What you propose will either (a) not work, or (b) get you into deep legal trouble, at least in the US, Canada, and as far as I know, all of the UK.
Actually, it'll work fine and is perfectly legal. There are at least two ways to do it:
1) Transfer the borrowed money into a company. Apply for a loan in company name - the bank will still ask for a personal guarantee, but the entity applying for the loan has a positive net worth.
2) Borrow the money and wait 6 months. After 6 months, the banks don't care where the deposit came from - remember, you're buying property, which serves as it's own collateral. The deposit is merely the banks safety margin, and they have first call on the property in the event you go under anyway.
How do I know this? Because I'm doing strategy number 1 - and I just got preapproval on two mortgages yesterday.
Not that I'd recommend doing it with Credit Cards! There are any number of non-bank institutions that will lend you the money at only a % or so more than the "standard" rate.
If you are actually trying to claim it's a hoax - I'm not sure - then think about the *difficulty* of pulling this off... what follows is quoted from another/. thread long ago. It covers most of the difficulties rather nicely I think.
# For the interest of Slashdot readers, national governments, and any other interested organization, I am posting instructions on how to fake a moon landing and not get caught for 30 years.
Before the Landing
Put out a request for tenders for a contract to build the lunar hardware to major aerospace companies. It would be pretty obvious after the fact that no one had built your launchers and landers. In the contracts, give a specification that would lead the 10,000 engineers who work on the project to reasonably believe that the equipment could be used to land on the moon. Engineers are smart people; they could easily spot holes in your assumptions if you make the requirements less stringent than they have to be. If it were obvious that the hardware couldn't land on the moon, you would be caught. Have the hardward manufactured and delivered. Again easy to spot if this wasn't done, especially for a Saturn V-class rocket and related assemblies. In summary: You would actually have to build stuff that would probably be able to land a man on the moon, with all the associated expenses.
During the missions:
You will actually have to launch the thing you contracted to build. You could launch something else -- but why bother? We've already established that you have to build a moon rocket, and you'd have to pay off everyone who was involved in its destruction and substitution. Besides, it would be big news, so news organizations would want to film the launch of the big rocket. So, the capsule could be suborbital, or stay in orbit, and the rest of the mission could be faked, right? Wrong. Antennas around the world will be tracking the radio signals from the capsule, including the continuous telemetry feeds. Something would have to go to the moon, on a realistic lunar trajectory, or this would be immediately spotted by legions of radio astronomers and HAM radio amateurs around the world, many of whom have advanced signal processing available to them (like Doppler analysis, etc.). They would also be able to tell the difference between a lunar trajectory and a different orbit, like a geosynchronous orbit, because of the moon's particular position in the sky. So, the capsule has to go to the moon. Does it have to orbit? Yes. The capsule must stay in the vicinity of the moon for several days (again checked by those with large radio antennas). The only foolproof way to do that is to orbit. So, the capsule has to orbit. Does it have to land? Yes. While in orbit, the capsule can't communicate with Earth from the far side of the moon. Yet a lander must be able to send continuous telemetry to the Earth. It would be pretty obvious fakery to have the "lander"'s telemetry fade out at the same time as the capsule's. Does it have to come back? Yes; for the same trajectory reason. The return trajectory could be tracked. Does the capsule/lander have to be manned? Not necessarily, but there would be many complications if it weren't. You would have to be able to carry on ground/capsule communications in a realistic manner even though the communications from the capsule would have to be recorded and beamed back (because your radio is being monitored). The "astronauts" would be unable to perform any diagnostic tasks aboard the spacecraft (because they're not aboard it), so the entire flight control team would have to be in on the hoax (dozens, even hundreds, of people to pay off). In summary: You would have to actually send something to the moon, which may as well be manned.
After the Landing
Bring back tons of "moon rocks" and other materials for analysis by independent scientists around the world. These rocks could not be obviously of terrestrial origin, implying some exotic materials science (or creative geology). Either that or pay off anyone who comes in contact with the "lunar samples". And if you're NASA - do this seven times, with one of the seven attempts turning into a remarkably realistic failure.
The upshot: It's equally easy and expensive to actually land a man on the moon than fake it convincingly. Furthermore, the evidence for fakery would not be found in trivial forms of evidence, like photographs, but in more obvious places, like contracts, accounting, radio monitoring, and the lunar samples themselves.
I think Literate programming is a wonderful idea, but I don't think it's a practical one in many (most?) real-world environments.
What like structured programming? Or - God forbid! - Object Oriented Programming! That crap's all just too damn slow to run and to damn difficult to code!
Sure, LP's not perfect now... but if you bother to learn only what's currently in fashion, by the time you've learnt it, the Next Bit Thing has arrived...
That said... you're perfectly correct when you say that no tool - be it OOP, LP, or any other silly acronym - can prevent bad programmers writing bad code - it can only help good programmers write better code.
And that's it. So under US laws the Jedi church would qualify (if it existed in the US).
