in the long run it's going to be cheaper than crushing $100 drives because you can't be sure they are erased, and it wouldn't raise the real cost by even $50. Feel free to use my idea, I'd rather stay poor than get involved in defense contracting again... 8-)
Agreed. The Supremes have made many bad decisions in 200+ years and thousands of cases--but I don't remember _any_ case where they made a bad decision that most politicians at that time wouldn't have done the same or worse. It's not perfect, but what else are you going to use as the ultimate authority, computers?
Good suggestion about staggered terms, but one problem. If they can get a second term, that subjects first-termers to political pressures. If retirement is mandatory after one term, then when a judge dies after 32 years, that leaves his replacement in a bad position (3 years and out). If you give the replacement a 35 year term starting from when he begins, eventually the system will get out of sync again...
Slashdot Mirror
The article did state that it was weighted towards the older albums. Mainly, he was trying to identify the groups that have sold platinums year after year. If you look at just the leading sellers in one year, in music, movies, or books -- most of the time, five years later it's "what were we thinking." If the band broke up about 30 years ago and two of them are dead, but their recordings are still selling, then it's likely there was something like real quality...
If I were to use a modified Linux kernel as the operating system for a VCR, plus an application program compiled with gnu C and using an LGPL'd library, then
1) I would have to send Linus a copy of the kernel source code.
2) I would have to provide the purchasers of the VCR some way to get the kernel source code.
3) The kernel would be GPL.
4) The application could be proprietary.
The main problem I can see in this (aside from execs who believe M$ FUD), is that if we made extensive hardware-related changes to the kernel, being able to download the source code would make it easier for our competitors to reverse-engineer the product. However, in most cases this could be avoided by putting the hardware dependent stuff in the proprietary modules, and doing kernel changes only to trim out the unused pieces. (e.g., no disk drive, so delete all that...)
Am I correctly interpreting the GPL?
You don't have to declare variables and their type before use. Sometimes that is good, sometimes it is bad. When I'm just playing around with a little code to see if something will work, I'd rather not have to go back to the top and insert a declaration every time I find I need a variable for a for loop or something. When I was first learning to program 30 years ago, as best I can remember FORTRAN didn't even _allow_ variable declarations except for arrays, and I think that worked out fairly well -- one less thing to worry about on the first few programs. The old unstructured BASIC was even better IMHO, someone with no idea at all of what they were doing could plunge in and start learning by writing programs that often worked. If Python will allow that approach to learning structured programming, even better. Many Computer Science professors will disagree about that ("they don't learn to plan ahead and map out their data structures"), but in my experience, no one can plan ahead _until_ they've written enough programs to get some idea of what they are doing.
But when I write a serious program that has to work right even under conditions I didn't quite anticipate, I certainly want to list all the variables ahead of time, and to have the compiler tell me when I misspell one or use the wrong type. So being able to choose whether or not definitions are mandatory is ideal. Sounds like Python does that.
pick any J. Random Licensing Agreement and you'll see words to the effect of "We promise nothing. We accept no responsibility. Don't be surprised if this software is utterly unfit for anything, including the task for which you are paying the license fee. Anything that happens is on your head. Now pay up, bitch."
You forgot: "You are not allowed to fix the defects in this program yourself. You may pay us extra for tech support, but we do not guarantee that even that will make it work, or that our tech support people will tell you when your problem comes from a known bug that affects many people. We may sue if you benchmark the program without our permission, because we don't want you publicizing the bugs. When we issue a bug fix, we may make you buy the program again to get it. We don't guarantee that the next version of the program (which you will pay for) will fix the known bugs."
This may be too complicated for pinhead executives to understand. Explain that it's like buying a used car "as-is" with the hood welded shut and a contract that does not allow you to fix it.
Virginia _did_ pass the UCITA, making it about the last place in the USA you would want to litigate a free software case. The Python license had a forum-selection clause specifying that litigation concerning it would be in VA; this is obviously a bad idea, but besides that the FSF doesn't allow any forum-selection clause in GPL's.
But there apparently were a couple of other issues which I don't understand -- there's a list of changes, but the only one explained in the posted letter from the FSF was eliminatiing the forum selection clause, and it's going to take more time than I am willing to put in even to find the original language.
