Re:What's wrong with giving TM holders first dibs?
on
Master Of Your Domain
·
· Score: 1
Giving TM holders universal first dibs is just plain wrong. TM law is only meant to apply to a certian product or class of products. That is, I could, if I really wanted to, TM Linux brand cola, or Microsoft brand toilet tissue. And, as Linux isn't a soft drink and Microsoft doesn't make toilet tissue, thats perfectly OK.
But, giving TM owners first dibs means that people with a direct interest in a name, but don't necessarily have a trademark, may be unable to register that name. If, for example, some company made OpenSSH brand widgets, and had a trademark on that, the proposed rules would allow them to snatch up all OpenSSH domains, even though they don't need all of them, and even though not all of them are proper. The OpenSSH project would just be SOL (although they appear to be anyway)
What would be better would to specify that a domain name may only be held if a) the holder has an external interest (foo.com goes to foo corp, etc) in that name, b) the TLD under which the name is held is appropriate (no network sites under.com and no small, local ISP's under.net for example) c) they aren't preventing another, more appropriate party from using their name (bar corp owning bar.sucks or barsucks.com is a nono) and finally d) they aren't attempting to fool people into thinking they are somebody else (check www.pgp.org for a good example of not tricking)
But SSH (more specifically, OpenSSH) will allow you to do encrypted tunnelling of any IP service. No SSL needed (if you aren't worried about ppl. spying on stuff as it goes from the source site to your tunneller).
I remember a game that came with netware or something like that called NetWar. It was a 3D-space combat game. Single-player mode bit, but wohooo was multi ever fun. Crappy graphics, but it would run on a 286. Man, that thing was better than any other game I've played since.
Yes it should compile. Whitespace shouldn't matter. The compiler shouldn't care if you use a [CR], a [TAB], a [SPACE], the number or combination of such characters you use, or whatever, just so long as there is at least one whitespace character where a whitespace character needs to be.
I'm really sorry, but...
on
SSH v. SRP
·
· Score: 1
>>With brute force you must have all primes available for checking
No you don't. A brute force search can just try all possible factors of n up to sqrt(n). BFI is very bad, as it runs in sqrt(n) time (so for a 1024 bit key that's ~ 2^512 computations).
>>The order of searches is about O( c* ( Q - (P+1) )/36 ). I'll look into the size of c.
Even if your analysis of your algorithm is accurate (I doubt it), it's still hideously poor. That's not any better than the BFI in big-oh notation. Also, the 'c' and the '36' don't matter worth a hill of beans in big-oh. Big-oh doesn't care about constant multipliers of the function. So, it would be more proper to say it's "O(Q-(P+1))".
The best factoring algorthim around is the general number field sieve, which runs in
e^(1.923 * (ln n)^(1/3) * (ln(ln n))^(2/3))
time. The e^ part makes it look bad, but the ln's about cancel it out. That's damn fast for factoring, but damn slow for an algorithm, as it still works out to ~ exponential time if you do it off of the # of bits in the key. It's way better than 2^(#bits/2), though.
#define RANT This is stuff from freshman year CS. There is no magic factoring algorithm. A lot of people way smarter than anybody on slashdot have spent a shitload of time working on this crap. I've heard that under a hundred people in the whole world understand the math behind the number field theory. You certianly ain't one of them. #undefine RANT
Several people have come up with saying that it costs a lot of money to defend a patent. I have one small question for them. Why? What is keeping me from going in and representing myself? It really doesn't cost me anything but my time to go in and argue that they are treading on my patent, especially if there are willing technical experts willing to volunteer as expert witnesses. Whatever big company is troubling me, however, is spending thousands of dollars a day on legal costs. I think it would make sense to patent our ideas, and then rabidly attack anybody who treads on them at all. No skin off my back to show up in court for a couple of weeks, but they can't afford every Tom Dick and Harry suing them for patent infringement. Somewhat off-topic, but I also fail to see how "the big boys" bullied Rio around so much. Couldn't Rio have just stuck one lone member of their legal department on handling all such stuff? If your case is pretty clear, it doesn't seem necessary (to me) to spend a fortune defending yourself. Let the other guy waste his money in repeated lawsuits. Just a thought. Comments?
