1) Send an unsolicited resume _everywhere_, not just to places you have researched and have some reason to think you might have a chance at...
2) When they ask you to stop sending resumes, respond with foul language and threats of lawsuit.
And this idiot did it repeatedly!
Really, even McDonald's won't hire you if they are aware of a history like that.
1.500.000.000 basepairs [are all the DNA coding for protein in the human genome]. So a 1.5 Gb file is enough to encode an entire human being.
The protein-coding chromosomal DNA is very, very far from encoding an entire human being. You've also got the DNA that controls which proteins are expressed (some unknown portion of that other 95% of the chromosomes), mitochondrial DNA, environmental effects during your whole life, and most of all some billions of neurons, each with up to a hundred semi-randomly connections to other neurons. No one yet has come anywhere near to giving a computer the equivalent of the life experience stored in your neurons. (Or at least my neurons -- some people never learn...)
Wrong. It said that it was not a disability because it did not affect her activities outside of the job. That is, the ADA has created a class system -- if you are _really_ disabled in normal life activities as well as the job, you get special privileges at work, but if you merely have problems with particular jobs, all you might get is workman's compensation, which doesn't pay all that much. Don't blame the courts, blame Congress; they passed the law, the judges have to struggle to interpret it.
If you really can't work due to carpal tunnel from your job, you certainly should be entitled to workmans compensation, but that doesn't get you into the big-money settlements. Obviously, the employer can try to find a way to keep you working -- usually this makes economic sense, as compared to paying you somewhat less for no work. But it sounds like this woman somehow was unable to keep up with any job they put her in, and was a real pain to supervise for years, so by now they might think it worthwhile to pay out the full WC plus all the legal fees to fight the ADA case and never see her again.
Except that it's bigger than a planet, and it's the second brown dwarf we've got an actual visual of. (Search upwards for the main discussion of Gliese 229B -- there is a "first" in a very narrow sense, but it's not the first visual.)
Gliese 229B is a brown dwarf orbiting a red dwarf of 0.46 Sol mass. The distance from the parent star is not given. Since 229 is pretty dim, it's easier to see the companion. Picking a brown dwarf out from around a bright star is much more difficult, because the bright star tends to wash out the picture if you are running long exposures to catch dim objects. However, CNN does not indicate the size of the new parent star, only that it is "like our sun".
Looking at the pictures, I think the new parent star is much brighter than 229, and the brown dwarf much closer, so this is indeed a step forward. It just isn't as big a step as you'd think from the CNN article. The CNN sub-head "the closest ever observed around a star through direct imaging" may be accurate, but the headline "In a first, object near a star caught on camera" is misleading, at least.
Correct -- if they were false. But I see a lot more lawsuits alleging NDA violations than alleging slander and libel, for a very obvious reason. Filing suit for libel and slander gives the defendant the opportunity to bring out evidence of corporate misconduct in open court, where major news outlets are likely to be watching. If there's any basis at all to the allegations, a libel suit will generally do more damage to the corporate image than the original allegations.
whats the current ratio? I would take the *zip algorithms as a standard. (I've seen commercial backup software that takes twice as long to compress the data as Winzip but leaves it 1/3 larger.) Zip will compress text files (ASCII such as source code, not MS Word) at least 50% (2:1) if the files are long enough for the most efficient algorithms to work. Some highly repetitive text formats will compress by over 90% (10:1). Executable code compresses by 30 to 50%. AutoCAD.DWG (vector graphics, binary format) compresses around 30%. Back when it was practical to use PKzip to compress my whole hard drive for backup, I expected about 50% average compression. This was before I had much bit-mapped graphics on it.
Bit-mapped graphic files (BMP) vary widely in compressibility depending on the complexity of the graphics, and whether you are willing to lose more-or-less invisible details. A BMP of black text on white paper is likely to zip (losslessly) by close to 100:1 -- and fax machines perform a very simple compression algorithm (sending white*number of pixels, black*number of pixels, etc.) that also approaches 100:1 ratios for typical memos. Photographs (where every pixel is colored a little differently) don't compress nearly as well; the JPEG format exceeds 10:1 compression, but I think it loses a little fine detail. And JPEG's compress by less than 10% when zipped.
IMHO, 100:1 as an average (compressing your whole harddrive, for example), is far beyond "pretty damn good" and well into "unbelievable". I know of only two situations where I'd expect 100:1. One is the case of a bit-map of black and white text (e.g., faxes), the other is with lossy compression of video when you apply enough CPU power to use every trick known.
people can sign away their right to freedom of speech in a contract. But a contract covering illegal activities is null and void -- so if the company tries to use an NDA to sue someone who publicly accused them of discrimination, fraud, illegal dumping, etc., the courts _ought_ to toss the case right out.
No. You go back to the original lunar mission plans (earth orbit assembly), from back before political considerations turned it into "Get two guys there the fastest way possible. And make this the first exploratory voyage in history with zero fatalities."
