but they didn't add everyone who called abroad to the no-fly list. That's the part I find hard to imagine being possible in todays USA. Despite being very critical of your current administration. (I actually don't know anyone over here that are enthusiastic about your Mr Bush)
But that's partly because people are capable of basic math. A 10% chance of dying, in exchange for immunity from dying from gunshots is simply a bad deal. People living normal stable lives in western countries without war has like 1 chance in 100.000 to die by being shot by a gun.
Swapping a 0.001 % risk of death against a 10% risk of death is simply dumb.
We do make such swaps though, when the odds are in our favour. Two examples will suffice.
Immunization saves a lot of lives. It also kills a few children, makes a handful of children permanently disabled, and make many more children temporarily sick. Swapping a 1% chance of dying of say polio with a 0.0001% chance of dying from the vaccine, a 0.001% chance of being permanently harmed by the vaccine and a 1% chance of getting temporarily sick from the vaccine is by most (excluding some US nutcases basically) regarded as a good deal.
In wartimes, and shortly before, its common to up the realism-level of the training the soldiers go trough to levels where soldiers actually start dying in training. It's considered a reasonable choice nevertheless, because the more realistic training *may* kill 1/10.000 of your soliders. But it'll save a lot more lives than that once the real war starts. Have a look at the loss-numbers in training prior to the allied invasion of normandie for an example.
Similarily, there are parts of the world where vaccine agains HIV would be a benefit even if the side-effects where severe. If you live somewhere where 30% of the adult population is infected, it's probably a good deal to get vaccinated at oh say age 12, even if the vaccine say makes 10% sick for a week, 0.1% permanently harmed and 0.01% die as a result of the vaccine. If you live in a place where HIV is rare, that'd probably be a bad deal.
I don't know about american cars, but what you're claiming is certainly not true for Japanese/European cars.
On the contrary, they are orders of magnitude more reliable, require less service, and go more years/miles before being worn out.
When I was young normal cars required an oil-change and basic service every 5000 miles, it was perfectly normal for the clutch to be worn out at 20000 miles, same for the register. A car that had 75000 miles on it after say 8 years was considered as near-scrap, many cars where scrapped earlier than that due to expensive repairs. Corrosion was a major problem (I livein western norway, it's wet and salty much of the year), many cars literally got holes in their floors in 5-7 years.
Today oil-change is only required every 10000 miles, the cars drink half as much petrol for the same performance, are much safer, enormously much more reliable. People expect to buy a new car, give it basic service once a year, and have it work pretty darn close to 100% of the time. Our previous car, a dirt-cheap one, went for 150.000 miles and 12 years with a total of *2* times having problems that made it un-drivable, none of those in the first 8 years.
Clutches frequently outlast the car, atleast if it's not a muscle-car and you don't drive very agressively, and corrosion is so seldom that most manufacturers give a 5-8 year *warranty* against it. My brother who works as an auto-mechanic has stopped recommending anti-corrosion undercarriage-treatment. This used to be a no-brainer. These days there's just little point.
But just how would the US government go about "blanket-wiretapping" a swedish ISP ?
Besides, even if they somehow acomplished this, I doubt that the USA is *that* close to a dictatorship.
It is true that compared to what most europeans are used to, and even more compared to what scandinavians are used to, your transparency, oversigth and democracy is utterly ridicolously weak, but Bush is still not quite a dictator for life. (even though he seems to claim rigths such as starting wars and wiretapping US citizens without a warrant, and getting away with it)
I could be wrong. Maybe you guys are even more screwed that I imagined.
To figure out that someguy@gmail.com was using the service, you'd first need to subpoena google for the ip-adresses of the person accessing that account. You'd get the ip-adress of the anonymizer.
Then you'd need to subpoena the anonymizer, in Sweden, which only works if you've got proof of a crime with a (in sweden) penalty of 2 years prison or more. And even then, you cannot subpoena information which the anonymizer simply do not have -- because they deliberately don't keep any logs.