I used to have a part time job at the local public library, and responsibility for maintaining our public access internet machines fell to me.
AOL has a most lovely Java install feature, that manages to bypass all of the security stuff that we had installed. For some reason, disallowing AOL's Java priviliges doesn't work. So, any asshole who wants to install AOL on our machines can. I spent more time fixing AOL-intall related problems than any other sort.
I left since before the new version came out, so it's not my headache. But all the people I know who still work there are going to hate it.
BTW, is there any way to convince McAffe or Norton to view the AOL install files as virii and nuke them accordingly?
>>In effect there are no 'native' crusoe instructions. Indeed, tm is reluctant to even release the current instructions for fear that major developers will code to them and then force backward compatibility on tm!
Not really. They can just have the code morpher for their later chips translate the old crusoe intructions into the new, updated one.
Really, they way they have it set up, they never have to worry about a compatability problem, backwards or otherwise. They can run any instruction set, including their own old one. They just have to be able to detect which instruction set it ought to be translating from.
It is true that today the antibiotics are being heavily abused, but the beginings of restant strains started way back during the development of penecillin.
When they were doing the first trials, some researchers noticed that some strains had become resistant, even theough the drugs were newly introduced. When penecillin was finally approved, penecillin-resistant tuberculosis was well-established.
There actually is little risk of bacteria/virii developing resistence to the drugs *IF* they are properly used. But during the first penecillin trials, they had no idea1 what proper dosing was, and therefore contributed to resistence by giving too low of a dose, or not keeping treatment up for a long enough period. At $50/$100 a 'scrip, there is a huge chance that people will take the drugs for a couple of days, and then stop because they feel better, and the drug is so expensive.
You make a good point about developing countries, and another big problem there is that many of the drugs available to them may be expired, or not fully potent. This makes resistence development more likely. Also, people in developing countries are poor (by defn.), and therefore more likely not to take a full course of treatment, but stop halfway through because they feel better.
I once knew somebody who had been on penecillin for over 10 years straight. She had had many recurrent ear-infections, so the doctor just said to stay on the drugs. She has faithfully made (nearly) every dose. Because of this, she probably isn't risking any bacteria developing resistence. On the other hand, I read about one clinic where they had a tubercular crack addict who would not ever finish their prescription. The people at the clinic were of the opinion that this one person was responsible for 3 new multi-resistant strains of tuberculosis.]
it just goes to show that how you use the drugs is very important.
That's what really frightens me. Back in the days when antibiotics were new, nobody thought of resistance (because it was a new concept), and used the new drugs too much. Now resistant strains of bacteria are everywhere, and the old antibiotics are nearly useless.
Perhaps if the FDA requires that any antivirals be used in combination with at least 1 other, resistance could be prevented.
I really have to disagree with point 2. Perhaps it is true that physics is taught harder (poorly) than CS or say, chemical engineering, but it isn't really intrically harder.
Just look at the degree requirements for physics, CS, and ChE (all in the engineering school). The math requirement for physics at Washington University St. Louis is Calc 1, 2, 3, differential equations, and engineering mathematics. For chemical engineering, the math requirement is Calc 1, 2, 3, differential equations, engineering mathematics, and probability and statistics. For CS, it's Calc 1, 2, 3, differential equations, engineering mathematics, probability and statistics, discrete math, and an algorithms class (it's applied math).
Notice anything about those lists. Physics requires the least math of any of the majors. CS actually requires the most. Now, that's because CS is really just applied mathematics, but still, physics doesn't even require a stats class. Someone mentioned that physics is hard because it has a lot of paradoxes. Raise your hand if you can define the difference between a countable and uncountable set, and also the different types of uncountable sets. Yep, that's right. Physics may deal with things that don't make sense physically, but theoretical CS deals with logical conundrums. Like there being more than one type of infinity. At least in physics the equations come out clean (or as clean as any diffy eq. do [shudder]).
Now, I'm not knocking physicists. As Dr. Lederman said, ~63% of the GDP comes from discoveries made in physics. But it isn't intrinsically harder than any other science.