Keep in mind that electronics we have now is immensely better than it was during the Apollo missions. You can do things with robots that they were scared to try with or without live crews in 1968. The following plan takes a decade of development work, but none of the pieces are too big, and there will be other applications for the technology created:
You use the shuttle to haul pieces of the mission ships into orbit. You don't actually need a space station, just an area where the shuttles bring the pieces and shuttle crews bolt them together. I think one shuttle brings up an orbital transfer robot "tug", the next one brings up a load of communication satellites. The tug has some very efficient low-thrust drive. The tug first takes the comsats out and drops them off in lunar orbit, so we have the needed com relays to the far side.
The tug returns, and meets one to three more shuttles with a disposable robot lunar lander, payload, and fuel. After the pieces are bolted together, the tug takes the lander out to lunar orbit, drops it, and returns for the next piece. The lander lands (of course). Subsequent landers home in on the first piece.
You better have two tugs, and you will need some spares of the landers and telescope sections, because some units will malfunction. But eventually all the pieces of the scope are landed on the moon, and you now know the transportation there is trustworthy. So now you send your man-rated lander out on one tug, with the second one following just in case. The crew lands, bolts things together, and comes home.
Alternately, send out some robot cranes, robot wrenches, etc., and have the final assembly done by robots. It will be slow and rather expensive, but maybe cheaper than sending men. And when you get done, you know how to make robots to tackle the biggest jobs...
Just think - if we hadn't been cursed with this weird coincidence (rotation period == orbital period) our scientists and philosophers might have cottoned on to the laws of orbital motion and the sphericity (?) of the Moon and Earth millennia ago.
It's not a weird coincidence. Because the Moon is so big and so close (relative to most planet/moon relations), tidal forces will affect it's rotational period until it is in sync with the orbital period. To not be tidelocked, the Moon would have to be so far away that it's unlikely that the rotation would be visible to the naked eye...
Would we actually know if a country was wiped out in 3,000 BC? Or maybe there was a middling-sized ocean impact around then: rain for 40 days and 40 nights might be a result of a strike in deep water boiling the ocean. And we don't have any records at all from civilizations on the Atlantic seacoasts until about 500 BC.
How many of the people responding to statusbar actually read his whole post? He's NOT claiming MS quality is OK or as good as the phone company:
Problems happen even with properly engineered systems. When an improperly designed system is put into place, all hell will break loose.
I'm not just talking Microsoft here, there is a real problem with companies/programmers seeing their system work once, and then assuming it is good enough to ship.
We've all seen examples of that. (And to be fair, MS does put it's products through quite a lot of testing -- the trouble is, they've made it possible to have far more configurations than it's possible to test, give the users few tools to figure out where things are going wrong, market the software as NOT requiring knowledgeable users or administrators, and create code that is beyond any one person's comprehension.)
But the phone company does maintain pretty good service, and no system can handle a 1000% overload well. But as someone else pointed out, they do have exchanges that handle a single-point overload without blocking other calls; the problem is that Ticket Master bought the wrong kind of service. I agree that phone availability in many parts of the US is less than what I'd really want for life-critical emergency services, however would you rather have a.99999 reliable telephone system that 1/3 of families can't afford, or a.999 reliable (in two tries at dialing) phone in every house?
Sorry about the bold-face, it was supposed to end after "Man-Month". I should have previewed, but I didn't think it was possible to mispell angle-bracket/b angle-bracket
For writing compilers, OS's, and major applications, you should get a CS degree. It's heavy on mathematics and fundamental principles. A few places also offer "Software Engineering", which seems to be CS with a less math and more large-project management. IF you can handle the math, I'd suggest you take the CS and learn project management later on -- I rather suspect that you'll learn more about that by reading The Mythical Man-month than in any college class anyhow.
But remember, this is the small end of the programming job market. You'd better be damned good if you want those jobs. But if you are good enough, I expect you get a lot of job satisfaction, and a chance at bodacious stock options.
I think the big three in employing programmers are databases, embedded programming, and web design. A lot of would-be programmers also end up as system administrators and tech support -- sometimes combined with coding databases and web pages... CS gives you much of the fundamentals for each of those jobs, but there are easier programs that are more directly oriented towards them.
Business databases have required by far the most programmer man-hours over the last 50 years. CS will give you the mathematics behind the database designs; MIS/CIS should give you more of the practice, more grounding on how to deal with the non-technical managers that are your customers, and maybe some background in system administration, because if you work for a medium sized corporation you will be expected to code the databases in between fixing network & server problems.
Embedded programming: This includes everything from washing machines and microwaves to Tivos and game consoles. 95% of the programs are really simple, but there are so many of them that this may include more than 50% of all coding manhours. However, most of them show up in job surveys as engineers rather than "programmers" -- the typical career path is to get an Electrical Engineering degree, design some hardware, then learn to program it. Some real training in programming (CS) would certainly help here, but you've also got to comprehend the hardware.
Web page design: Are there any college majors that are really effective at teaching this? It takes a little artistic talent, a lot of ergonomics engineering, and a tiny bit of coding. Maybe Industrial Engineering would be a better major, it covers ergonomics and artistry. Just don't let them brainwash you into thinking that _looking_ good is better than giving the users the info they are looking for without waiting for large graphics to download and then hunting around. The good news: if you rate a C overall for your web design ability, you're far above the average... At most places, you'll also be expected to keep the server running, so whatever training you can get in that will be useful too.