If they can't use it in court, why are they collecting it? To post on the internet? Under J. Edgar Hoover, the FBI collected lots of data that was useless in court, but maybe could be used for politicking, blackmail, or just gave JEH a kick to read. Allegedly his position and the FBI budget were untouchable for 40 years because his secret files contained scandalous info on so many politicians. And in the 60's, he was always trying to get some newspaper to print his files on Martin Luther King, but they didn't consider the reverend's affairs newsworthy. If he'd had the internet so he could publish them anonymously himself...
They claim the FBI has changed. I'm not convinced. At best, they still seem to prefer good publicity to good police work. At worst, they seem to tolerate and cover up their own incompetence (Ruby Ridge, Waco, the FBI labs). Let them secretly troll Congressional e-mail for blackmail material, and they could become completely untouchable...
Right and wrong. The Ninth Amendment: The enumeration in the Constitution, of certain rights, shall not be construed to deny or disparage others retained by the people. It sounds like an invitation for the Supreme court to go discover new rights, but as far as I know they didn't until the 1950's in a case involving a Massachusetts law against contraceptives. The justices decided that this law violated some unenumerated right, which they called the "right to privacy", but which is more accurately described as "the right to control your own body." That is, they've applied it to contraceptives and abortion, but in regards to wiretapping and other snooping, their decisions have been based on the 4th Amendment (Unreasonable search and seizure), not the 9th.
The application of the 4th Amendment when only electrons are "seized" has always made some strict constructionists uncomfortable. So it is possible that with a few more justices who are unwilling to go beyond the actual words in the Constitution, the gov't would be free to wiretap at will, look through the walls of your house with infrared, and order fat people to go on a diet. So I would like to see two amendments specifically codifying a true right to privacy (no snooping without a warrant), and the right to control your own body (which would also invalidate drug laws and many of the FDA's powers).
Darn, you beat me to it! Definitely like the final rocket launch in The Iron Dream.
The incandescent light bulb (with a filament of charred linen) was simultaneously invented by Edison's lab and by some British high-school teacher. They sued each other over patents, and each one won -- in his own country. However, maybe the real credit ought to go to whoever invented a vacuum pump good enough to keep the filament from burning up in residual oxygen. A frenchman (I think) had the basic idea 10 years earlier, but his vacuum wasn't good enough and the bulb died in minutes. The bulbs we use now have a tungsten filament (for a 10x improvement in life-expectancy), sometimes nitrogen gas fill instead of a vacuum, or even a bit of bromine and iodine to combine with evaporating tungsten atoms and re-deposit them on the filament ("halogen" bulbs); I remember that last invention was made by aircraft engineers trying to shrink the size of the wingtip running lights, but I have no idea who made the others.
There was also the electric arc lamp, which was around decades earlier than the incandescent bulb. But in the 19th century it resembled an arc-welder, and someone had to stay there and keep adjusting it as the electrodes burned down. Many decades later, it was tamed by being built into a bulb with automatic control circuits -- neon, mercury arc, and flourescent lights.
I never heard of that airplane in New Zealand. Are there more details on the web? It's possible that simple remoteness cheated the Kiwis out of proper recognition. But it's quite clear why the Wright's got the recognition out of all the Europeans and Americans who were riding wobbly powered kites up in the air at that time: they managed to crash the plane gently enough to be able to repair it and go up again the next day, and to keep it in one piece while they taught themselves how to _land_. Lilienthal (at least) was flying several times before the Wrights, but he always came down hard. He apparently assumed that when he got the airframe right, he could just automatically fly it, while the Wrights did everything possible to prepare themselves and then took baby steps. Even with all that preparation, the first flight ended in a tailspin -- but at 12 feet and maybe 20mph it didn't break the airplane much.
Er... How Tenochtitlan ever managed to import enough food to survive is hard to explain (unless the Aztec army marched slaves into the city as a food supply, but as far as I can recall that was about the only conceivable evil the Spanish didn't accuse the Aztecs of). The city probably was in the middle of extremely productive fields (they had water, and deserts usually are extremely productive when you first start irrigation), but a meat supply would have been darned difficult and the Aztecs don't seem likely to be vegetarians. Did runners wear themselves out bringing in bits of fresh meat for the king, while everyone else lived on corn, beans, and a little jerky?
