Okay, so it's "offensive". If you read Paul Grahams essay though, you should've encountered the idea that stuff that is offensive migth nevertheless be *true*.
My father studied history and norwegian, and did quite well in both. He now teaches the same subjects. I studied cryptography.
Fact is, if we where to change jobs for say a month, I'd be able to do basically perform his job, while he wouldn't know how to even start doing any of the things I do in my job.
This doesn't by itself imply that history is easier than crypto. It just means that crypto is less forgiving -- if you know little about it it is not possible to fake being half-decent in it. While even if you know literally nothing about history, you'll be able to fake being a teacher of it pretty well simply by always reading a few chapters ahead of what you're currently teaching, afterall the ones you teach also doesn't know much.
In general, areas where there is precisely one very obviously correct answer, and no two ways about it are harder to fake than areas where the answer is fuzzy, open to debate and generally not so much "wrong" or "correct" as more or less good.
You see it in grades too. In maths there are more people who fall trough, than in history. That ain't nessecarily because math is "harder", it migth also be because when someone does not know math, it is very obvious and no amount of handwaving gets around the fact that 2 + 2 is not, infact 5.
Rigth. Reads like dreams of the accounting-department for some game-developer.
Recurring revenue, in addition to up-front revenue. Perhaps also the possibility of controlling (read destroying) the second-hand market aswell as "turning off" old games that are no longer profitable.
Most people I know are simply not interested in any model where they're supposed to continue being nickeled and dimed after having bougth a game. It's just bad value.
You can buy a new game for like $50, and if you get tired of it after a while simply resell it. If you don't absolutely *have* to have the game the very first month it's out, (Actually, what is a fun game in say january 2004 is very likely to provide you with the same amount of fun if you choose to play it 6 months later) you can buy it used for maybe $25.
It's not that bad for game-companies either. Knowing that I can always resell a game if it doesn't appeal to me makes me much more likely to impulse-buy a unknown game. And the money I occasinaolly regain by selling older games are mostly reinvested in newer ones.
I realise you're being tongue-in-cheek. But your point is a good one. Actually I'm absolutely sure that ten times as many children are abused (in various ways, not just sexually) by family-members as by "strangers-from-the-internet".
The bogey-man in the bushes is what the media cry about, mostly, but most abused children are still abused by parents, grandparents, uncles, aunts, friends of the family, whatever.
That ain't an argument against families. Just as the fact that the internet enables communications between everyone *including* the small minority that are child-abusers isn't really much of an argument against the internet.
If you actually read through, he's claiming 99.9% of the top 60 public channels on IRC are largely illegal behavior.
No. If you actually read the article he does not even claim this. What he claims is that if you join the top 6 channels on each of the top ten networks (not the same as the "top 60 channels"), log the "conversation" for 36 hours, grep this for "Microsoft", "Norton", "Jasc", "Symantec" then of the resulting lines of text, 99.9% will appear to be warez-trafficking. (i.e. only 0.1% of those lines are people legitimately discussing these companies/software.
This is no analysis of IRC. It's not even an analysis of the top channels in IRC. It's only an analysis of 6 words used on the top 6 channels in the top 10 networks.
With similar research you could prove that there is no warez on irc:
Join 60 medium-large channels on IRC. (say from the middle of the channel-list when sorted by users)
Log for 36 hours.
grep for "hello", "perhaps", "linux", "stupid"
Manually ascertain what percentage of those lines appear to be about warez and what percentage appears to be normal chat.
It doesn't take a genius to see that the result would be rather different. This is just an example that when you're allowed to freely choose your parameters, you'll arrive at whatever conclusion you set out to "proove".
No. It's definately *not* a "very difficult" problem. Actually, you gave the solution to this "problem" yourself.
Have the velocity decay. For applications like this, "writing numbers in the air" any acceleration that is more than something like 3-5 seconds "old" is irrelevant.
This *migth* still give problems if you try to use this in say a moving vehicle in city-traffic or very bumpy road, but it'll *certainly* not be thrown off just 'cos the phone thinks it's still moving since yesterday.
If you're trying to figure out if the user just wrote "3" or "5" it's simply completely irrelevant what accelerations took place yesterday. Now that I think about it you'll probably want even faster decay, essentially any movement that took place before the time needed to enter a single digit is irrelevant, that's probably more like one second tops.
The complainers claimed that the sticker unfairly singled out evolution among literally thousands of other scientific theories that are equally much "just a theory".
This is perfectly true. Evolution is no less proven than most other scientific theories. Infact evolution has seen an *enormous* amount of time and energy spent trying to discredit it, without success. That's the best any scientific theory can hope for.
Noone claims that evolution, or any other scientific theory is "absolute proven truth", all we can claim is that these are theories which match all observed evidence better than any other theory we where able to think of.