#8 is a *very* important one - it prohibits excessive bail and cruel/unusual punishment. So, would trucking Dmitri Skylarov (and many others, I suspect) around the country for two weeks before bringing him up before a judge for a bail hearing count as violating #8? Personally, I'd rate that as kidnapping across state lines...
It's been the worst year in almost a decade for _most_ industries... Except BS production, but if you could charge for that all/.ers would be millionaires...
I lived in VA for a while... never again. But a couple more stories like this, and it will get tempting. Especially every time I see the guy that's apparently glued to the local congressional seat drooling on TV....
it isn't the facial recgnition that needs work - it's the facial database! If you are trying to ID foreign terrorists, that's true, but pictures of the visa holders won't help much. If you've got one grainy picture of a guy in turban and beard, with his face half turned away, no ID system is going to tell you that the cleanshaven guy in western clothes waiting in the immigration line is the same guy.
But in Tampa apparently they were after ordinary criminals. Most of these have been arrested and mug-shots taken multiple times. If that doesn't give them good enough pictures to work with, there's something seriously wrong in the mug shots, face-matching software, or both. I suspect that it's really the false-positives problem; the cops got tired of running over to look at the latest "bingo" when the pictures don't look at all alike to humans, and missed the few times when it really was the same guy.
they actually bent over backwards to make sure the KLS did not record any of his online keystrokes.
Wiretap warrants require more probable cause, because they will capture _everything_, not just the particular conversation the cops are after. So, instead of getting a wiretap warrant, they got a warrant to go after the encryption key and configured the KLS to discard keystrokes when the modem was active, therefore it didn't catch any keystrokes that were being _directly_ transmitted.
That's highly Jesuitical reasoning. Quite obviously if Scarpo typed e-mail off-line, then dialed in to send it, the KLS would capture that. Sounds like a wiretap to me. More to the principle of the laws, KLS captures everything typed in whenever the modem is off, not just the item specified in the warrant.
Note that although the FBI insisted and finally convinced the judge that the KLS system was "secret" and so the court and Scarfo's lawyers could only see an edited version of the specs, they did let out how to beat it. Keep that modem running! (Wouldn't an ethernet connection also do this? It's continually active on an external cable, and so under their definition of "wiretapping" KLS would have to stay off.)
To me, it looks like the courts are going nuts over tiny technical details, which they hardly understand, while missing the big picture. The FBI has lied and concealed evidence about Waco, protected one of their agents who turned out to be spying for many years (Hansen), and at least one field office (Boston organized crime task force) has become difficult to distinguish from the mobsters. And it's pretty clear by now that if anyone is ever disciplined for Waco, it will be a letter of reprimand sent to their retirement home, and I have no reason to expect any significant firings over the other misdeeds, let alone agents going to jail. Yet, the judge will take the FBI's word for it that the KLS has to be secret and the sanitized description released is sufficiently accurate.
If I could trust the cops to obey the laws and their procedures, I wouldn't worry much about technicalities...
Yes, the poster certainly needs to go back and review the amendments. Ashcroft respects #2 (the right to bear arms) and it's been a very long time since anyone threatened #3. But he's certainly trying to rip up #1 (freedom of speech), #4-6 (search and seizure, self-incrimination, fair trials, etc.), and #9 IIRC (powers not specifically given to the federal gov't are reserved to the states and the people --prosecuting medical marijuana in the states that allow it, or for that matter, prosecuting non-interstate drug use and trading at all). I can't remember what # 7 and 8 were, so I can't definitely say that he's working on violating them, but I think we've already established a pattern...
Um, that leaves #10, which is the extremely vague catch-all about "unenumerated rights", from which the Supreme Court occasionally and inconsistently pulls a "right to privacy". How can I accuse anyone of attempting to break that one?
I would _hope_ that the patent defines the terms. I don't want to spend the time reading it to find out, if that's how they write. And of course, it is quite possible that between 1994 when they filed the patent and 1997, they discovered which way the RDF team was going and added the appropriate definitions to the patent...
If I was on a jury concerning this patent, that language would certainly put me on my guard and predispose me to find something wrong with it. In my experience, made-up words like these occur in four ways:
1) The inventor or scientist is doing something so new that it is necessary to invent words to describe it. However, serious inventors and scientists don't make up words like "Endo-dynamic information node."
2) The "inventor" is a crack-pot.
3) The "inventor" is utterly unaware of existing work in the area, so does not know the proper terminology and makes up his own. He also doesn't know about prior art. And, to make up words like that, he is either a crack-pot or his marketing side is a lot stronger than his technical side, so I'd wonder about whether he really can invent anything.
4) The non-standard terminology is deliberately used so that no one doing a patent search, is likely to find it. This lets the patent holder wait until someone has committed their business to using this technology, instead of working around it by changing the implementation, and then spring the patent on them. If that's not fraud, it ought to be.
As badly as the patent office is doing, (and at least the USPTO never allowed a patent on the wheel), we might consider simply removing the "checking" function from them entirely and just have them run a public on-line database. Applications get posted.