So, just how would you go about adding people to no-fly lists based on the usage of an anonymizer such as this ?
First, I explicitly said some reasons where noble, and some less so. That's perfectly fine with me. I don't *want* to live in a world where only noble actions are possible.
Anonymity can also be used to spread slander, to threaten or intimidate someone without being traced, to further pump-and-dump frauds in the stock-market, to organize criminal activity and so on.
I still prefer to live in a world with strong anonymity. We've had that for a long time. It's bloody damn hard to track who the sender of a letter is. It's near-impossible to figure out who calls in a tip with a greymarket prepaid mobile phone.
Yes. It helps criminals. So does cars, houses, telephones, indeed *everything* that is useful is useful also to those with illegal intentions. That's no argument at all in my opinion.
Consider yourself excused. Happens to the best of us. A friend of mine made that particular gaffe at the oral examination for his master thesis in cryptography. (on improvements to elliptic curve cryptography making the keys more compact without reducing the computational complexity of breaking them)
He still managed to somehow utter that 2^128 was twice as large a number as 2^64. Didn't even fail him, as closer asking showed that he didn't really believe that, it's just that it sorta "looks" that way, cause everyone knows 128 is twice as large as 64.
Sure there are going to be "allegations". So what ?
Telephones, letters, ads in the New York Times, cars, bicycles and pigeons are also all usable (and doubtlessly used!) by people with nefarious purposes. So what ? That's neither here nor there, certainly not a reason to forbid any of them.
There's many reasons, some noble, some less noble. Anyone that wants to can easily think of reasons one migth want to be anonymous on the internet.
For example;
You may want to blow the whistle on some illegal or unethical behaviour where you work, but fear losing your job.
You may want to critisize government, but fear negative consequences. (depending on where you live this may be from none up to and including imprisonment, torture or execution)
You may want to lobby for an unpopular position.
You may want to send email to your friends in the near east (I've got several friends in the UaE and Saudi-arabia) without ending up on the "no-fly" list.
You may want to help the police figthing corruption or mafia without yourself or your family ending up dead.
You may want to post anonymized nude photos online (for fun or profit) without exposing your identity.
If you're a boss, you may want to pose as an "outsider" and f.ex. contact your own support-department to get a picture of how said department works, as seen from the outside.
You may not *want* people to know you're collecting pictures of rhinos.
You may want to be able to discuss personal problems online without those you discuss it with knowing who you are.
You may want to look for love online without risking stalkers.
There's a zillion reasons really. But more importantly, you shouldn't need any reason at all. The simple fact is, there exist people who would prefer, atleast sometimes, being anonymous online.
Terrorism isn't even one of the top ten problems we face as a nation.
Certainly not. Probably not even in the top 100 if you measure it along, for example, the axis of life-years lost.
Diabetes, car-accidents, circulatory problems and cancer all kill more people a *week* than terrorism kills in a *year*.
I'm not saying we shouldn't try to reduce terrorism -- sure we should.
All I'm saying is that we shouldn't give up all rational thought, panic and throw the baby out with the water. Giving up on fundamental liberties for the reason of marginally (if at all) more effectively catching terrorists is not a good deal.
Actually, just stopping military support to israel would do tons more if the goal was having less americans blown up.
Thats pretty close to how it works in Norway. For marketing of any sort adressed directly to you. There is a single govnerment-maintained list where you can opt to not receive direct marketing.
Companies that do direct marketing send their lists in, and get them back without those persons who have opted out. They learn nothing new about you in the process, other than the fact that you've opted out.
For electronic marketing (email, sms, fax) it's opt-in rather than opt-out. In other words, they cannot legally do it unless you've given prior, informed consent to that. The logic is that this in this type of marketing, the recipient typically pays a large part of the cost. Marketers are less likely to abuse say paper-based marketing as that actually costs them to print and distribute. (compare the quality of the marketing in the average paper-based marketing and the average spam you receive to see what I mean..)