Yeah sure. Except HDTV isn't, as a matter of fact, anywhere near any practical, much less physical, limit. A common computer-monitor is capable of resolutions significantly higher than HDTV.
Just because you call it "TV" doesn't make it more amazing. What's really the difference between a "monitor" and a "tv" other than that the tv has a tuner anyway ?
Default to blocking outbond connection port 25, but provide an option, say in your typical web-based configuration-interface say "Use third-party smtp-server" or something.
That way the 1% of people who need it can turn it on, and the 99% of people who never even knew about this can be prevented from having their computers spew spam at the rest of us.
If you want to fit 1000Mbit in 100Mhz, you'll need to encode 10 bits on each period, which means your signals must be 2**10 == 1024 times stronger than the noise.
That's about 30db. (10**3 == 1000)
Doesn't tell you much aslong as you know nothing about the noise in that band on the lines offcourse.
data rates in excess of 1 Gigabit require bandwidth in excess of 2 GigaHertz
Sorry, but no.
There is no reason why you cannot modulate several bits onto each phase of a signal using a combination of amplitude and phase-modulation.
Here's a clue for you: 56K modems are possible. This is *not* because the telephone-system transfers frequencies up to 100Khz. Infact the phone-system cuts off at aproximately 4Khz.
The possible transmit-rate is limited by the width of the band *and* by the signal to noise ratio. A bit simplified, if your signal is 128 times as strong as the noise, you can distinguish up to 128 different signals from another, so you can send 7 bits at a time.
That depends. Some entities perform very thorough verification. For example, my bank Skandiabanken does indeed provide you with a client-certificate for SSL and use this along with other measures to identify you.
The reason I suggested returning the signed public-key is that that would make it possible for entities to trust oneanother, if they so choose.
For example, the public library could choose to trust that I am who I say without further verification because say the tax-department has signed my public key.
Thus with a public-key signed by multitude highly trusted entities, you would sometimes be able to identify for low-security applications without further steps.
No. On the contrary it's a stupid suggestion that show primarily that you don't know much about crypto.
First, using one of 8 algorithms randomly provides at *best* 3 extra bits of keyspace to search (2^3 = 8), you could gain the same by adding 3 bits to your key.
Secondly, if flaws are found in *any* of the 8 algorithms (8 times as likely as if you used only one) you'll instantly be able to break 1/8th of the encryptions.
Thirdly it is not always impossible to tell from ciphertext alone what algorithm is likely used. Especially not if you occasionally have the possibility of doing known or even choosen plaintext attacks.
Fourth, implementing 8 different algorithms makes your program more complex. You'll have more chanses of doing mistakes. In general complexity is the enemy of security.
Fifth, either you include info about what algorithm is used, or you force the receiver to brute-force it by testing with all 8 algorithms until it finds one that works. If the first, then you've defeated the purpose of using 8 in the first place, if the second, you've just multiplied the work of the receiver by a worst-case of 8, average of 4.
Sixth, using the same key with multiple different algorithms isn't generally always secure. Sometimes using the same key in different algorithms can reveal info about the key or weaknesses that wouldn't be there if you used it in only 1. You could also have 8 different keys, but then you've increased key-handling headaches.
Seventh, I'm bored now, but I could list another dozen or so reasons why this is a fundamentally bad idea if I wanted to. If you cannot, this only means that you're not understanding enough of what you are trying to argue.
Remember: Anyone can make a cryptosystem that he himself cannot break.
The fees aren't the problem. They are a mere symptom. The problem is handing over control over large pieces of infrastructure to a single company that has already been convicted of abusing it's monopolistic situation in another area.
Single sign-on is desireable. But it needs to be desentralised, and someone needs to think very carefully about what exactly happens when that single sign-on gets compromised.
One simple idea would simply be to have each user equipped with a secret/public keypair. Sigining up to a site would then consist of having the site in question somehow validate whatever they want to validate about you, and thereafter sign your public key.
Login would go somewhat like this:
User presents signed public key and a signed text of type "Login UserName my.current.ip timestamp" to the website.
Website verifies that it's own signature on the key is valid.
Website verifies that the users signature on the message is valid, and that timestamp and current.ip match the reality.
Loosing your secret key would be a disaster, but that is true for *any* system with single signon. You can try to prevent this for example by storing the secret key on a smartcard that does it's own signing of messages. (i.e. no API exists for extracting the key from the smartcard, only an API for saying to the card: "Please sign this with the secretkey")
That'd still not be inhackable, but atleast you'd need physical posession of the card and some funky hardware to be able to learn the secret-key.
Sure. But you know, the thing is, this is equally true for boys. Really.
It's just that, for whatever reason, it seems there are more boys who don't care what their peers thing about their pursuits than girls. Somehow the opinion of the other girls in the class counts more for the girls.