There is a public comment period during which anyone can see the patent and send in examples of prior art, or whatever tends to limit or invalidate the patent. The office appends this to the on-line file. At the end of this period, the filer may withdraw the application and only be out a few $ for the posting fee, but it stays in the database as public-domain prior art. The filer can amend and re-post it just once, but the history is retained. Or the filer can assert that it _is_ a valid patent. There is no government certificate that the patents are any good.
If you still think the patent is invalid, you can sue in a (special, technically savvy) court. You don't have to possibly break the law first by infringing and wait for them to sue you. Both challenges and infringement suits are judged under these rules:
1) Obfuscated language will be interpreted to the advantage of the challenger. (See also "defend it or lose it")
2) If the patent does not contain a specific implementation that was workable at the time of filing, it's invalid.
3) Claims (defining the actual reach of the patent) may be broader than that specific implementation, but overly broad claims are penalized so that someone searching the database doesn't have to manually examine dozens of patents for whether the patent is validly that broad. Any claim that applies to something that was either obvious or had prior art or publication at the time of filing is completely invalid. Two claims invalidated will invalidate the whole patent. And it costs money -- see "Loser Pays".
4) Loser Pays: If any part of the patent is invalidated, the filer must pay all court costs, legal fees, and a penalty to the challenger. (Bad patents are a big enough public nuisance to make it worthwhile allowing lawyers to profit from knocking them out.) Invalidated patent applications, or parts thereof, stay in the database, marked either as public domain or with a cross-reference to the earlier patent. If the patent is upheld, the challenger pays, and the database notes that it was upheld and points to the court record.
5) Defend it or Lose it:
(a) Statute of limitations for damages for infringement: 1 year before the defendant is formally notified that he is in infringement, or 2 years befor suit is filed, whichever is less. (Exception -- if it is proven that the defendant knew of the infringement and attempted to conceal it while continuing to infringe.)
(b) Delay in defending patent: If defendant was openly in infringement for two or more years before patent holder notified them to cease, defendant gets one year of royalty-free operations for every two years before notification.
(c) Obfuscating the search process: If it is shown that the use of non-standard terminology in the patent kept it from being found in a reasonably thorough automated search, this is not only a valid defense against damages for infringement, but it will also allow the defendant to continue producing infringing products for at least two years.
One of the challenges in bottle cap technology is designing a device that is guaranteed to seal despite rather gross variations in the bottle top. And to do it very, very cheaply. That's where the real inventiveness comes in. Anyone can seal a bottle with $1.00 worth of rubber, but to handle variations in the bottle top for half a cent, that's tricky.
Anyway, these bottle top patents apply only to a very specific design. If you can improve upon a patented cap at all, you've probably changed it enough to be non-infringing. To get a general patent on the idea of sealing a bottle, you'd have had to file in 5,000 BC or so...
However, software patents often claim to be about that general, in most cases without much justification. IANAL, but what I've seen looking at those that I could to some extent read is:
1) The RSA patent (now expired) would once have been questionable because it patents a mathematical algorithm (if performed by computer), but it did cover a specific and genuinely new idea. Not too bad.
2) Patents describing one specific implementation of a not-especially-new idea, coupled to claims alleging to cover all implementations of that idea. (I haven't looked up the hyperlinks patent, but what else could it be?) The implementation is probably patentable, but easy to work around. The claims are, IMO, way too broad and cover lots of prior art. Bring these guys to court with evidence of prior art, and their claims will get trimmed way back. But why in heck should it be necessary for you to pay lawyers to take the case to court just to continue doing what you were doing before the patent was filed? The problem is, (1) the patent office is out of its depth in evaluting these claims, and (2) there is no penalty for ridiculous claims. So why not make the claims ridiculously broad and see if you can scare someone into paying royalties? My suggestion: loser pays, plus you can lose the valid part of the patent by over-reaching with overly broad claims. That is, if two claims in a patent are invalidated, the whole patent is invalid.
3) Patents filed for ideas the applicant cannot actually implement at the time of original filing, in the hopes that eventually someone will implement it, then discover that they owe you money. This ought to be disallowed in total, on the grounds that if you can't make it work, you haven't _invented_ anything yet! But the USPTO has not been enforcing that in the last few decades.
In addition, the way US patents were handled until recently, it was possible to file an application and keep it in limbo for decades with continual small amendments, changing it to match technical developments, and then finish it up and spring it after the technology had matured. And it stayed secret until then. Hence we got applications first filed about 1960 and 1970, becoming patents in the 1980's claiming to cover integrated circuits (in general) and microprocessors (in general). These were not from the engineers who actually designed the first IC's and uP's in production. And somebody had to go to court to fight them.
4) This "RDF" patent seems to be a new class of "stealth" patent: whatever the actual history and whether or not there was prior art, it looks to me like they used obfuscating language to ensure that no one would realize they were infringing on it until the patent owners started sending out threatening letters. Looks pretty close to fraud to me. (It probably also wasn't truly implemented -- since technology like RDF isn't much use until it is public.)
Slashdot Mirror
1) Send an unsolicited resume _everywhere_, not just to places you have researched and have some reason to think you might have a chance at...