For unadressed "distributed to all" marketing there's a small sticker you can put on your mailbox, and you won't get any.
In short, you can eliminate receiving any marketing by following 3 simple steps:
Register yourself to opt-out of direct marketing. (one phone-call or one visit to the opt-out list.
Do not agree to receive direct marketing when companies ask.
Excel is mindbogglingly dumb from a CompSci or mathemathical POV. It's incomprehensible to make a spreadsheet and then fail to understand even the basics of floating-point arithmethic.
Demonstrating it to the customer shouldn't be a problem. Make a *trivial* Excel-spreadsheet that gives an obviously wrong result, you need nothing more than a table of numbers and a single call to SUM() for this.
Used to work correctly in 123 for DOS 6.22 by the way, (and works correctly in OpenOffice Calc) so it's not as if this is something new....
Sure. There's many ways of solving the problem, none of them very hard to grasp once you've done the fundamental task: which is to be aware of and understand the problem.
Lots of programmers though, are unaware of the finer details of floating-point numbers.
As evidenced by MS-Excel failing to give the correct answer, even when as we've now demonstrated, there's multiple, simple, correct algorithms for doing so.
It *is* surprising to do the equivalent of 1000 + (1+1+1...[10000 times] +1) and get 1000 as the answer. The answer *should* be on the order of 11000. Anything else is a bug.
Thus we can conclude that MS-Excel does not even manage to get its SUM() function correct. One of the simplest functions there can be on a set of floats.
It's worse than that. Actually with 2^512 kernels and 2^128 possible hash-values there are (about) 2^384 kernels matching each possible hash. That is 39402006196394479212279040100143613805079739270465 44666794829340424572177149721061141426625488491564 0806627990306816 which you will notice is *sligthly* more than your "4".
There's other funkyness too, besides the precision. For example, if you're adding up a lot of floating-point numbers, it makes a difference what sequence you do the additions in.
For example, if your input consist of one large number, and tons of small ones, then rounding-errors mean that starting with the large number gives a much smaller result than starting with the small ones.
If I scale it down to smaller numbers, you see why:
1.0*10^5 + 1.0*10^1 = 1.0*10^5
So, adding a "small" number to a "large" number gives you simply the large number.
If you repeat this, a million times, your result is still simply the large number.
So you could end up concluding that 1.0*10^5 + (1.0*10^1 + 1.0*10^1..[1000000 times]...) = 1.0*10^5
That is an order of magnitude wrong. The correct result is 1.1*10^6
Practical result ? You need to think about your input. If it *may* look like this, you need to add up by repeatedly adding the two smallest numbers. Easy to do with a priority-tree. pseudocode like this:
Insert all numbers in priority-tree.
Extract two smallest numbers from tree.
Add the two numbers, producing a new number.
Push this single new number into the tree.
Repeat from step 2 until you're left with a single number.
MS-Excel, by the way, does *NOT* do this in it's SUM() function, if you feed it a "large" number and *many* "small" numbers, you get horrendously wrong results. Because of the relatively high precision of floats and doubles though, you need to use larger numbers than in my example here.
It's not more "accurate". It's just that the inputs and outputs are often decimal, in which case using something other than decimal can give "unexpected" results.
For example for finance, floating-point is useless, people generally do something like use a single int to store number-of-cents.
The issue ain't accuracy per se, it's accuracy with *certain* numbers (thous representable in base10).
In a financial program people expect $0.40 * 1000000000 to come out as *precisely* 400000000 and not 399999999.99
There's lots of numbers that base10 can't accurately represent, such as 1/3. The thing is, you'll never have those as inputs if the inputs are stuff like, for example, prices.
government, at least in some form, is absolutely necessary. How do I know? I'm human. And by and large, humans are greedy, amoral, unethical creatures that left to their own devices lie, steal, cheat, murder and rape their way through life.