But you'll notice that a large number of "normal" religions would not... for example:
The organization may not intervene in political campaigns, and
Large segments of the Christian world fail this test... c.f. abortion and the whole Religious Right bull.
No substantial part of its activity may be attempting to influence legislation,
See above.
Net earnings may not inure to the benefit of any private individual or shareholder,
And I guess the Catholic church fails this one too... you can't tell me that the officials in the Catholic church haven't become inured to their benefits... all those US20 million mansions the Cardinals live in... the golf courses and convention centres...
Ahh hell, who am I kidding. Your average American is, and always will be, a religious hypocrite. You'd have more luck reasoning with a rock. (NB: This probably doesn't include the vast majority of/. readers, or geeks in general).
You've got 7 hours during which you can't be interrupted. Seven hours free from your boss. Seven whole hours of entirely interruption free bliss. And you want to use it to play games? Dammit man! Code, or work, or do something productive... those 7 interruption-free hours are worth a week of "normal" work!
And your asking/. for advice?! Come on! If you want an opinion that's not based on someone else's opinion, for the love of God don't do an AskSlashdot!
please, dear god, offer RING BOUND versions of your books!
Hell Yes! And if you can't ring bind them, at least bind them properly - I've had two books (albeit probably not yours) fall apart on me already just *this* year - and we're only one month in!
Re:Chinese Rooms and Software Guys
on
Arguing A.I.
·
· Score: 1
For example, has there ever been a good refutation of Searle's Chinese Room argument?
Yep, and it runs essentially something like this: If it walks like a duck, and quacks like a duck, then it is a duck. That is, to say, while neither the books, nor the person are individually a sentient mind that "knows" Chinese, together they comprise a (sentient) system that does.
As an analogy, lets say I can seperate the neurons that receive visual information, and those that process it into mental pictures. Arguably, on their own, neither set of nerves can "see" in a useful way. Searle's argument would have you believe that when you combine these two sets, what you get is not actually vision, but merely a convincing simulation of it. But if it looks like vision, and works like vision, then I'm afraid I'm gonna call it vision.
For them the mind is completely separate from the brain, an idea that has been pretty much discarded by the thinking public
This is a strawman argument. Dualism argues that the mind is seperate, partially or completely, from the physical world. I'd be interested who "them" are, since this is not a view held by many (any??) strong AI believers - after all, strong AI and dualism do contradict...
I think you'll find that what your actually talking about is a form of materialism, namely the variety that is characterised by the quote "the mind is software that runs on the brain". From this point of view, the mind and the brain are different, but very tightly related, in the same way that a compiled program and a CPU are.
If you want to say that this is in fact dualism, then be my guest. But redefining terms in this arbitrary manner is rather pointless if you actually want to try and debate an issue.
and think: when was the last time you changed a candy bar to a credit card? by itself? there is a point when paying by electronic means is perceived to be too much of a hassle.
Umm... if I can't charge it to credit or a debit card (via EFTPOS or the like) I get annoyed. Hell, I haven't been forced to pay cash in New Zealand for at least 2 years... Occasionally I still do get cash out and end up with small change - but only as a result of the goddamned buses here. (And yes... only the buses. Even taxis have mobile EFTPOS terminals).
This is very much a cultural thing, and frankly, given how much more techie Taiwan, Japan, Korea etc are than even NZ, I wouldn't be surprised at all if you could be as physically cashless as you liked.
Using random data as opposed to zeroes is more secure because writing zeroes may leave a readable residual magnetic signature on the media whereas random data tends to obscure the mag sig.
Actually, this is only a partial solution. Because of little movements in the read/write heads, you actually have to a one, then a zero, then a one... and so on, depending on how securely you want to wipe out the data. It's the flipping of the polarisation of the little bits of oxide back and forth that actually wipes it out... anything else, will as you say, leave a residual magnetic signature which is recoverable with an oscilloscope and very fine motor control (still not easy though!).
But it's sort of in the nature of these things that "near misses" will be very common compared to actual hits. Let's look at the numbers:
The earth has a radius of about 6300 km.
This gives a volume of about 10^12 km^3
This asteroid was at the closest about 830000 km from earth
A sphere with a radius of 830000 has a volume of about 2.5*10^18
If we divide these numbers, we find that an object will be this close to earth on the average something a bit more than 2 million times as often as it actually hits the earth.
Yup, you are simplifying... as I will be shortly. However you are actually being too tricky... the question is not whether the asteroid will intersect the sphere which makes up the earth, but whether it's trajectory will intersect the disk that makes up the earth (you've got one dimension to many I'm afraid!). Hence the chances are more likely to be in the region calculated as follows:
Radius of Earth=6300 km
Therefore the "target area" the earth presents is approximately pi*6300^2 = 125,000,000 sq. km
Closest point of asteroid approach = 830,000 km
Therefore the size of the disk that the asteroid actually hit is approximately = 2,164,000,000,000 sq. km
This gives odds nearer to 17000 to 1 against. Gravitational attraction and the fact that most asteroids lie in or near a plane which intersects the Earth will slightly increase the odds of a collision, but probably not significantly.