2) When they ask you to stop sending resumes, respond with foul language and threats of lawsuit.
And this idiot did it repeatedly!
Really, even McDonald's won't hire you if they are aware of a history like that.
1.500.000.000 basepairs [are all the DNA coding for protein in the human genome]. So a 1.5 Gb file is enough to encode an entire human being.
The protein-coding chromosomal DNA is very, very far from encoding an entire human being. You've also got the DNA that controls which proteins are expressed (some unknown portion of that other 95% of the chromosomes), mitochondrial DNA, environmental effects during your whole life, and most of all some billions of neurons, each with up to a hundred semi-randomly connections to other neurons. No one yet has come anywhere near to giving a computer the equivalent of the life experience stored in your neurons. (Or at least my neurons -- some people never learn...)
So, can anyone translate that to English, please?
Wrong. It said that it was not a disability because it did not affect her activities outside of the job. That is, the ADA has created a class system -- if you are _really_ disabled in normal life activities as well as the job, you get special privileges at work, but if you merely have problems with particular jobs, all you might get is workman's compensation, which doesn't pay all that much. Don't blame the courts, blame Congress; they passed the law, the judges have to struggle to interpret it.
If you really can't work due to carpal tunnel from your job, you certainly should be entitled to workmans compensation, but that doesn't get you into the big-money settlements. Obviously, the employer can try to find a way to keep you working -- usually this makes economic sense, as compared to paying you somewhat less for no work. But it sounds like this woman somehow was unable to keep up with any job they put her in, and was a real pain to supervise for years, so by now they might think it worthwhile to pay out the full WC plus all the legal fees to fight the ADA case and never see her again.
Except that it's bigger than a planet, and it's the second brown dwarf we've got an actual visual of. (Search upwards for the main discussion of Gliese 229B -- there is a "first" in a very narrow sense, but it's not the first visual.)
Gliese 229B is a brown dwarf orbiting a red dwarf of 0.46 Sol mass. The distance from the parent star is not given. Since 229 is pretty dim, it's easier to see the companion. Picking a brown dwarf out from around a bright star is much more difficult, because the bright star tends to wash out the picture if you are running long exposures to catch dim objects. However, CNN does not indicate the size of the new parent star, only that it is "like our sun".
Looking at the pictures, I think the new parent star is much brighter than 229, and the brown dwarf much closer, so this is indeed a step forward. It just isn't as big a step as you'd think from the CNN article. The CNN sub-head "the closest ever observed around a star through direct imaging" may be accurate, but the headline "In a first, object near a star caught on camera" is misleading, at least.
Correct -- if they were false. But I see a lot more lawsuits alleging NDA violations than alleging slander and libel, for a very obvious reason. Filing suit for libel and slander gives the defendant the opportunity to bring out evidence of corporate misconduct in open court, where major news outlets are likely to be watching. If there's any basis at all to the allegations, a libel suit will generally do more damage to the corporate image than the original allegations.
whats the current ratio? I would take the *zip algorithms as a standard. (I've seen commercial backup software that takes twice as long to compress the data as Winzip but leaves it 1/3 larger.) Zip will compress text files (ASCII such as source code, not MS Word) at least 50% (2:1) if the files are long enough for the most efficient algorithms to work. Some highly repetitive text formats will compress by over 90% (10:1). Executable code compresses by 30 to 50%. AutoCAD .DWG (vector graphics, binary format) compresses around 30%. Back when it was practical to use PKzip to compress my whole hard drive for backup, I expected about 50% average compression. This was before I had much bit-mapped graphics on it.
Bit-mapped graphic files (BMP) vary widely in compressibility depending on the complexity of the graphics, and whether you are willing to lose more-or-less invisible details. A BMP of black text on white paper is likely to zip (losslessly) by close to 100:1 -- and fax machines perform a very simple compression algorithm (sending white*number of pixels, black*number of pixels, etc.) that also approaches 100:1 ratios for typical memos. Photographs (where every pixel is colored a little differently) don't compress nearly as well; the JPEG format exceeds 10:1 compression, but I think it loses a little fine detail. And JPEG's compress by less than 10% when zipped.
IMHO, 100:1 as an average (compressing your whole harddrive, for example), is far beyond "pretty damn good" and well into "unbelievable". I know of only two situations where I'd expect 100:1. One is the case of a bit-map of black and white text (e.g., faxes), the other is with lossy compression of video when you apply enough CPU power to use every trick known.
people can sign away their right to freedom of speech in a contract. But a contract covering illegal activities is null and void -- so if the company tries to use an NDA to sue someone who publicly accused them of discrimination, fraud, illegal dumping, etc., the courts _ought_ to toss the case right out.
Was that one e-mail to 35,000 employees, or 1,000 e-mails to 35 employees?
No. You go back to the original lunar mission plans (earth orbit assembly), from back before political considerations turned it into "Get two guys there the fastest way possible. And make this the first exploratory voyage in history with zero fatalities."