No actually. It's more like some small fraction of humans will behave like that some small fraction of the time. Which means your conclusion is rigth, we do need some sort of concerted response to certain behaviour, we tend to name that response "government".
Notice I've friended you -- I don't mind people disagreeing with me. Infact I very much *appreciate* people with other viewpoints, aslong as they're able and willing to argue sensibly for them.
It's just that I've had the "government is evil" discussion one too many times. I've bored of it. Sorry. You can keep your police/courts/military only government.
Agreed. But, on the other hand, having cameras work that way is an excellent idea.
If police knew that cameras tended to upload the pictures to secure storage unaccessible to them seconds after they pictures are taken, they'd be a lot less interested in confiscating the cameras in the first place. (and if they did, the picures already taken would be safe)
No really. I assume no such thing. Let's play with your example. You have 4 delays. Let's say the accuracy in timing is that your stated numbers (and those we could measure) are accurate to the nearest 10ms.
Let's say I want to encode 8 bits of secret information on these 4 delays. I want to send the bitstring '11010010'.
First, we discard the on the end of the timings as meaningless.
That changes the sequence to 10 11 9 4
Since I want to send 2 bits on each delay, I need 4 different values (00 01 10 11), so I use mod4, and map so that 0=00 1=01 2=10 and 3=11.
I use delays to acomplish this. I want 3,1,0,2 because that maps to '11010010' which is what I want to transmit.
The original was: 100 110 90 40
I delay that to get: 110 130 120 60
To decode that, remove the 0s as insignificant and you have 11 13 12 6. Use mod4 and you get 3,1,0,2 replace those with their binary equivalent and voila: '11010010'
In this example, the delays where all lengthened. That was made to make the explanation simpler for you. In real life it could just aswell shorten. This seems counterinituitive, but really, trust me, it goes like this:
Original timings (timings, not delay) 100 500 1000 (which means delays are 400 500)
Imagine you need the first delay to be *500* rather than *400*, you do this by delaying the packet containing the second keypress, you now have keypresses at 100 600 1000. Notice how the second gap is now only 400. Less than it originally was. If 400 matches the delay you want (a 1:4 chance if you're encoding 2 bits for every keypress) then you need not do anything to the third keypress.)
Frankly, I couldn't care less what you think of our education-system. Norway is a democratic country. Most people living in Norway are quite happy with the current education-system. If it is contrary to the religious beliefs of some US minority or not is irrelevant.
So, with a sensible document-format, all you need to exchange documents is any compliant software on either end.
With MS-Office, on the other hand, it's not enough that all participants have some MS-Office compliant software. It's not enough that they buy Microsoft Office, the very same office-suite. It's not even enough that they buy "MS-Office 2007", no, that's not enough to ensure compatibility.
It needs to be: "MS-Office 2007, running on MS-Windows, variant for 32bit Windows"
A naive user would expect software that saves in a format that the software itself refers to as (for example) "Microsoft Office Excel Workbook" to be openable by any software that is "Microsoft office Excel"
A sophisticated user would know this is a lie, and that in reality they mean "Microsoft Office Excel 2003 Workbook"
Instead of improving, this shows that in the future, that must be read as: "Microsoft Office Excel 2007, for x86, running MS-Windows, Workbook"
Anything less, and support is at best incomplete. Ridicolous. Completely ridicolous.
Slashdot Mirror
but they didn't add everyone who called abroad to the no-fly list. That's the part I find hard to imagine being possible in todays USA. Despite being very critical of your current administration. (I actually don't know anyone over here that are enthusiastic about your Mr Bush)
But that's partly because people are capable of basic math. A 10% chance of dying, in exchange for immunity from dying from gunshots is simply a bad deal. People living normal stable lives in western countries without war has like 1 chance in 100.000 to die by being shot by a gun.
Swapping a 0.001 % risk of death against a 10% risk of death is simply dumb.