I think that the way it works is that if you use a patented technique in a mod to the kernel, either it must be royalty-free for derivitive copies (when you distribute the mod and other people redistribute it), or you must not distribute it at all. So by licencing a copy of licence to the NSA with their patented modifications, they revoke to right to charge the NSA a royalty, and they cannot charge anybody that the NSA redistributes it to a royalty. But, since the NSA isn't very likely to redistribute it, they don't have to worry.
On the other hand, if the NSA does decide to redistribute it, they cannot restrict the NSA's redistribution without violating the GPL. The claus quoted above is for things like PGP, which are royalty free in every country besides the US. Technically, if a person in, say, Germany, uses the RSA cryptosystem in a GPL-ed project, they cannot distributed it to the US, and they may place a limit in the licence forbidding redistribution to the US. But if RSA licenced something under the GPL in the US, then they can't charge a royalty for copies/mods of it.
I'd highly recommend that everybody just sit down and read the text of the GPL slooowly several times. It's really interesting the provisions that are built into it to prevent a contractual paradox.
These things always remind me of a unusual consequense of the Godel Incompleteness theorem (which is a proof that you can't prove everything).
The consequence is essentially that any sufficiently powerful computer system cannot be made virus/cracker proof. No matter how good your AV software and how tight your security procedures, unless you limit the power of your machine (not how fast it runs, but what sorts of things can it do) you cannot ensure your security.
I've decided that it's really not worth the bother to run a totally secure system, and I don't even run a virus scanner anymore.
Before everybody jumps on my back, I'd better clarify "sufficiently powerful". You could say that a machine that is stored in a locked room w/out any connection to any external network that requires a swipecard and 128=byte password to access has perfect security. But, such a machine is not "sufficiently powerful" to be crackable. It is less powerful than my pokey 'ol 486, because my 486 can connect to the internet. If I wanted, I could set it up as a web server. But a machine in a locked room can't do this, and therefore is less powerful.
Even if you have an internet connection, if it refuses all external connections and is behind a good firewall, it may be impossible to break into. But again, it is less powerful than any web-server, even one that just displays static pages. It is once you cross a certain threshold of useability that it becomes "sufficiently powerful". If you have an open telnet port and 1 user account, it is probably "sufficiently powerful". That's not a very high threshold.
The threshold for virii is even lower than cracking. If I want to run outside software, I have to expose myself to virii. If I have good AV software installed and running, I may be able to detect all virii. But, then I can't run any programs that appear sufficiently virus-like because the AV software will flag it, and Godel's Incompleteness theorem shows that if my software catches all virii, it MUST catch some non-virii.
So, security is an impossible goal.
It's still pretty cool to have AV software that automatically looks for 'new' virii though.
Superheating would be fall under "special tricks", since last I checked you had to be really carefull to get water to superheat. I doubt it is even possible to get water to superheat in a McDonalds coffeepot even if you tried.
While the whole checks and balances thing is nice, in reality, congress has all of the power. They have the power to propose constitutional amendments (per article V) and, since they are the head member of the political parties, they have the power to get them shoved through all of the state legislatures.
If the NSA really pissed congress off, they could constitutionally ban it and require the president to do away with them or face charges of treason. And in addition, congress has exclusive power of the purse, as set forth in article I, section 9, clause 7: Clause 7: No Money shall be drawn from the Treasury, but in Consequence of Appropriations made by Law; and a regular Statement and Account of the Receipts and Expenditures of all public Money shall be published from time to time.
Actually, I think it'd be neat if somebody wrote a program to recursively fetch the whole site (except for the catalog part) every hour or so and search the results for "etoy.com". I mean, it is illegal to ping a site every second to take it down. But it isn't illegal to ping it to see if it's up, or even to ping it every hour or so to get an uptime history. And seeing if they post anything about the etoy business on their site is a lot more useful than just pinging them.
Well, yes. That's called a DOS attack, and while it might be justified, it's illegal.
Now, if you just sent one complaint automatically, that be better, but not perfect. Not that it's illegal to bitch about spam to the sender, but often spam is fakemail, and if you trust a software to complain automatically, someday your software will be tricked into complaining to somebody other than the spammer. Multiply your one complaint by 150K and you get quite the mail bomb. Of course maybe the spammer could be fingered for launching a DOS attack against that person, but unless somebody tracks them down manually, it won't help.