Encouraging girls to think for themselves, make their own choises and do less group-think would indeed be a good thing. I don't mean only technology.
Doing what your everyone else expects of you (or what you imagine they expect, which isn't always the same thing) isn't more important than doing what you *want* to do.
The math is actually *worse* than that here. (Cottbus, Germany) Let's say I compare watching in the theatre with renting a DVD.
Theatre:
Tickets: 2*7 euro + 2*4 euro = 22 euro.
Popcorn, Soda is something like 15 euro.
Total price: 37 euro.
If the movie is supposed to start at 19, rarely are you even let in before 19:10 or so.
Then follows 20 - 30 minutes of advertising. Coming late is not an option.
Then 5-10 minutes of trailers.
Then one or two anti-piracy flicks for good measure. So, they punish those who *don't* pirate but go to the theatres with boring and insulting shit. Offcourse if I pirated the movies instead I wouldn't have to watch this...
Seats are bad. Air-quality is inadequate, especially if the theatre is full, the movie is long, or it's just a warm day in summer.
If instead I rent a movie on DVD, then:
The price is 1.50 euro.
Soda and popcorn is atmost another 2 euro or so.
Total price: 3.50 -- 10% of the theatre-option.
Movie starts when we want it to. There is no advertising of any sort.
The seats are better, the air is good.
The selection of movies is *much* better. The Cinema here is owned by UCI -- you guess how impartially they select movies. They don't even show Oscar-vinners if they aren't produced in Hollywood...
Basically the only reason anyone ever watches stuff in the theatres is if they a) Absolutely want to watch the movie now, and not in a year when it comes in rental. and b) are so honest they won't download the thing.
No. Makes no sense. Or more precisely makes "sense" only for a person with extremely limited knowledge of crypto and in this spesific case, cryptographically secure hashes.
First, the fact that md5 is a published algorith does *not* imply that finding different files with identical hash is easy. Yes, you can just try different files until two matches, this will happen at the latest when you've tested 2^hashsize+1 files, and on the average when you've tested sqrt(2^hashsize) files. That would take forever.
Your suggestion to randomly check say 10% of the file is dumb. This would mean if I changed some small thing in a file so that the file became garbled (or malicious) then in 90% of the cases you wouldn't discover it. It would also mean that you add a 10% overhead to all your downloads. (no show-stopper, I agree, but also not desireable)
The solution is to use a secure hash. md5 has been suspect for quite some time now, sha1 is the current favourite. A secure hash must have the following properties:
Knowledge of the hash must convey no knowledge of the file beyond "could it be a hash from:this: file?"
It must be computationally hard to create two files with the same hash.
It must be computationally easy to compute the hash for any given file.
All of the above must be true even with complete knowledge of the hashing-algorithm.
If we didn't have hashes with these properties we could toss away PGP, infact pretty much forget about digital signatures alltogether (since signing a complete message is very computationally expensive, what is typically done is signing the *hash* of a message). Same with SSL, SSH, and infact something like 95% of all cryptographic products in use today.
Luckily there are secure hashes. Or atleast hashes we have no idea how to break. Ofcourse in principle sometime in the future a break for some of those hashes can be found, in which case we'll need to design better ones.
But there's no indication that this is a eternal arms-race. There are problems that where thougth to be "hard" 30 years ago which are still pretty much equally hard. For example, factoring large primes is still a hard problem now, and there's no indication that will ever change. Some algorithmic progress is being made, and ofcourse Moores law means we'll need a few extra bits of hash every decade, but nothing I'd worry about.
my actual point, which I assume you agree with is that it is not really relevant if you knew you where infringing, nor if the infringing method was programmed by yourself or is simply a part of a program you bougth from an outside source.
Re:Mostly on target..
on
Sun-isms Debunked
·
· Score: 2, Informative
Sure.
But there's a big difference between:
there is no legal basis for Kodak to sue end users over their use of the JRE or JDK. End users did not infringe upon Kodak's patents
which is a direct quote from the article and:
As an end-user you are fully liable, but unless you have very deep pockets, are disliked by the patent-holder, are a very high-profile user, or the patent-holder is having a particularily bad day, they are probably not going to choose to sue you.
which, as you point out is the reality.
The first claim, the one in the article, that no legal basis exists to sue end users is simply wrong. When an author displays such ignorance of patent-law it weakens his credibility overall.
The article seems mostly on target, but on a few occasions the author does go overboard in critisizing as nonsensical things which are not. For example:
there is no legal basis for Kodak to sue end users over their use of the JRE or JDK. End users did not infringe upon Kodak's patents -- they downloaded the Java software in good faith that it was perfectly legal, and they presumably abided by the license terms. Kodak would have absolutely no right to try to recover any damages from an end user or anyone else who was not a party to adding the allegedly infringing code to the Java source code.