We do make such swaps though, when the odds are in our favour. Two examples will suffice.
Immunization saves a lot of lives. It also kills a few children, makes a handful of children permanently disabled, and make many more children temporarily sick. Swapping a 1% chance of dying of say polio with a 0.0001% chance of dying from the vaccine, a 0.001% chance of being permanently harmed by the vaccine and a 1% chance of getting temporarily sick from the vaccine is by most (excluding some US nutcases basically) regarded as a good deal.
In wartimes, and shortly before, its common to up the realism-level of the training the soldiers go trough to levels where soldiers actually start dying in training. It's considered a reasonable choice nevertheless, because the more realistic training *may* kill 1/10.000 of your soliders. But it'll save a lot more lives than that once the real war starts. Have a look at the loss-numbers in training prior to the allied invasion of normandie for an example.
Similarily, there are parts of the world where vaccine agains HIV would be a benefit even if the side-effects where severe. If you live somewhere where 30% of the adult population is infected, it's probably a good deal to get vaccinated at oh say age 12, even if the vaccine say makes 10% sick for a week, 0.1% permanently harmed and 0.01% die as a result of the vaccine. If you live in a place where HIV is rare, that'd probably be a bad deal.
On the contrary, they are orders of magnitude more reliable, require less service, and go more years/miles before being worn out.
When I was young normal cars required an oil-change and basic service every 5000 miles, it was perfectly normal for the clutch to be worn out at 20000 miles, same for the register. A car that had 75000 miles on it after say 8 years was considered as near-scrap, many cars where scrapped earlier than that due to expensive repairs. Corrosion was a major problem (I livein western norway, it's wet and salty much of the year), many cars literally got holes in their floors in 5-7 years.
Today oil-change is only required every 10000 miles, the cars drink half as much petrol for the same performance, are much safer, enormously much more reliable. People expect to buy a new car, give it basic service once a year, and have it work pretty darn close to 100% of the time. Our previous car, a dirt-cheap one, went for 150.000 miles and 12 years with a total of *2* times having problems that made it un-drivable, none of those in the first 8 years.
Clutches frequently outlast the car, atleast if it's not a muscle-car and you don't drive very agressively, and corrosion is so seldom that most manufacturers give a 5-8 year *warranty* against it. My brother who works as an auto-mechanic has stopped recommending anti-corrosion undercarriage-treatment. This used to be a no-brainer. These days there's just little point.
Besides, even if they somehow acomplished this, I doubt that the USA is *that* close to a dictatorship.
It is true that compared to what most europeans are used to, and even more compared to what scandinavians are used to, your transparency, oversigth and democracy is utterly ridicolously weak, but Bush is still not quite a dictator for life. (even though he seems to claim rigths such as starting wars and wiretapping US citizens without a warrant, and getting away with it)
I could be wrong. Maybe you guys are even more screwed that I imagined.
To figure out that someguy@gmail.com was using the service, you'd first need to subpoena google for the ip-adresses of the person accessing that account. You'd get the ip-adress of the anonymizer.
Then you'd need to subpoena the anonymizer, in Sweden, which only works if you've got proof of a crime with a (in sweden) penalty of 2 years prison or more. And even then, you cannot subpoena information which the anonymizer simply do not have -- because they deliberately don't keep any logs.
So, just how would you go about adding people to no-fly lists based on the usage of an anonymizer such as this ?
Anonymity can also be used to spread slander, to threaten or intimidate someone without being traced, to further pump-and-dump frauds in the stock-market, to organize criminal activity and so on.
I still prefer to live in a world with strong anonymity. We've had that for a long time. It's bloody damn hard to track who the sender of a letter is. It's near-impossible to figure out who calls in a tip with a greymarket prepaid mobile phone.
Yes. It helps criminals. So does cars, houses, telephones, indeed *everything* that is useful is useful also to those with illegal intentions. That's no argument at all in my opinion.