It's not that they're stupid, but they're overworked and don't have the necessary training. Most CS-type patents are probably handled by somebody whose training was in electrical engineer, not CS. They issue patents that are obvoius to programmers because they aren't programmers. They issue patents that any fool in the IT biz would know of prior art for because they aren't IT ppl.
Whenever somebody tells me that they are thinking of getting a law degree, I tell them that they should go for patent law. Not so that they can be one of the jerkoff lawyers who file these things, but so that they can be a patent examiner. I encourage anybody out there who is good at computers but not that terriffic of a programmer to work for the patent office. They need your help badly.
I'm getting sick of all these trivial lawsuits. Toshiba, Wizards of the Coast, etc. These lawsuits are increasing my costs of living for silly reasons. If anybody decides to file a class action against one of these law firms for filing expesive but frivilous lawsuits, count me in.
I do have to say that if Trek continues on now as it has been, it won't last long.
Paramount's only cash cow is Star Trek. As a business they cannot exist without it. Unfortunately the cow is going dry. I'd put the beginning of the end when the networks got fed up with being required to air paramounts other shows in order to air Trek. When DS9 stopped being aired regularly on network TV was the day that I stopped watching regularly. It's back now, but I don't have the time to catch up with what I missed. The series really needs a rest, and get a fresh start, or at least some new visionary writers.
It isn't that any of the new Trek stuff is bad, it's just that it is all old-hat now. They are covering stuff that has been done before in Trek and in other places. Star Trek used to be real science fiction. You know, the good stuff that wins Nebula Awards, not those pulp series books (the Trek books actually come to mind). Since its comercial success, they don't have the guts to try anything new. Science fiction isn't dead, and it's been around a lot longer than Start Trek. As long as there are new, revolutionary ideas being written about in SF, there is a chance for Trek.
It's really a pity that Nimoy won't be doing any more movies. If a good movie was written with a good Spock part, I'd hope that he'd agree to do the part. Of course, nothing with any artistic value is going to be put out, simply because Paramount wants a money maker, not a cultural treasure. Really, that's the big problem. Paramount wants money, and they fail to realize that the reason that Trek has been so successfull is that it is a cultural icon. Not on par with Shakespeare, but it certainly is something that I'd like my grandchildren to be exposed to some distant day in the future. Unfortunately they'll have to live with Kirk dying an ignoble death, while Spock, Bones, and Scotty just fade away...
I remember reading somewhere about a small-time publisher that Coca-Cola tried to stomp all over for using the words "It's the real thing" in an advertisement for a book they published. The president of said publisher wrote a nice little reply about the tradition of the small press to resist being bullied into censoring what they put out, and how Ben Franklin was a printer and helped write the first amendment and so on. And then he ended his letter pointing out that they were quoting what a book reviewer had said about their book.
Several people have said that you can't loose a copyright if it is abused. That isn't strictly true. If the owner of the copyright knows of the abuse and fails to take reasonable action defending it, they can loose their copyright, and the material enters public domain. A specific example of this happened when David Eddings had to shut down a MUD based on his books. The creators of the MUD wrote him to tell him what they did, and he sadly had to tell them to shut it down. He did, however, very kindly explain why he had to defend his copyright, and pointed out that if he hadn't heard about the violation he wouldn't have been forced to shut them down. So to sum up, kyes Virginia, you can loose your copyright.
and it just so happens that he is a co-sponsor of the SAFE bill (the original, not the DoD version). Not that the SAFE bill will ever get by a presidential veto, but it still gives me a warm fuzzy feeling.
I can tell you several things that the NSA does/has done.
They sponsor a mathematics competition (the USAMTS) for high-school students.
They helped design the DES algorithm, specifying components (I think it was the S-boxes specifically) in such a way to protect DES against differential and logarithmic attacks. Public cryptanalysis didn't develop these attack techniques until well after DES was released.
They do a lot of crytographic research, and a lot of research in mathematics in general.
They also have a very informative web page (unlike Transmeta's) which tells about several more of thier activities.
The NSA isn't some mysterious organization that never gives out any information. It just happens that most of the information that they collect is sensitive, and can't be public. But what they do is no big secret, it's just the results of thier activities that have to be hidden.