Keep in mind that electronics we have now is immensely better than it was during the Apollo missions. You can do things with robots that they were scared to try with or without live crews in 1968. The following plan takes a decade of development work, but none of the pieces are too big, and there will be other applications for the technology created:
You use the shuttle to haul pieces of the mission ships into orbit. You don't actually need a space station, just an area where the shuttles bring the pieces and shuttle crews bolt them together. I think one shuttle brings up an orbital transfer robot "tug", the next one brings up a load of communication satellites. The tug has some very efficient low-thrust drive. The tug first takes the comsats out and drops them off in lunar orbit, so we have the needed com relays to the far side.
The tug returns, and meets one to three more shuttles with a disposable robot lunar lander, payload, and fuel. After the pieces are bolted together, the tug takes the lander out to lunar orbit, drops it, and returns for the next piece. The lander lands (of course). Subsequent landers home in on the first piece.
You better have two tugs, and you will need some spares of the landers and telescope sections, because some units will malfunction. But eventually all the pieces of the scope are landed on the moon, and you now know the transportation there is trustworthy. So now you send your man-rated lander out on one tug, with the second one following just in case. The crew lands, bolts things together, and comes home.
Alternately, send out some robot cranes, robot wrenches, etc., and have the final assembly done by robots. It will be slow and rather expensive, but maybe cheaper than sending men. And when you get done, you know how to make robots to tackle the biggest jobs...
Just think - if we hadn't been cursed with this weird coincidence (rotation period == orbital period) our scientists and philosophers might have cottoned on to the laws of orbital motion and the sphericity (?) of the Moon and Earth millennia ago.
It's not a weird coincidence. Because the Moon is so big and so close (relative to most planet/moon relations), tidal forces will affect it's rotational period until it is in sync with the orbital period. To not be tidelocked, the Moon would have to be so far away that it's unlikely that the rotation would be visible to the naked eye...
Would we actually know if a country was wiped out in 3,000 BC? Or maybe there was a middling-sized ocean impact around then: rain for 40 days and 40 nights might be a result of a strike in deep water boiling the ocean. And we don't have any records at all from civilizations on the Atlantic seacoasts until about 500 BC.
How many of the people responding to statusbar actually read his whole post? He's NOT claiming MS quality is OK or as good as the phone company:
.99999 reliable telephone system that 1/3 of families can't afford, or a .999 reliable (in two tries at dialing) phone in every house?
Problems happen even with properly engineered systems. When an improperly designed system is put into place, all hell will break loose.
I'm not just talking Microsoft here, there is a real problem with companies/programmers seeing their system work once, and then assuming it is good enough to ship.
We've all seen examples of that. (And to be fair, MS does put it's products through quite a lot of testing -- the trouble is, they've made it possible to have far more configurations than it's possible to test, give the users few tools to figure out where things are going wrong, market the software as NOT requiring knowledgeable users or administrators, and create code that is beyond any one person's comprehension.)
But the phone company does maintain pretty good service, and no system can handle a 1000% overload well. But as someone else pointed out, they do have exchanges that handle a single-point overload without blocking other calls; the problem is that Ticket Master bought the wrong kind of service. I agree that phone availability in many parts of the US is less than what I'd really want for life-critical emergency services, however would you rather have a
Sorry about the bold-face, it was supposed to end after "Man-Month". I should have previewed, but I didn't think it was possible to mispell angle-bracket /b angle-bracket
For writing compilers, OS's, and major applications, you should get a CS degree. It's heavy on mathematics and fundamental principles. A few places also offer "Software Engineering", which seems to be CS with a less math and more large-project management. IF you can handle the math, I'd suggest you take the CS and learn project management later on -- I rather suspect that you'll learn more about that by reading The Mythical Man-month than in any college class anyhow.
But remember, this is the small end of the programming job market. You'd better be damned good if you want those jobs. But if you are good enough, I expect you get a lot of job satisfaction, and a chance at bodacious stock options.
I think the big three in employing programmers are databases, embedded programming, and web design. A lot of would-be programmers also end up as system administrators and tech support -- sometimes combined with coding databases and web pages... CS gives you much of the fundamentals for each of those jobs, but there are easier programs that are more directly oriented towards them.
Business databases have required by far the most programmer man-hours over the last 50 years. CS will give you the mathematics behind the database designs; MIS/CIS should give you more of the practice, more grounding on how to deal with the non-technical managers that are your customers, and maybe some background in system administration, because if you work for a medium sized corporation you will be expected to code the databases in between fixing network & server problems.
Embedded programming: This includes everything from washing machines and microwaves to Tivos and game consoles. 95% of the programs are really simple, but there are so many of them that this may include more than 50% of all coding manhours. However, most of them show up in job surveys as engineers rather than "programmers" -- the typical career path is to get an Electrical Engineering degree, design some hardware, then learn to program it. Some real training in programming (CS) would certainly help here, but you've also got to comprehend the hardware.
Web page design: Are there any college majors that are really effective at teaching this? It takes a little artistic talent, a lot of ergonomics engineering, and a tiny bit of coding. Maybe Industrial Engineering would be a better major, it covers ergonomics and artistry. Just don't let them brainwash you into thinking that _looking_ good is better than giving the users the info they are looking for without waiting for large graphics to download and then hunting around. The good news: if you rate a C overall for your web design ability, you're far above the average... At most places, you'll also be expected to keep the server running, so whatever training you can get in that will be useful too.