He still managed to somehow utter that 2^128 was twice as large a number as 2^64. Didn't even fail him, as closer asking showed that he didn't really believe that, it's just that it sorta "looks" that way, cause everyone knows 128 is twice as large as 64.
Telephones, letters, ads in the New York Times, cars, bicycles and pigeons are also all usable (and doubtlessly used!) by people with nefarious purposes. So what ? That's neither here nor there, certainly not a reason to forbid any of them.
For example;
There's a zillion reasons really. But more importantly, you shouldn't need any reason at all. The simple fact is, there exist people who would prefer, atleast sometimes, being anonymous online.
Certainly not. Probably not even in the top 100 if you measure it along, for example, the axis of life-years lost.
Diabetes, car-accidents, circulatory problems and cancer all kill more people a *week* than terrorism kills in a *year*.
I'm not saying we shouldn't try to reduce terrorism -- sure we should.
All I'm saying is that we shouldn't give up all rational thought, panic and throw the baby out with the water. Giving up on fundamental liberties for the reason of marginally (if at all) more effectively catching terrorists is not a good deal.
Actually, just stopping military support to israel would do tons more if the goal was having less americans blown up.
Companies that do direct marketing send their lists in, and get them back without those persons who have opted out. They learn nothing new about you in the process, other than the fact that you've opted out.
For electronic marketing (email, sms, fax) it's opt-in rather than opt-out. In other words, they cannot legally do it unless you've given prior, informed consent to that. The logic is that this in this type of marketing, the recipient typically pays a large part of the cost. Marketers are less likely to abuse say paper-based marketing as that actually costs them to print and distribute. (compare the quality of the marketing in the average paper-based marketing and the average spam you receive to see what I mean..)
For unadressed "distributed to all" marketing there's a small sticker you can put on your mailbox, and you won't get any.
In short, you can eliminate receiving any marketing by following 3 simple steps:
Demonstrating it to the customer shouldn't be a problem. Make a *trivial* Excel-spreadsheet that gives an obviously wrong result, you need nothing more than a table of numbers and a single call to SUM() for this.
Used to work correctly in 123 for DOS 6.22 by the way, (and works correctly in OpenOffice Calc) so it's not as if this is something new....
Lots of programmers though, are unaware of the finer details of floating-point numbers.
As evidenced by MS-Excel failing to give the correct answer, even when as we've now demonstrated, there's multiple, simple, correct algorithms for doing so.
It *is* surprising to do the equivalent of 1000 + (1+1+1 ...[10000 times] +1) and get 1000 as the answer. The answer *should* be on the order of 11000. Anything else is a bug.
Thus we can conclude that MS-Excel does not even manage to get its SUM() function correct. One of the simplest functions there can be on a set of floats.
It's worse than that. Actually with 2^512 kernels and 2^128 possible hash-values there are (about) 2^384 kernels matching each possible hash. That is 39402006196394479212279040100143613805079739270465 44666794829340424572177149721061141426625488491564 0806627990306816 which you will notice is *sligthly* more than your "4".
For example, if your input consist of one large number, and tons of small ones, then rounding-errors mean that starting with the large number gives a much smaller result than starting with the small ones.
If I scale it down to smaller numbers, you see why:
1.0*10^5 + 1.0*10^1 = 1.0*10^5
So, adding a "small" number to a "large" number gives you simply the large number.
If you repeat this, a million times, your result is still simply the large number.
So you could end up concluding that 1.0*10^5 + (1.0*10^1 + 1.0*10^1 ..[1000000 times]...) = 1.0*10^5
That is an order of magnitude wrong. The correct result is 1.1*10^6
Practical result ? You need to think about your input. If it *may* look like this, you need to add up by repeatedly adding the two smallest numbers. Easy to do with a priority-tree. pseudocode like this:
MS-Excel, by the way, does *NOT* do this in it's SUM() function, if you feed it a "large" number and *many* "small" numbers, you get horrendously wrong results. Because of the relatively high precision of floats and doubles though, you need to use larger numbers than in my example here.