We probably all wish for it, but that is not how patent law infact works.
Using something in good faith is no defence against a patent-lawsuit. Neither does it save you that the patented algorithm was added to the software you use by someone else, without your knowledge.
If this was a valid defence, then most Linux-users would also be equally safe, afterall they *also* tend to use Linux in good faith, abiding by its license terms, and they *also* had the hypotetical patented technique added by someone else without their knowledge.
Sadly, that's not how patent-law works. There are basically only 3 relevant questions in a patent-infringement-lawsuit:
Is the patent valid ? (i.e. no prior art, applies in your country, not expired, non-obvious ?)
Are you doing something, or using a product that is doing something covered by the patent ?
Do you have a valid license from the patent-holder to do so ?
If the answers to those are yes, yes and no, then you are guilty. Even if you didn't *know* the patent existed. Even if you had absolutely no idea that your software was doing this. Even if the software infringing on the patent was written by someone else. Hell, even if the software is closed-source and you thus reasonably *couldn't* know that it was doing this. Those are all irrelevant.
That would depend on which wavelength it uses, and how much those wavelengths are absorbed by various atmospheric conditions.
The site would probably be selected for good, stable weather anyway, and the most likely result of unsuitable weather would be the climbers need to slow down, in extreme cases, like the lasers going offline for whatever reason the climbers would need to stop and wait for them to return.
Climbers carrying people would certainly have some mechanism for safely climbing back down even without external power in any case.
So yes, weather is a factor in this scheme, but I don't think it's a critical factor.
You are rigth that 1$/pound is a long way away, and you are rigth that currently we cannot produce the nessecary length of nanotubes, certainly not at volume and at an acceptable price.
Luckily you are wrong about just about everything else. For example:
Gravity is not a constant over the entire 36.000 km to geosynchronous orbit. Rather it decreases with the *square* of the distance from the earths center. This alone makes your estimate aproximately 1 order of magnitude too high. So, cost of electricity would be $13, not $135.
Secondly, electricity is cheaper in larger quantities. Perhaps you can get another factor of 2 or so here.
Thirdly it is quite likely that the space elevator itself will contribute to lowering energy-prices. Space is a pretty ideal place for a solar collector; no atmosphere, no clouds, no nigth, no land-price. Only currently it costs too much to get there.
Add up all this, and energy-price of $1/pound doesnt sound completely out of the question in the longer run, but I agree with you it's unrealistic for a start. (it's just not *quite* as unrealistic as you make it out to be)
Then to the question of the ribbon. It is correct that currently we can only make nanotubes a few mm long. Fortunately we do *not* need to make nanotubes 36000 km long to build a space-elevator.
Rather we'll braid together, in NASAs design they use epoxy aditionally, multiple tubes to make a "ribbon". Sort of like how you can twist together fibers that individually are 10-20cm long into a rope and have that rope have a large fraction of the strength of the individual fibers.
How long the individual fibers need to be, and how the braiding/epoxying must be done for the finished ribbon to be strong enough is an area of intense research at the moment. I've seen some people saying we need individual tubes 1 cm long, while others are saying they'd need to individually be one or even two orders of magnitude more than that. (i.e up to a meter)
Now, making a single molecule 1cm, or even 1m long is certainly a challenge, but its still simpler than making one thousands of kilometres long.
The 40m bridge over the campus-stream is thus certainly *NOT* 5 orders of magnitude away. It's more likely the braids we are making today, consisting of nanotubes a mm or so long are more than strong enough for that job. It's just that noone are likely to build that bridge aslong as the needed nanotube-ropes would cost millions and a suitable steel-cable is available for a few orders of magnitude less.
Don't get me wrong, I also don't think we'll have a working space-elevator, capable of lifting cargo at $1/pound in a decade.
But I do think that the prospects look a fair bit brigther than you make them out to be.
This experiement is interesting, but unfortunately it does not help too much toward a space-elevator. Probably not a single part or technique from this climber can be used on a space-elevator climber. For example:
This one gets energy from a battery-stack. A battery-stack will not have enough energy to climb 36000 km to geostationary orbit. Infact current batteries are atleast 2 orders of magnitude too weak for that.
Climbing-mechanism is here based on gripping the ribbon. Thing is, climbing to geosynch is a 36000 km travel straigth up, even if a 36 hour climb-time is acceptable you'll need to climb at 1000km/h gripping this (or similar) gripping-mechanisms are not up to that, infact this thing climbs 3 orders of magnitude slower.
So, it has a energy-storage and a climbing-mechanism, none of which can climb to space, even with improvements. Instead both components will need to be made fundamentally different.