#8 is a *very* important one - it prohibits excessive bail and cruel/unusual punishment. So, would trucking Dmitri Skylarov (and many others, I suspect) around the country for two weeks before bringing him up before a judge for a bail hearing count as violating #8? Personally, I'd rate that as kidnapping across state lines...
It's been the worst year in almost a decade for _most_ industries... Except BS production, but if you could charge for that all /.ers would be millionaires...
I lived in VA for a while ... never again. But a couple more stories like this, and it will get tempting. Especially every time I see the guy that's apparently glued to the local congressional seat drooling on TV....
it isn't the facial recgnition that needs work - it's the facial database! If you are trying to ID foreign terrorists, that's true, but pictures of the visa holders won't help much. If you've got one grainy picture of a guy in turban and beard, with his face half turned away, no ID system is going to tell you that the cleanshaven guy in western clothes waiting in the immigration line is the same guy.
But in Tampa apparently they were after ordinary criminals. Most of these have been arrested and mug-shots taken multiple times. If that doesn't give them good enough pictures to work with, there's something seriously wrong in the mug shots, face-matching software, or both. I suspect that it's really the false-positives problem; the cops got tired of running over to look at the latest "bingo" when the pictures don't look at all alike to humans, and missed the few times when it really was the same guy.
they actually bent over backwards to make sure the KLS did not record any of his online keystrokes.
Wiretap warrants require more probable cause, because they will capture _everything_, not just the particular conversation the cops are after. So, instead of getting a wiretap warrant, they got a warrant to go after the encryption key and configured the KLS to discard keystrokes when the modem was active, therefore it didn't catch any keystrokes that were being _directly_ transmitted.
That's highly Jesuitical reasoning. Quite obviously if Scarpo typed e-mail off-line, then dialed in to send it, the KLS would capture that. Sounds like a wiretap to me. More to the principle of the laws, KLS captures everything typed in whenever the modem is off, not just the item specified in the warrant.
Note that although the FBI insisted and finally convinced the judge that the KLS system was "secret" and so the court and Scarfo's lawyers could only see an edited version of the specs, they did let out how to beat it. Keep that modem running! (Wouldn't an ethernet connection also do this? It's continually active on an external cable, and so under their definition of "wiretapping" KLS would have to stay off.)
To me, it looks like the courts are going nuts over tiny technical details, which they hardly understand, while missing the big picture. The FBI has lied and concealed evidence about Waco, protected one of their agents who turned out to be spying for many years (Hansen), and at least one field office (Boston organized crime task force) has become difficult to distinguish from the mobsters. And it's pretty clear by now that if anyone is ever disciplined for Waco, it will be a letter of reprimand sent to their retirement home, and I have no reason to expect any significant firings over the other misdeeds, let alone agents going to jail. Yet, the judge will take the FBI's word for it that the KLS has to be secret and the sanitized description released is sufficiently accurate.
If I could trust the cops to obey the laws and their procedures, I wouldn't worry much about technicalities...
Yes, the poster certainly needs to go back and review the amendments. Ashcroft respects #2 (the right to bear arms) and it's been a very long time since anyone threatened #3. But he's certainly trying to rip up #1 (freedom of speech), #4-6 (search and seizure, self-incrimination, fair trials, etc.), and #9 IIRC (powers not specifically given to the federal gov't are reserved to the states and the people --prosecuting medical marijuana in the states that allow it, or for that matter, prosecuting non-interstate drug use and trading at all). I can't remember what # 7 and 8 were, so I can't definitely say that he's working on violating them, but I think we've already established a pattern...
Um, that leaves #10, which is the extremely vague catch-all about "unenumerated rights", from which the Supreme Court occasionally and inconsistently pulls a "right to privacy". How can I accuse anyone of attempting to break that one?
I would _hope_ that the patent defines the terms. I don't want to spend the time reading it to find out, if that's how they write. And of course, it is quite possible that between 1994 when they filed the patent and 1997, they discovered which way the RDF team was going and added the appropriate definitions to the patent...
If I was on a jury concerning this patent, that language would certainly put me on my guard and predispose me to find something wrong with it. In my experience, made-up words like these occur in four ways:
1) The inventor or scientist is doing something so new that it is necessary to invent words to describe it. However, serious inventors and scientists don't make up words like "Endo-dynamic information node."
2) The "inventor" is a crack-pot.
3) The "inventor" is utterly unaware of existing work in the area, so does not know the proper terminology and makes up his own. He also doesn't know about prior art. And, to make up words like that, he is either a crack-pot or his marketing side is a lot stronger than his technical side, so I'd wonder about whether he really can invent anything.
4) The non-standard terminology is deliberately used so that no one doing a patent search, is likely to find it. This lets the patent holder wait until someone has committed their business to using this technology, instead of working around it by changing the implementation, and then spring the patent on them. If that's not fraud, it ought to be.