One thing is that unlike European cities, Tenochtitlan did not have to be self-supporting economically. European cities traded manufactured and imported goods for the food they needed. The Aztec capital apparently depended on a harsh system of conquest, taxation, and slavery. There may have been more slaves carrying dried food in on their backs than there were people living in the city, plus vast numbers of tributary tribesmen (not quite slaves if they behaved themselves, but close enough) growing or catching and drying that food. Then in the city, a lot of slaves who probably got only corn and beans and eventually died of deficiency diseases, a number of nobles who visited for a few days then went back to the country where they could get fresh food, and the kings inner circle and personal guard who got the best of the local garden produce and occasionally a relatively fresh haunch of deer brought in by relays of runners. Maybe they imported dogs for the king's greatest feasts. In any case, the cost of supporting that city must have been a great burden upon the countryside, and the tributary tribes were ripe for revolt when Cortez showed up.
Humankind: the only species capable, and dumb enough to wipe out ... themselves. Not true. Bacteria will wipe themselves out quite well in the right conditions. You fill a dish with rich bacteria food, drop in a few bacteria, keep it warm and wait. Within days (or maybe hours) there will be millions of dead bacteria, poisoned by their own wastes. What _is_ disheartening is that, as smart as one human is, a million humans together are usually no smarter than bacteria...
It's not like ancient man had the tools nessecery to kill off a sufficient quantity of any animal as to drive it to extinction. Yes they did, at least for herd herbivores (horses, camels, mammoths). You follow the herd until it is near a cliff, then the whole tribe comes out screaming and waving torches or whatever, and scare the whole herd over the cliff. Everyone eats until they puke for a few days, then when the remaining 99% of the meat smells too bad to eat you go looking for the next herd. Of course, you do have to live off of roots and bugs for a week or two until the next opportunity for a herd kill arises, but on the average you are well fed.
I think the bison ("buffalo") avoided going with the rest of the large North American herbivores because they mostly live where there aren't enough roots, bugs, and field mice to keep a man alive until the next bison herd comes along. Indians living on the fringes of the great plains where there is barely sufficient sustenance only bagged one or two buffalo herds a year before Europeans brought back the horse, and they did it by driving them over a cliff.
The large predators that also died when man moved into North America are more of a mystery, but it is not inexplicable. The large animals they depended on for most of their meat became scarce. Strange new animals were around and looked like easy prey teetering on their hind legs, but few of those who tried eating them lived to learn better. And in the end, "trophy hunting" may have become a factor; grandpa keeps showing off his necklace with dozens of saber-tooth tiger fangs, and you really want some, but he won't share so you go hunt down the last saber-tooth in the area...
The basic difference is: the early americans ate all the horses. Someplace near present-day Ukrainia, some of my ancestors learned to ride them instead, and the horse-riding peoples spread over Eurasia in waves of conquest (Celts, Hellenes, Huns, Goths, Normans, Mongols... Similarly, tameable breeds of camels, cows, sheep, goats, and donkeys were domesticated in Eurasia or Africa, but eaten in North America. The result 13,000 years later was that the europeans easily overwhelmed the native americans and took their land. Horses and a wider variety of farmable plants helped, but maybe the biggest difference was that with no tame meat animal except the dog, Indian towns had to stay small enough to allow farms nearby to grow the corn needed and the men to _walk_ to their hunting grounds, while much larger European cities were fed by cattle driven into town and grain hauled in by horse-drawn wagons. City life gave Europeans metal, gunpowder, government, large armies, measles, and smallpox, and all the Indians had to counter with was skill in the woods.
So when we drive another hundred species to extinction today, just what possible uses might we have missed?
FORTRAN does fit pure number-crunching applications (science and engineering) fairly well. The syntax and (since 1977) program structure are a pretty good match to specifications written in math, and the compiled executables are _fast_. There are thousands of libraries for mathematical and statistical applications. Finally, it's easier to write a compiler that will automatically divide up the work among multiple CPU's in FORTRAN than in most procedural languages. (This is primarily due to a lack of those features that make C/C++ so flexible but also allow all sorts of non-obvious side effects.)