Most serious designs I've seen use energy from an external source, because if you are carrying your own energy on the climber, then you use most of your power to lift the energy-storage. (sorta like rockets are mostly lifting rocket-fuel) Ideas include powerful lasers shining on the thing from below, being converted to electricity by efficient photocells. (cells tuned to a single frequency like laser can be more efficient than full-spectrum cells) The laser will get weaker as the climber gains heigth, but so will gravity and thus the required energy.
For the actual climbing a non-contact method would be preferable, perhaps something involving magnetism. (essentially a vertical maglev) The trick is to manage that without making the ribbon itself much heavier. (and thus more expensive)
On the face of it, this claim seems pretty dubious.
It's obvious that a wind-generator slows down the passing air, i.e. makes the wind weaker. Afterall it has to take the energy it delivers from somewhere.
What is pretty hard to believe is that wind-generators are in any way special in this sense.
When we remove forest, and replace it with cropland, we take away a lot of wind-braking. A forest is a more efficient brake for the lower air than any conceivable windmill-density. And we have removed a *LOT* of forest the last few hundred years.
To make this plausible they would have to argue that the net sum of human activites act more to erect brakes for the wind than it does to remove them. This seems a pretty unreasonable conclusion on the face of it. And like they say, extreme claims require extreme evidence.
Sure. But this elaborate scheme doesn't bring you much. What you are saying is that it migth be possible to design a system where atleast one person needs to purchase the rigth to listen to a certain song to aquire the key for decoding that song. After *one* hacker has done this, he can liberate the song, i.e. use the key to produce and distribute an unencumbered copy.
So, you've managed to inconvenience the reasonable customer that wants to, say, make a copy of a song in an unencumbered format that works on for example the kids DRM-incapable CD-player while doing pretty close to nothing at all to block the person who's learnt the lesson, stopped buying from you and download his songs unencumbered from whatever filesharing-du-jour instead.
I don't think we really disagree that much. My point is that even only one liberated device is capable of producing unencumbered files from encumbered ones, and those unencumbered files can be distributed easily enough.
Stopping most of the copiers most of the time doesn't actually buy you much at all in a world where the people who copy also share with oneanother.
For that matter, there's always the analog-hole. The large majority of music-consumers are happy with 128kbps crappily-encoded mp3s. Those will also settle for a well-produced analog-hole copy. And again, not everyone needs to do this, it's enough that *one* person does it and shares the resulting unencumbered file somehow.
Slashdot Mirror
My father studied history and norwegian, and did quite well in both. He now teaches the same subjects. I studied cryptography.
Fact is, if we where to change jobs for say a month, I'd be able to do basically perform his job, while he wouldn't know how to even start doing any of the things I do in my job.
This doesn't by itself imply that history is easier than crypto. It just means that crypto is less forgiving -- if you know little about it it is not possible to fake being half-decent in it. While even if you know literally nothing about history, you'll be able to fake being a teacher of it pretty well simply by always reading a few chapters ahead of what you're currently teaching, afterall the ones you teach also doesn't know much.
In general, areas where there is precisely one very obviously correct answer, and no two ways about it are harder to fake than areas where the answer is fuzzy, open to debate and generally not so much "wrong" or "correct" as more or less good.
You see it in grades too. In maths there are more people who fall trough, than in history. That ain't nessecarily because math is "harder", it migth also be because when someone does not know math, it is very obvious and no amount of handwaving gets around the fact that 2 + 2 is not, infact 5.
Recurring revenue, in addition to up-front revenue. Perhaps also the possibility of controlling (read destroying) the second-hand market aswell as "turning off" old games that are no longer profitable.
Most people I know are simply not interested in any model where they're supposed to continue being nickeled and dimed after having bougth a game. It's just bad value.
You can buy a new game for like $50, and if you get tired of it after a while simply resell it. If you don't absolutely *have* to have the game the very first month it's out, (Actually, what is a fun game in say january 2004 is very likely to provide you with the same amount of fun if you choose to play it 6 months later) you can buy it used for maybe $25.
It's not that bad for game-companies either. Knowing that I can always resell a game if it doesn't appeal to me makes me much more likely to impulse-buy a unknown game. And the money I occasinaolly regain by selling older games are mostly reinvested in newer ones.
The bogey-man in the bushes is what the media cry about, mostly, but most abused children are still abused by parents, grandparents, uncles, aunts, friends of the family, whatever.
That ain't an argument against families. Just as the fact that the internet enables communications between everyone *including* the small minority that are child-abusers isn't really much of an argument against the internet.
No. If you actually read the article he does not even claim this. What he claims is that if you join the top 6 channels on each of the top ten networks (not the same as the "top 60 channels"), log the "conversation" for 36 hours, grep this for "Microsoft", "Norton", "Jasc", "Symantec" then of the resulting lines of text, 99.9% will appear to be warez-trafficking. (i.e. only 0.1% of those lines are people legitimately discussing these companies/software.