For example for finance, floating-point is useless, people generally do something like use a single int to store number-of-cents.
The issue ain't accuracy per se, it's accuracy with *certain* numbers (thous representable in base10).
In a financial program people expect $0.40 * 1000000000 to come out as *precisely* 400000000 and not 399999999.99
There's lots of numbers that base10 can't accurately represent, such as 1/3. The thing is, you'll never have those as inputs if the inputs are stuff like, for example, prices.
No actually. It's more like some small fraction of humans will behave like that some small fraction of the time. Which means your conclusion is rigth, we do need some sort of concerted response to certain behaviour, we tend to name that response "government".
My website is currently irrelevant. 2004 I think :-)
Notice I've friended you -- I don't mind people disagreeing with me. Infact I very much *appreciate* people with other viewpoints, aslong as they're able and willing to argue sensibly for them.
It's just that I've had the "government is evil" discussion one too many times. I've bored of it. Sorry. You can keep your police/courts/military only government.
changing ips *AND* blocking of cookies is needed to avoid leaving a single continous trail at any site you visit regularily.
Transparency is absolutely crucial if a democracy shall be able to control a government. You guys basically don't have that.
Transparent my ass....
If police knew that cameras tended to upload the pictures to secure storage unaccessible to them seconds after they pictures are taken, they'd be a lot less interested in confiscating the cameras in the first place. (and if they did, the picures already taken would be safe)
Let's say I want to encode 8 bits of secret information on these 4 delays. I want to send the bitstring '11010010'.
First, we discard the on the end of the timings as meaningless.
That changes the sequence to 10 11 9 4
Since I want to send 2 bits on each delay, I need 4 different values (00 01 10 11), so I use mod4, and map so that 0=00 1=01 2=10 and 3=11.
I use delays to acomplish this. I want 3,1,0,2 because that maps to '11010010' which is what I want to transmit.
The original was: 100 110 90 40
I delay that to get: 110 130 120 60
To decode that, remove the 0s as insignificant and you have 11 13 12 6. Use mod4 and you get 3,1,0,2 replace those with their binary equivalent and voila: '11010010'
In this example, the delays where all lengthened. That was made to make the explanation simpler for you. In real life it could just aswell shorten. This seems counterinituitive, but really, trust me, it goes like this: Original timings (timings, not delay) 100 500 1000 (which means delays are 400 500)
Imagine you need the first delay to be *500* rather than *400*, you do this by delaying the packet containing the second keypress, you now have keypresses at 100 600 1000. Notice how the second gap is now only 400. Less than it originally was. If 400 matches the delay you want (a 1:4 chance if you're encoding 2 bits for every keypress) then you need not do anything to the third keypress.)
Frankly, I couldn't care less what you think of our education-system. Norway is a democratic country. Most people living in Norway are quite happy with the current education-system. If it is contrary to the religious beliefs of some US minority or not is irrelevant.
So, with a sensible document-format, all you need to exchange documents is any compliant software on either end.
With MS-Office, on the other hand, it's not enough that all participants have some MS-Office compliant software. It's not enough that they buy Microsoft Office, the very same office-suite. It's not even enough that they buy "MS-Office 2007", no, that's not enough to ensure compatibility.
It needs to be: "MS-Office 2007, running on MS-Windows, variant for 32bit Windows"
A naive user would expect software that saves in a format that the software itself refers to as (for example) "Microsoft Office Excel Workbook" to be openable by any software that is "Microsoft office Excel"
A sophisticated user would know this is a lie, and that in reality they mean "Microsoft Office Excel 2003 Workbook"
Instead of improving, this shows that in the future, that must be read as: "Microsoft Office Excel 2007, for x86, running MS-Windows, Workbook"
Anything less, and support is at best incomplete. Ridicolous. Completely ridicolous.