So if you are going to write programs that will tie up the FPU for hours, days, weeks, or even months, and don't need a user interface or care much what the printout looks like as long as the numbers are accurate, then FORTRAN is the right tool (assuming the job is too big for math worksheet tools like MATLAB). If you need a decent user interface, or have to produce nice looking reports for managers, then using FORTRAN is like building a car body with a hammer and anvil. I did have to do a short hardware-driver type of program in FORTRAN back in engineering school (it was the only compiler the obsolete minicomputer in that lab had), but if there'd been an _assembler_ available it would have been easier. I knew of companies that did all their programming in FORTRAN through the 60's and 70's, from corporate accounting to machine control programs, but it sure wasn't a smart choice.
I haven't looked at Python hard enough yet, but at first glance it does look like a pretty good candidate for the first language. One of the big questions here is just what should be taught in the first course: how computers work (C or even assembly), or how to program in a clean and easy language (Python or some other very high level OO language). I'd get them programming first, then do the low-level stuff in the second term, then cover the languages you'll need on the job the second and subsequent years (C++/Java and database programming). C++ is just too darned hard for the first programming course. Java is better, but I think (I haven't used it) there are still too many traps for novices.
By the way, I've been wondering if my grandson (a very smart and hyper 3rd grader) could get started on programming this summer. 10 years ago it would have been BASIC, but I wouldn't inflict Visual Basic on any beginner. Python maybe? Or does someone have other suggestions.
The most important part of this article may be the photo. If they can actually make a tank that big that holds together under 10,000 psi and make it light enough to lift, that is quite an accomplishment.
Not that I'd be particularly happy to see tanks like that traveling down the road. Even filled with a non-reactive gas such as nitrogen, a 10,000 psi tank is a bomb waiting for a hard bump to set it off. Filled with hydrogen, which makes an explosive mixture with air... 8-(
I wonder if they plan to include a small turbo-generator to recover all the energy used to compress the gas.
If you get the electricity from solar panels and store the excess daytime power from them as hydrogen from seawater, then you may get some real savings. I say "may", not "will", because making those solar cells has its own cost in energy and pollution, deploying them and the electrolysis plant has significant environmental impacts, etc. But without spending a few months trying to find all the figures, it does seem reasonable that putting solar cells on all the southward facing roof-tops in the sunny part of the USA would in the long run have a positive environmental impact as compared to burning fossil fuels at an ever-increasing rate.
However, this plan is extremely expensive. At present (I suspect artificially low) US electricity rates, a solar cell probably won't pay back it's purchase cost in it's expected life of 30 years. Maybe higher volume production of solar cells or better technology will cut the price, but they've been saying that ever since 1973... Besides requiring millions of individuals to make large and probably losing investment in solar cells, the power companies have to make some hardware investments (not big ones, but they budget for everything to last 20+ years so replacing one-way meters, etc., is going to go slowly) and (far more difficult) change the way they do business: instead of buying power at a few big plants and selling it to thousands of consumers, they have to both buy and sell from consumers, and sell to a few big electrolysis plants. (For a couple of decades they'd also be running some old fossil fuel plants every night -- otherwise they'd lose too much of the money they put into building them.)
Do you truly have a cable network that can deliver VOD to everyone at once, or do you have VOD that works as long as just a few people use it? If you really have a network capable of 1 channel per user, you are mighty lucky to be served by one of the few forward-looking companies. Probably you won't know until the system does overload -- unless you can get a few hundred neighbors to cooperate in a test and bring it down NOW. I've heard too many stories of cable executives enthusiastically pushing cable modems while being utterly clueless about the network segmentation ("Huh?") that is needed to make those work for more than a few people. Chances are the same thing will happen with VOD.
VOD and PVR are not the same thing, no matter how much nCube tries. Yep. Well, you could deliver PVR from a server over a VOD-capable network, but you can also crack nuts by setting off a hand grenade near them. That doesn't make the hand grenade a nutcracker... 8-)
"Worst non-technical article about a technical subject", or maybe "Most errors in one page".
The worst error of all: for this to work at all like Tivo, the cable company would have to dedicate one channel to each subscriber. That means cable loops with less than 100 subscribers on each, which will usually require running fiber further into neighborhoods and installing more fiber-to-cable units. Also the fiber-to-cable units have to be upgraded to select the channels instead of just dumping everything they get to the cable, the fiber bandwidth has to be increased to carry thousands of subscriber channels, and the central office needs lots of high-powered servers.
I do expect all those hardware upgrades to happen in about 10 years, but it's not going to happen just for timeshifting -- it will happen because the hardware is necessary for (1) good high-speed internet service to homes, and (2) to enable the cable companies to sell video rentals. It's going to take a long time to work out the details (mainly how the servers and the content providers split up the money), but on-line video rentals are going to be _big_ someday.