Slashdot Mirror
Giving TM holders universal first dibs is just plain wrong. TM law is only meant to apply to a certian product or class of products. That is, I could, if I really wanted to, TM Linux brand cola, or Microsoft brand toilet tissue. And, as Linux isn't a soft drink and Microsoft doesn't make toilet tissue, thats perfectly OK.
But, giving TM owners first dibs means that people with a direct interest in a name, but don't necessarily have a trademark, may be unable to register that name. If, for example, some company made OpenSSH brand widgets, and had a trademark on that, the proposed rules would allow them to snatch up all OpenSSH domains, even though they don't need all of them, and even though not all of them are proper. The OpenSSH project would just be SOL (although they appear to be anyway)
What would be better would to specify that a domain name may only be held if a) the holder has an external interest (foo.com goes to foo corp, etc) in that name, b) the TLD under which the name is held is appropriate (no network sites under .com and no small, local ISP's under .net for example) c) they aren't preventing another, more appropriate party from using their name (bar corp owning bar.sucks or barsucks.com is a nono) and finally d) they aren't attempting to fool people into thinking they are somebody else (check www.pgp.org for a good example of not tricking)
I think that covers everything.
But SSH (more specifically, OpenSSH) will allow you to do encrypted tunnelling of any IP service. No SSL needed (if you aren't worried about ppl. spying on stuff as it goes from the source site to your tunneller).
I remember a game that came with netware or something like that called NetWar. It was a 3D-space combat game. Single-player mode bit, but wohooo was multi ever fun. Crappy graphics, but it would run on a 286. Man, that thing was better than any other game I've played since.
Yes it should compile. Whitespace shouldn't matter. The compiler shouldn't care if you use a [CR], a [TAB], a [SPACE], the number or combination of such characters you use, or whatever, just so long as there is at least one whitespace character where a whitespace character needs to be.
>>With brute force you must have all primes available for checking
/36 ). I'll look into the size of c.
No you don't. A brute force search can just try all possible factors of n up to sqrt(n). BFI is very bad, as it runs in sqrt(n) time (so for a 1024 bit key that's ~ 2^512 computations).
>>The order of searches is about O( c* ( Q - (P+1) )
Even if your analysis of your algorithm is accurate (I doubt it), it's still hideously poor. That's not any better than the BFI in big-oh notation. Also, the 'c' and the '36' don't matter worth a hill of beans in big-oh. Big-oh doesn't care about constant multipliers of the function. So, it would be more proper to say it's "O(Q-(P+1))".
The best factoring algorthim around is the general number field sieve, which runs in
e^(1.923 * (ln n)^(1/3) * (ln(ln n))^(2/3))
time. The e^ part makes it look bad, but the ln's about cancel it out. That's damn fast for factoring, but damn slow for an algorithm, as it still works out to ~ exponential time if you do it off of the # of bits in the key. It's way better than 2^(#bits/2), though.
#define RANT
This is stuff from freshman year CS. There is no magic factoring algorithm. A lot of people way smarter than anybody on slashdot have spent a shitload of time working on this crap. I've heard that under a hundred people in the whole world understand the math behind the number field theory. You certianly ain't one of them.
#undefine RANT
Several people have come up with saying that it costs a lot of money to defend a patent. I have one small question for them. Why? What is keeping me from going in and representing myself? It really doesn't cost me anything but my time to go in and argue that they are treading on my patent, especially if there are willing technical experts willing to volunteer as expert witnesses. Whatever big company is troubling me, however, is spending thousands of dollars a day on legal costs. I think it would make sense to patent our ideas, and then rabidly attack anybody who treads on them at all. No skin off my back to show up in court for a couple of weeks, but they can't afford every Tom Dick and Harry suing them for patent infringement. Somewhat off-topic, but I also fail to see how "the big boys" bullied Rio around so much. Couldn't Rio have just stuck one lone member of their legal department on handling all such stuff? If your case is pretty clear, it doesn't seem necessary (to me) to spend a fortune defending yourself. Let the other guy waste his money in repeated lawsuits. Just a thought. Comments?
I used to have a part time job at the local public library, and responsibility for maintaining our public access internet machines fell to me.