This is no analysis of IRC. It's not even an analysis of the top channels in IRC. It's only an analysis of 6 words used on the top 6 channels in the top 10 networks.
With similar research you could prove that there is no warez on irc:
It doesn't take a genius to see that the result would be rather different. This is just an example that when you're allowed to freely choose your parameters, you'll arrive at whatever conclusion you set out to "proove".
Have the velocity decay. For applications like this, "writing numbers in the air" any acceleration that is more than something like 3-5 seconds "old" is irrelevant.
This *migth* still give problems if you try to use this in say a moving vehicle in city-traffic or very bumpy road, but it'll *certainly* not be thrown off just 'cos the phone thinks it's still moving since yesterday.
If you're trying to figure out if the user just wrote "3" or "5" it's simply completely irrelevant what accelerations took place yesterday. Now that I think about it you'll probably want even faster decay, essentially any movement that took place before the time needed to enter a single digit is irrelevant, that's probably more like one second tops.
The complainers claimed that the sticker unfairly singled out evolution among literally thousands of other scientific theories that are equally much "just a theory".
This is perfectly true. Evolution is no less proven than most other scientific theories. Infact evolution has seen an *enormous* amount of time and energy spent trying to discredit it, without success. That's the best any scientific theory can hope for.
Noone claims that evolution, or any other scientific theory is "absolute proven truth", all we can claim is that these are theories which match all observed evidence better than any other theory we where able to think of.
Just because you call it "TV" doesn't make it more amazing. What's really the difference between a "monitor" and a "tv" other than that the tv has a tuner anyway ?
Default to blocking outbond connection port 25, but provide an option, say in your typical web-based configuration-interface say "Use third-party smtp-server" or something.
That way the 1% of people who need it can turn it on, and the 99% of people who never even knew about this can be prevented from having their computers spew spam at the rest of us.
If you want to fit 1000Mbit in 100Mhz, you'll need to encode 10 bits on each period, which means your signals must be 2**10 == 1024 times stronger than the noise.
That's about 30db. (10**3 == 1000)
Doesn't tell you much aslong as you know nothing about the noise in that band on the lines offcourse.
Sorry, but no.
There is no reason why you cannot modulate several bits onto each phase of a signal using a combination of amplitude and phase-modulation.
Here's a clue for you: 56K modems are possible. This is *not* because the telephone-system transfers frequencies up to 100Khz. Infact the phone-system cuts off at aproximately 4Khz.
The possible transmit-rate is limited by the width of the band *and* by the signal to noise ratio. A bit simplified, if your signal is 128 times as strong as the noise, you can distinguish up to 128 different signals from another, so you can send 7 bits at a time.
The reason I suggested returning the signed public-key is that that would make it possible for entities to trust oneanother, if they so choose.
For example, the public library could choose to trust that I am who I say without further verification because say the tax-department has signed my public key.
Thus with a public-key signed by multitude highly trusted entities, you would sometimes be able to identify for low-security applications without further steps.
First, using one of 8 algorithms randomly provides at *best* 3 extra bits of keyspace to search (2^3 = 8), you could gain the same by adding 3 bits to your key.
Secondly, if flaws are found in *any* of the 8 algorithms (8 times as likely as if you used only one) you'll instantly be able to break 1/8th of the encryptions.
Thirdly it is not always impossible to tell from ciphertext alone what algorithm is likely used. Especially not if you occasionally have the possibility of doing known or even choosen plaintext attacks.
Fourth, implementing 8 different algorithms makes your program more complex. You'll have more chanses of doing mistakes. In general complexity is the enemy of security.
Fifth, either you include info about what algorithm is used, or you force the receiver to brute-force it by testing with all 8 algorithms until it finds one that works. If the first, then you've defeated the purpose of using 8 in the first place, if the second, you've just multiplied the work of the receiver by a worst-case of 8, average of 4.
Sixth, using the same key with multiple different algorithms isn't generally always secure. Sometimes using the same key in different algorithms can reveal info about the key or weaknesses that wouldn't be there if you used it in only 1. You could also have 8 different keys, but then you've increased key-handling headaches.
Seventh, I'm bored now, but I could list another dozen or so reasons why this is a fundamentally bad idea if I wanted to. If you cannot, this only means that you're not understanding enough of what you are trying to argue.
Remember: Anyone can make a cryptosystem that he himself cannot break.
Single sign-on is desireable. But it needs to be desentralised, and someone needs to think very carefully about what exactly happens when that single sign-on gets compromised.
One simple idea would simply be to have each user equipped with a secret/public keypair. Sigining up to a site would then consist of having the site in question somehow validate whatever they want to validate about you, and thereafter sign your public key.