As badly as the patent office is doing, (and at least the USPTO never allowed a patent on the wheel), we might consider simply removing the "checking" function from them entirely and just have them run a public on-line database. Applications get posted.
There is a public comment period during which anyone can see the patent and send in examples of prior art, or whatever tends to limit or invalidate the patent. The office appends this to the on-line file. At the end of this period, the filer may withdraw the application and only be out a few $ for the posting fee, but it stays in the database as public-domain prior art. The filer can amend and re-post it just once, but the history is retained. Or the filer can assert that it _is_ a valid patent. There is no government certificate that the patents are any good.
If you still think the patent is invalid, you can sue in a (special, technically savvy) court. You don't have to possibly break the law first by infringing and wait for them to sue you. Both challenges and infringement suits are judged under these rules:
1) Obfuscated language will be interpreted to the advantage of the challenger. (See also "defend it or lose it")
2) If the patent does not contain a specific implementation that was workable at the time of filing, it's invalid.
3) Claims (defining the actual reach of the patent) may be broader than that specific implementation, but overly broad claims are penalized so that someone searching the database doesn't have to manually examine dozens of patents for whether the patent is validly that broad. Any claim that applies to something that was either obvious or had prior art or publication at the time of filing is completely invalid. Two claims invalidated will invalidate the whole patent. And it costs money -- see "Loser Pays".
4) Loser Pays: If any part of the patent is invalidated, the filer must pay all court costs, legal fees, and a penalty to the challenger. (Bad patents are a big enough public nuisance to make it worthwhile allowing lawyers to profit from knocking them out.) Invalidated patent applications, or parts thereof, stay in the database, marked either as public domain or with a cross-reference to the earlier patent. If the patent is upheld, the challenger pays, and the database notes that it was upheld and points to the court record.
5) Defend it or Lose it:
(a) Statute of limitations for damages for infringement: 1 year before the defendant is formally notified that he is in infringement, or 2 years befor suit is filed, whichever is less. (Exception -- if it is proven that the defendant knew of the infringement and attempted to conceal it while continuing to infringe.)
(b) Delay in defending patent: If defendant was openly in infringement for two or more years before patent holder notified them to cease, defendant gets one year of royalty-free operations for every two years before notification.
(c) Obfuscating the search process: If it is shown that the use of non-standard terminology in the patent kept it from being found in a reasonably thorough automated search, this is not only a valid defense against damages for infringement, but it will also allow the defendant to continue producing infringing products for at least two years.
One of the challenges in bottle cap technology is designing a device that is guaranteed to seal despite rather gross variations in the bottle top. And to do it very, very cheaply. That's where the real inventiveness comes in. Anyone can seal a bottle with $1.00 worth of rubber, but to handle variations in the bottle top for half a cent, that's tricky.
Anyway, these bottle top patents apply only to a very specific design. If you can improve upon a patented cap at all, you've probably changed it enough to be non-infringing. To get a general patent on the idea of sealing a bottle, you'd have had to file in 5,000 BC or so...
However, software patents often claim to be about that general, in most cases without much justification. IANAL, but what I've seen looking at those that I could to some extent read is:
1) The RSA patent (now expired) would once have been questionable because it patents a mathematical algorithm (if performed by computer), but it did cover a specific and genuinely new idea. Not too bad.
2) Patents describing one specific implementation of a not-especially-new idea, coupled to claims alleging to cover all implementations of that idea. (I haven't looked up the hyperlinks patent, but what else could it be?) The implementation is probably patentable, but easy to work around. The claims are, IMO, way too broad and cover lots of prior art. Bring these guys to court with evidence of prior art, and their claims will get trimmed way back. But why in heck should it be necessary for you to pay lawyers to take the case to court just to continue doing what you were doing before the patent was filed? The problem is, (1) the patent office is out of its depth in evaluting these claims, and (2) there is no penalty for ridiculous claims. So why not make the claims ridiculously broad and see if you can scare someone into paying royalties? My suggestion: loser pays, plus you can lose the valid part of the patent by over-reaching with overly broad claims. That is, if two claims in a patent are invalidated, the whole patent is invalid.
3) Patents filed for ideas the applicant cannot actually implement at the time of original filing, in the hopes that eventually someone will implement it, then discover that they owe you money. This ought to be disallowed in total, on the grounds that if you can't make it work, you haven't _invented_ anything yet! But the USPTO has not been enforcing that in the last few decades.
In addition, the way US patents were handled until recently, it was possible to file an application and keep it in limbo for decades with continual small amendments, changing it to match technical developments, and then finish it up and spring it after the technology had matured. And it stayed secret until then. Hence we got applications first filed about 1960 and 1970, becoming patents in the 1980's claiming to cover integrated circuits (in general) and microprocessors (in general). These were not from the engineers who actually designed the first IC's and uP's in production. And somebody had to go to court to fight them.
4) This "RDF" patent seems to be a new class of "stealth" patent: whatever the actual history and whether or not there was prior art, it looks to me like they used obfuscating language to ensure that no one would realize they were infringing on it until the patent owners started sending out threatening letters. Looks pretty close to fraud to me. (It probably also wasn't truly implemented -- since technology like RDF isn't much use until it is public.)