AOL has a most lovely Java install feature, that manages to bypass all of the security stuff that we had installed. For some reason, disallowing AOL's Java priviliges doesn't work. So, any asshole who wants to install AOL on our machines can. I spent more time fixing AOL-intall related problems than any other sort.
I left since before the new version came out, so it's not my headache. But all the people I know who still work there are going to hate it.
BTW, is there any way to convince McAffe or Norton to view the AOL install files as virii and nuke them accordingly?
>>In effect there are no 'native' crusoe instructions. Indeed, tm is reluctant to even release the current instructions for fear that major developers will code to them and then force backward compatibility on tm!
Not really. They can just have the code morpher for their later chips translate the old crusoe intructions into the new, updated one.
Really, they way they have it set up, they never have to worry about a compatability problem, backwards or otherwise. They can run any instruction set, including their own old one. They just have to be able to detect which instruction set it ought to be translating from.
It is true that today the antibiotics are being heavily abused, but the beginings of restant strains started way back during the development of penecillin.
When they were doing the first trials, some researchers noticed that some strains had become resistant, even theough the drugs were newly introduced. When penecillin was finally approved, penecillin-resistant tuberculosis was well-established.
There actually is little risk of bacteria/virii developing resistence to the drugs *IF* they are properly used. But during the first penecillin trials, they had no idea1 what proper dosing was, and therefore contributed to resistence by giving too low of a dose, or not keeping treatment up for a long enough period. At $50/$100 a 'scrip, there is a huge chance that people will take the drugs for a couple of days, and then stop because they feel better, and the drug is so expensive.
You make a good point about developing countries, and another big problem there is that many of the drugs available to them may be expired, or not fully potent. This makes resistence development more likely. Also, people in developing countries are poor (by defn.), and therefore more likely not to take a full course of treatment, but stop halfway through because they feel better.
I once knew somebody who had been on penecillin for over 10 years straight. She had had many recurrent ear-infections, so the doctor just said to stay on the drugs. She has faithfully made (nearly) every dose. Because of this, she probably isn't risking any bacteria developing resistence. On the other hand, I read about one clinic where they had a tubercular crack addict who would not ever finish their prescription. The people at the clinic were of the opinion that this one person was responsible for 3 new multi-resistant strains of tuberculosis.]
it just goes to show that how you use the drugs is very important.
That's what really frightens me. Back in the days when antibiotics were new, nobody thought of resistance (because it was a new concept), and used the new drugs too much. Now resistant strains of bacteria are everywhere, and the old antibiotics are nearly useless.
Perhaps if the FDA requires that any antivirals be used in combination with at least 1 other, resistance could be prevented.
I really have to disagree with point 2. Perhaps it is true that physics is taught harder (poorly) than CS or say, chemical engineering, but it isn't really intrically harder.
Just look at the degree requirements for physics, CS, and ChE (all in the engineering school). The math requirement for physics at Washington University St. Louis is Calc 1, 2, 3, differential equations, and engineering mathematics. For chemical engineering, the math requirement is Calc 1, 2, 3, differential equations, engineering mathematics, and probability and statistics. For CS, it's Calc 1, 2, 3, differential equations, engineering mathematics, probability and statistics, discrete math, and an algorithms class (it's applied math).
Notice anything about those lists. Physics requires the least math of any of the majors. CS actually requires the most. Now, that's because CS is really just applied mathematics, but still, physics doesn't even require a stats class. Someone mentioned that physics is hard because it has a lot of paradoxes. Raise your hand if you can define the difference between a countable and uncountable set, and also the different types of uncountable sets. Yep, that's right. Physics may deal with things that don't make sense physically, but theoretical CS deals with logical conundrums. Like there being more than one type of infinity. At least in physics the equations come out clean (or as clean as any diffy eq. do [shudder]).
Now, I'm not knocking physicists. As Dr. Lederman said, ~63% of the GDP comes from discoveries made in physics. But it isn't intrinsically harder than any other science.
I think that the way it works is that if you use a patented technique in a mod to the kernel, either it must be royalty-free for derivitive copies (when you distribute the mod and other people redistribute it), or you must not distribute it at all. So by licencing a copy of licence to the NSA with their patented modifications, they revoke to right to charge the NSA a royalty, and they cannot charge anybody that the NSA redistributes it to a royalty. But, since the NSA isn't very likely to redistribute it, they don't have to worry.