Login would go somewhat like this:
User presents signed public key and a signed text of type "Login UserName my.current.ip timestamp" to the website.
Website verifies that it's own signature on the key is valid.
Website verifies that the users signature on the message is valid, and that timestamp and current.ip match the reality. Loosing your secret key would be a disaster, but that is true for *any* system with single signon. You can try to prevent this for example by storing the secret key on a smartcard that does it's own signing of messages. (i.e. no API exists for extracting the key from the smartcard, only an API for saying to the card: "Please sign this with the secretkey")
That'd still not be inhackable, but atleast you'd need physical posession of the card and some funky hardware to be able to learn the secret-key.
It's just that, for whatever reason, it seems there are more boys who don't care what their peers thing about their pursuits than girls. Somehow the opinion of the other girls in the class counts more for the girls.
Encouraging girls to think for themselves, make their own choises and do less group-think would indeed be a good thing. I don't mean only technology.
Doing what your everyone else expects of you (or what you imagine they expect, which isn't always the same thing) isn't more important than doing what you *want* to do.
Theatre:
If instead I rent a movie on DVD, then:
Basically the only reason anyone ever watches stuff in the theatres is if they a) Absolutely want to watch the movie now, and not in a year when it comes in rental. and b) are so honest they won't download the thing.
First, the fact that md5 is a published algorith does *not* imply that finding different files with identical hash is easy. Yes, you can just try different files until two matches, this will happen at the latest when you've tested 2^hashsize+1 files, and on the average when you've tested sqrt(2^hashsize) files. That would take forever.
Your suggestion to randomly check say 10% of the file is dumb. This would mean if I changed some small thing in a file so that the file became garbled (or malicious) then in 90% of the cases you wouldn't discover it. It would also mean that you add a 10% overhead to all your downloads. (no show-stopper, I agree, but also not desireable)
The solution is to use a secure hash. md5 has been suspect for quite some time now, sha1 is the current favourite. A secure hash must have the following properties:
If we didn't have hashes with these properties we could toss away PGP, infact pretty much forget about digital signatures alltogether (since signing a complete message is very computationally expensive, what is typically done is signing the *hash* of a message). Same with SSL, SSH, and infact something like 95% of all cryptographic products in use today.
Luckily there are secure hashes. Or atleast hashes we have no idea how to break. Ofcourse in principle sometime in the future a break for some of those hashes can be found, in which case we'll need to design better ones.
But there's no indication that this is a eternal arms-race. There are problems that where thougth to be "hard" 30 years ago which are still pretty much equally hard. For example, factoring large primes is still a hard problem now, and there's no indication that will ever change. Some algorithmic progress is being made, and ofcourse Moores law means we'll need a few extra bits of hash every decade, but nothing I'd worry about.
The actual study says they've found a correlation, the braindead /. editor writes (or accepts, whatever) a title which would indicate causation.
Say it again boys and girls, real loud, maybe even the editors will hear it; Correlation is not causation.
my actual point, which I assume you agree with is that it is not really relevant if you knew you where infringing, nor if the infringing method was programmed by yourself or is simply a part of a program you bougth from an outside source.
But there's a big difference between:
there is no legal basis for Kodak to sue end users over their use of the JRE or JDK. End users did not infringe upon Kodak's patents
which is a direct quote from the article and:
As an end-user you are fully liable, but unless you have very deep pockets, are disliked by the patent-holder, are a very high-profile user, or the patent-holder is having a particularily bad day, they are probably not going to choose to sue you.
which, as you point out is the reality.
The first claim, the one in the article, that no legal basis exists to sue end users is simply wrong. When an author displays such ignorance of patent-law it weakens his credibility overall.
there is no legal basis for Kodak to sue end users over their use of the JRE or JDK. End users did not infringe upon Kodak's patents -- they downloaded the Java software in good faith that it was perfectly legal, and they presumably abided by the license terms. Kodak would have absolutely no right to try to recover any damages from an end user or anyone else who was not a party to adding the allegedly infringing code to the Java source code.
We probably all wish for it, but that is not how patent law infact works.
Using something in good faith is no defence against a patent-lawsuit. Neither does it save you that the patented algorithm was added to the software you use by someone else, without your knowledge.
If this was a valid defence, then most Linux-users would also be equally safe, afterall they *also* tend to use Linux in good faith, abiding by its license terms, and they *also* had the hypotetical patented technique added by someone else without their knowledge.
Sadly, that's not how patent-law works. There are basically only 3 relevant questions in a patent-infringement-lawsuit:
If the answers to those are yes, yes and no, then you are guilty. Even if you didn't *know* the patent existed. Even if you had absolutely no idea that your software was doing this. Even if the software infringing on the patent was written by someone else. Hell, even if the software is closed-source and you thus reasonably *couldn't* know that it was doing this. Those are all irrelevant.