As far as scheduled programming goes, everything the article claimed as a reason for consumers to buy the service looks to me like a reason to avoid it: give me targeted advertisements embedded in the program or with fast-forward locked-out, and I'll spend a lot more time reading books!
Finally, "I can scale it, depending on how popular the service is. It's all under my roof. And I don't have to send a truck out to every customer who wants it. It presents great efficiencies for cable operators." WTF? No one had to send out a truck to put in my VCR or my DVD player. No one has to send out a truck to install Tivo boxes with the hard drive. But if they do implement the proposed scheme, they'll have to send out lots of trucks to do the network upgrades.
No, a DAC step is the change you get when you flip the last bit in the position code -- like 100010001000 to 100010001001. This is a small fraction of a track (because on modern drives you have to compensate for thermal expansion and wear or you won't find the track at all). So when the feedback loop flips the last bit, the head is going to jump a small fraction of a track, and this will leave unerased areas on the fringes.
Stationary erase heads: those would work for a complete wipe, but moving erase heads would also allow for selective erase of one file at a time, making this feature useful through the working life of the machine rather than just when it was moved to a different role. Either way, you would have to give the drive the capability of low-level formatting itself after the wipe, I forgot that. I'd guess that the military would prefer the selective wipe capability, because in the situation where the "instant full erase" was really needed, they'd probably prefer just to package a thermite grenade with the drive -- you might not have power to run an erase cycle but yank that pin when you've got to bug out, and the disk (and much else) is a puddle of melted glass and metal 5 seconds later... And finally, yes I did mean what you call a "random write treatment a few times." There is a mil-spec for that, something like write all 0's, write all 1's, write 010101...., write 101010..., and repeat X times. That is, the erase head would have to be capable of writing 1's and 0's both, and doing this harder and wider than the write head.
The reason businesses would store their data at a service bureau is that they don't want the hassles of managing the server themselves. People who truly know what they are doing are both rare and hard for a layman to distinguish from all those that just think they know what they are doing, so contracting the whole thing out to a company with apparent expertise is quite appealing. And with a big enough service bureau, you do get a certain assurance that somewhere in the organization there must be someone that is really competent, and if you yell loud enough about problems you experience they are going to assign that competent person to straighten things out.
How about a service bureau that administers your network and servers at your site instead of dragging your data off to their site?
Actually, most banks in 1800 were reasonably reliable, because most people lived in small towns and their banker was someone they knew from childhood and had good reason to trust. Also, in a small town you rip off one customer and soon you'll have no customers. But there were enough fools who couldn't figure out who to trust to make a few crooked bankers enormously rich, even if they had to leave town immediately afterwards, and there were many more bankers that were honest but stupid and lost their customers' money along with their own... But now, you can walk into a bank and hand your money to a total stranger, without even checking the financial condition of the bank, and be pretty confident about getting your money back. But the reason for that isn't that the _banks_ found solutions, it is that the legislatures imposed very heavy regulations on the banks. There are strict limits to how much risk the banks can take. Their books are kept in very specific ways. Inspectors from both the state and the federal government (if the bank is federally insured) will come in and check the books. People go to jail if the regulations aren't followed...
It's pretty hard to imagine software thriving in an atmosphere like that. Legislators understand money pretty well, but are clueless about technical issues. Even if they understood, they couldn't change the laws fast enough to keep up with technology that changes significantly even while the legislature is in session, but banks still do pretty much the same thing they did 200 years ago -- they just use computers and phone lines to do it faster.
If you've paid money, that changes the story entirely, and you'll be able to get action immediately. Most of these companies ensure that you've somehow agreed to a contract that completely covers their rear and leave you no right to sue for anything. A smart businessman would insist on changes in that contract, but if your account is too small to be worthwhile bringing in lawyers, it's probably going to be "take it or leave it."