On the other hand, if the NSA does decide to redistribute it, they cannot restrict the NSA's redistribution without violating the GPL. The claus quoted above is for things like PGP, which are royalty free in every country besides the US. Technically, if a person in, say, Germany, uses the RSA cryptosystem in a GPL-ed project, they cannot distributed it to the US, and they may place a limit in the licence forbidding redistribution to the US. But if RSA licenced something under the GPL in the US, then they can't charge a royalty for copies/mods of it.
I'd highly recommend that everybody just sit down and read the text of the GPL slooowly several times. It's really interesting the provisions that are built into it to prevent a contractual paradox.
These things always remind me of a unusual consequense of the Godel Incompleteness theorem
(which is a proof that you can't prove everything).
The consequence is essentially that any sufficiently powerful computer system cannot be made virus/cracker proof. No matter how good your
AV software and how tight your security procedures, unless you limit the power of your machine (not how fast it runs, but what sorts of things can it do) you cannot ensure your security.
I've decided that it's really not worth the bother to run a totally secure system, and I don't even run a virus scanner anymore.
Before everybody jumps on my back, I'd better clarify "sufficiently powerful". You could say that a machine that is stored in a locked room w/out any connection to any external network that requires a swipecard and 128=byte password to access has perfect security. But, such a machine is not "sufficiently powerful" to be crackable. It is less powerful than my pokey 'ol 486, because my 486 can connect to the internet. If I wanted, I could set it up as a web server. But a machine in a locked room can't do this, and therefore is less powerful.
Even if you have an internet connection, if it refuses all external connections and is behind a good firewall, it may be impossible to break into. But again, it is less powerful than any web-server, even one that just displays static pages. It is once you cross a certain threshold of useability that it becomes "sufficiently powerful". If you have an open telnet port and 1 user account, it is probably "sufficiently powerful". That's not a very high threshold.
The threshold for virii is even lower than cracking. If I want to run outside software, I have to expose myself to virii. If I have good AV software installed and running, I may be able to detect all virii. But, then I can't run any programs that appear sufficiently virus-like because the AV software will flag it, and Godel's Incompleteness theorem shows that if my software catches all virii, it MUST catch some non-virii.
So, security is an impossible goal.
It's still pretty cool to have AV software that automatically looks for 'new' virii though.
Superheating would be fall under "special tricks", since last I checked you had to be really carefull to get water to superheat. I doubt it is even possible to get water to superheat in a McDonalds coffeepot even if you tried.
While the whole checks and balances thing is nice, in reality, congress has all of the power. They have the power to propose constitutional amendments (per article V) and, since they are the head member of the political parties, they have the power to get them shoved through all of the state legislatures.
If the NSA really pissed congress off, they could constitutionally ban it and require the president to do away with them or face charges of treason. And in addition, congress has exclusive power of the purse, as set forth in article I, section 9, clause 7:
Clause 7: No Money shall be drawn from the Treasury, but in Consequence of Appropriations made by Law; and a regular Statement and Account of the Receipts and Expenditures of all public Money shall be published from time to time.
Actually, I think it'd be neat if somebody wrote a program to recursively fetch the whole site (except for the catalog part) every hour or so and search the results for "etoy.com". I mean, it is illegal to ping a site every second to take it down. But it isn't illegal to ping it to see if it's up, or even to ping it every hour or so to get an uptime history. And seeing if they post anything about the etoy business on their site is a lot more useful than just pinging them.
Well, yes. That's called a DOS attack, and while it might be justified, it's illegal.
Now, if you just sent one complaint automatically, that be better, but not perfect. Not that it's illegal to bitch about spam to the sender, but often spam is fakemail, and if you trust a software to complain automatically, someday your software will be tricked into complaining to somebody other than the spammer. Multiply your one complaint by 150K and you get quite the mail bomb. Of course maybe the spammer could be fingered for launching a DOS attack against that person, but unless somebody tracks them down manually, it won't help.