The site would probably be selected for good, stable weather anyway, and the most likely result of unsuitable weather would be the climbers need to slow down, in extreme cases, like the lasers going offline for whatever reason the climbers would need to stop and wait for them to return.
Climbers carrying people would certainly have some mechanism for safely climbing back down even without external power in any case.
So yes, weather is a factor in this scheme, but I don't think it's a critical factor.
Luckily you are wrong about just about everything else. For example:
Gravity is not a constant over the entire 36.000 km to geosynchronous orbit. Rather it decreases with the *square* of the distance from the earths center. This alone makes your estimate aproximately 1 order of magnitude too high. So, cost of electricity would be $13, not $135.
Secondly, electricity is cheaper in larger quantities. Perhaps you can get another factor of 2 or so here.
Thirdly it is quite likely that the space elevator itself will contribute to lowering energy-prices. Space is a pretty ideal place for a solar collector; no atmosphere, no clouds, no nigth, no land-price. Only currently it costs too much to get there.
Add up all this, and energy-price of $1/pound doesnt sound completely out of the question in the longer run, but I agree with you it's unrealistic for a start. (it's just not *quite* as unrealistic as you make it out to be)
Then to the question of the ribbon. It is correct that currently we can only make nanotubes a few mm long. Fortunately we do *not* need to make nanotubes 36000 km long to build a space-elevator.
Rather we'll braid together, in NASAs design they use epoxy aditionally, multiple tubes to make a "ribbon". Sort of like how you can twist together fibers that individually are 10-20cm long into a rope and have that rope have a large fraction of the strength of the individual fibers.
How long the individual fibers need to be, and how the braiding/epoxying must be done for the finished ribbon to be strong enough is an area of intense research at the moment. I've seen some people saying we need individual tubes 1 cm long, while others are saying they'd need to individually be one or even two orders of magnitude more than that. (i.e up to a meter)
Now, making a single molecule 1cm, or even 1m long is certainly a challenge, but its still simpler than making one thousands of kilometres long.
The 40m bridge over the campus-stream is thus certainly *NOT* 5 orders of magnitude away. It's more likely the braids we are making today, consisting of nanotubes a mm or so long are more than strong enough for that job. It's just that noone are likely to build that bridge aslong as the needed nanotube-ropes would cost millions and a suitable steel-cable is available for a few orders of magnitude less.
Don't get me wrong, I also don't think we'll have a working space-elevator, capable of lifting cargo at $1/pound in a decade.
But I do think that the prospects look a fair bit brigther than you make them out to be.
So, it has a energy-storage and a climbing-mechanism, none of which can climb to space, even with improvements. Instead both components will need to be made fundamentally different.
Most serious designs I've seen use energy from an external source, because if you are carrying your own energy on the climber, then you use most of your power to lift the energy-storage. (sorta like rockets are mostly lifting rocket-fuel) Ideas include powerful lasers shining on the thing from below, being converted to electricity by efficient photocells. (cells tuned to a single frequency like laser can be more efficient than full-spectrum cells) The laser will get weaker as the climber gains heigth, but so will gravity and thus the required energy.
For the actual climbing a non-contact method would be preferable, perhaps something involving magnetism. (essentially a vertical maglev) The trick is to manage that without making the ribbon itself much heavier. (and thus more expensive)
It's obvious that a wind-generator slows down the passing air, i.e. makes the wind weaker. Afterall it has to take the energy it delivers from somewhere.
What is pretty hard to believe is that wind-generators are in any way special in this sense.
When we remove forest, and replace it with cropland, we take away a lot of wind-braking. A forest is a more efficient brake for the lower air than any conceivable windmill-density. And we have removed a *LOT* of forest the last few hundred years.
To make this plausible they would have to argue that the net sum of human activites act more to erect brakes for the wind than it does to remove them. This seems a pretty unreasonable conclusion on the face of it. And like they say, extreme claims require extreme evidence.
So, you've managed to inconvenience the reasonable customer that wants to, say, make a copy of a song in an unencumbered format that works on for example the kids DRM-incapable CD-player while doing pretty close to nothing at all to block the person who's learnt the lesson, stopped buying from you and download his songs unencumbered from whatever filesharing-du-jour instead.
I don't think we really disagree that much. My point is that even only one liberated device is capable of producing unencumbered files from encumbered ones, and those unencumbered files can be distributed easily enough.
Stopping most of the copiers most of the time doesn't actually buy you much at all in a world where the people who copy also share with oneanother.
For that matter, there's always the analog-hole. The large majority of music-consumers are happy with 128kbps crappily-encoded mp3s. Those will also settle for a well-produced analog-hole copy. And again, not everyone needs to do this, it's enough that *one* person does it and shares the resulting unencumbered file somehow.