Iomega didn't get away clean because this concerned hardware sold to consumers. Judges understand hardware (at least that it is a tangible item, which is sold, not licensed). And when hardware is sold to consumers, the manufacturer cannot disavow the "implied warranty of merchantability", including "incidental and consequential damages." (Note that when a business buys the hardware, it is presumed that they are capable of understanding and agreeing to the warranty, which in most cases severely limits the recourses available.) I forget which is "incidental" and which is "consequential", but in the Iomega case one would have been the damage from loss of data, the other the time you'd spend trying to get the problem fixed. [Begin off-topic rant] However, IMO Iomega got off extremely lightly: they paid the lawyers millions, and gave the consumers coupons good for a few $ off on Iomega products. The coupons were worthless unless you spent _more_ money on Iomega, and why would anyone burned by the click of death do that? It's like a tobacco company settling a lung-cancer suit with 25% off on a truckload of cigarettes -- wait, I know people dumb enough to fall for that!
But for software and other computer-related intangibles, the industry has so far managed to hoodwink the courts into treating what definitely looks like a "sale" as a sort of indefinite rental instead. You try to sue Microsoft because Windows crashed on your home computer, you spent 10 hours re-installing it, and you lost the pictures of your wedding, and they will point you at that EULA which says (1) you don't own your copy of Windows, you just "licensed" it, and (2) they aren't responsible for incidental and consequential damages or anything else beyond supplying a disk with the object code properly written on it. That the object code is obviously defective doesn't matter -- until some judge decides that if it looks like a sale, it is a sale, and the UCC's implied warranty does apply. Once again note that this only applies to consumers, not businesses.
[End off-topic rant]
Anyway, with an ASP, you get a service agreement, which is a contract between you and the ASP, and among other things it says what rights you have when something goes wrong and the ASP can't deliver the agreed-upon service. If you let the ASP's lawyers write this, you probably don't even get your money back. If your lawyer writes it, then they'll owe you plenty. Of course, if the service fails because the power and comm lines were shut off when they ran out of money to pay their bills, they might owe you plenty but you aren't going to get it! So you need to send in both a lawyer and a geek who knows how to specify backups that will enable you to re-build the services you need even if the entire ASP is unavailable. Then you need to keep that geek around to verify that the backups are really being made, along with negotiating other issues that come up during implementation. Of course, you might start wondering why you pay someone who's capable of doing the job himself just to check up on whether someone else is doing it properly...
Bootable Linux floppy would be fine, if the kernel plus the erase program would fit. With DOS, the kernel is tiny (msdos.sys, io.sys, command.com) so I _know_ it would fit.
"Stepping motors" was indeed a mis-statement, sorry. I'm not sure what the author of the article meant by "microstepper", but I believe what any modern disk drive has for head positioning is a tiny electromagnet ("audio coil") driven by a DAC, with DAC steps fine enough that you can use servo data to center the head on the track. The audio coil has a continuous response curve, but the DAC has a limited number of bits and thus a finite step size. So the head is still going to wander a little (by 1/2 a DAC step at least), so there will be data left unerased along the edges here and there.
Maybe you missed this: the article is not concerned with data recovery using the drive as is, but with taking it apart and using a very fine and very sensitive magnetic probe, which would be able to search the fringes for data, besides searching the center track in case the overwrites left ghosts of the original data. If you can get direct control of the DAC, you could control the head position to ensure the entire surface got erased, even between the tracks. I don't know if IDE and SCSI interfaces allow this; the article seemed to assume you can, but that "microstepper" blooper makes me wonder if the author understands the difference between MFM and IDE.
Incidentally, there may be a business opportunity here. Make a disk drive in which a somewhat larger and stronger erase head rides in front of the read/write head. Build in commands to use this to erase a sector or the whole drive, using that big erase head to sweep everything, and following the mil-spec for erasing classified data. Sell it to DOD for $300 each -- in the long run it's going to be cheaper than crushing $100 drives because you can't be sure they are erased, and it wouldn't raise the real cost by even $50. Feel free to use my idea, I'd rather stay poor than get involved in defense contracting again... 8-)
Agreed. The Supremes have made many bad decisions in 200+ years and thousands of cases--but I don't remember _any_ case where they made a bad decision that most politicians at that time wouldn't have done the same or worse. It's not perfect, but what else are you going to use as the ultimate authority, computers?
Good suggestion about staggered terms, but one problem. If they can get a second term, that subjects first-termers to political pressures. If retirement is mandatory after one term, then when a judge dies after 32 years, that leaves his replacement in a bad position (3 years and out). If you give the replacement a 35 year term starting from when he begins, eventually the system will get out of sync again...