It's not that they're stupid, but they're overworked and don't have the necessary training. Most CS-type patents are probably handled by somebody whose training was in electrical engineer, not CS. They issue patents that are obvoius to programmers because they aren't programmers. They issue patents that any fool in the IT biz would know of prior art for because they aren't IT ppl.
Whenever somebody tells me that they are thinking of getting a law degree, I tell them that they should go for patent law. Not so that they can be one of the jerkoff lawyers who file these things, but so that they can be a patent examiner. I encourage anybody out there who is good at computers but not that terriffic of a programmer to work for the patent office. They need your help badly.
Moo. www.distributed.net
I'm getting sick of all these trivial lawsuits. Toshiba, Wizards of the Coast, etc. These lawsuits are increasing my costs of living for silly reasons. If anybody decides to file a class action against one of these law firms for filing expesive but frivilous lawsuits, count me in.
I do have to say that if Trek continues on now as it has been, it won't last long.
Paramount's only cash cow is Star Trek. As a business they cannot exist without it. Unfortunately the cow is going dry. I'd put the beginning of the end when the networks got fed up with being required to air paramounts other shows in order to air Trek. When DS9 stopped being aired regularly on network TV was the day that I stopped watching regularly. It's back now, but I don't have the time to catch up with what I missed. The series really needs a rest, and get a fresh start, or at least some new visionary writers.
It isn't that any of the new Trek stuff is bad, it's just that it is all old-hat now. They are covering stuff that has been done before in Trek and in other places. Star Trek used to be real science fiction. You know, the good stuff that wins Nebula Awards, not those pulp series books (the Trek books actually come to mind). Since its comercial success, they don't have the guts to try anything new. Science fiction isn't dead, and it's been around a lot longer than Start Trek. As long as there are new, revolutionary ideas being written about in SF, there is a chance for Trek.
It's really a pity that Nimoy won't be doing any more movies. If a good movie was written with a good Spock part, I'd hope that he'd agree to do the part. Of course, nothing with any artistic value is going to be put out, simply because Paramount wants a money maker, not a cultural treasure. Really, that's the big problem. Paramount wants money, and they fail to realize that the reason that Trek has been so successfull is that it is a cultural icon. Not on par with Shakespeare, but it certainly is something that I'd like my grandchildren to be exposed to some distant day in the future. Unfortunately they'll have to live with Kirk dying an ignoble death, while Spock, Bones, and Scotty just fade away...
Would the Mathematica lab happen to be near the 6-pack? :)
I remember reading somewhere about a small-time publisher that Coca-Cola tried to stomp all over for using the words "It's the real thing" in an advertisement for a book they published. The president of said publisher wrote a nice little reply about the tradition of the small press to resist being bullied into censoring what they put out, and how Ben Franklin was a printer and helped write the first amendment and so on. And then he ended his letter pointing out that they were quoting what a book reviewer had said about their book.
Coca-cola stopped harrassing them.
Several people have said that you can't loose a copyright if it is abused. That isn't strictly true. If the owner of the copyright knows of the abuse and fails to take reasonable action defending it, they can loose their copyright, and the material enters public domain. A specific example of this happened when David Eddings had to shut down a MUD based on his books. The creators of the MUD wrote him to tell him what they did, and he sadly had to tell them to shut it down. He did, however, very kindly explain why he had to defend his copyright, and pointed out that if he hadn't heard about the violation he wouldn't have been forced to shut them down. So to sum up, kyes Virginia, you can loose your copyright.
and it just so happens that he is a co-sponsor of the SAFE bill (the original, not the DoD version). Not that the SAFE bill will ever get by a presidential veto, but it still gives me a warm fuzzy feeling.
I can tell you several things that the NSA does/has done.
They sponsor a mathematics competition (the USAMTS) for high-school students.
They helped design the DES algorithm, specifying components (I think it was the S-boxes specifically) in such a way to protect DES against differential and logarithmic attacks. Public cryptanalysis didn't develop these attack techniques until well after DES was released.
They do a lot of crytographic research, and a lot of research in mathematics in general.
They also have a very informative web page (unlike Transmeta's) which tells about several more of thier activities.
The NSA isn't some mysterious organization that never gives out any information. It just happens that most of the information that they collect is sensitive, and can't be public. But what they do is no big secret, it's just the results of thier activities that have to be hidden.
That's my $.02