And I do agree... these are still long odds. But odds worth doing something about:)
But this does more to help MS than it does Linux, since it will remove yet another barrier to exit for people running Linux on servers. (Run C# and.Net on Linux, and it's easier to convert to Windows.) And remember, MS has concluded that Linux is not a threat on the desktop, but a very serious threat on servers. (I agree with both parts of that, FWIW.)
Oh but of course. And that's why Excel wasn't successful until it was both forwards and backwards compatible with Lotus 123?
And as much as I hate to say it, this also provides ammunition to the people who claim that open source is very good at copying other projects' work, but terrible at innovating. Honestly, of all the high profile open source projects, how many of them are a significant innovation, and how many are merely an attempt to produce an equivalent of feature Z of Windows or Unix or Mac OS on Linux?
Yeah? And what proportion of "high profile" closed-source projects really innovate?
PostgreSQL provides inheritance for tables [postgresql.org] (which is why they call it an object-relational database). I haven't used that feature yet, but it looks perfect for persistent storage of OO data.
Hmmm... very sexy. I will have to look at that... I can see it being incredibly useful for a project I am planning at the moment.
While you may be correct, I can't help but wonder how you are meant to implement any form of generalisation or inheritance in an RDBMS, without a huge mess of tables and complex relationships.
I mention this, because it is something I find myself wanting to do all the time, for example, when storing data that originates in OO programs. Being able to store it in an RDBMS has heaps of advantages for me (primarily that it is easier and less buggy to load and save data) - but I can't easily store the different info of different derived classes.
If you have any suggestions on how to solve this, I'd definitely appreciate it!
Err-hmm...umm uber-no. It is trivial to get the information out of the experts, if you are trained in getting information out of people. Most code-monkeys are not trained to do this properly Moving someone down a decision tree is not hard. It can be time consuming if the process is complex. It its complexity quotient is too high, then you make a design decision not to use expert systems. (Most code-monkeys aren't trained in making decision decisions either.)
I agree with all of what you say. Talking someone through a decision tree is not at all complex, and sometimes expert systems are not the right solution.
The problem comes in constructing the decision tree itself - this is the type of expert knowledge which someone can know, but not enunciate. The situation is also made more complex when you have decision trees with feedback loops, more than one parallel tree, etc. I probably should have been more careful in my previous post, and done a better job of defining this kind of implicit knowledge.
That's an interesting point. However, I think a silicon sentience built by humans is going to be very biased toward human psychology, given its "parentage".
I think in large it depends on how we create the silicon-based intelligence. If we program it directly and then just "tweak" it, then yes, I suspect you might be right. If on the other hand we evolve an intelligence on silicon, then I guess the result would be an intellectual architecture that is much more alien, and much more suited to the environment in a computer. But this is all very hand-wavy hypothesising:)
Futhermore, when I refer to the Turing test, I'm perhaps abusing the term a bit to refer to a vague, hand-wavy notion of "Does this program appear to be sentient to the end-user?" rather than the gimmicky implementation of the Loebner test. I know it's poor science, but when it comes to a sentient program, my criteria would be "I'll know it when I see it."
Indeed, defining sentience is not easy! I think a part of the problem arises intelligence and sentience are confused - I don't think they are the same thing. Solving calculus problems requires, in some shape or form, intelligence. However it doesn't require sentience. Perhaps you could define sentience along the lines of an ability to introspect (cf one definition of consciousness: "mental states that monitor mental states") and self-modify. From this it would seem that sentience requires intelligence, but not vice versa? Or have I simply argued circularly around the topic? I think my brain just blew a tube.
All of this is trivial for anyone with even rudimentry programming experience to implement, it's not especially profound, and it'll never pass the Turing test, but it is a legitimate part of the AI field.
Err-hmm... ummm no. It's not trivial. Writing the code is trivial. Implementing an expert system is a pain in the arse. Why? Because you have to get the expert knowledge out of the experts. This can be hard when the experts don't know what they know - to them it feels like they are going on hunches and "feelings" and not using rules, which they most often are. So you have a long, drawn out interview process which goes something like this:
"Well, if X, Y, Z, A, B, C, D, E, and F are present, well it's gotta be Q."
"I see... do you think you could break that down into steps for me. For example, if you consider just X, Y, and Z, does that mean anything?"
And you get to do this for hours on a topic that your often struggling to understand yourself. It might be easy for the genuine polymaths out there... but I think you'll find they are pretty damn rare!
Finally, to call the Turing Test a test of AI is, well, it's stupid. It may be the best test we have (I don't know), but frankly, the ability of an AI to imitate a human is not a good indication of sentience. After all, you would have to assume that the vast majority of alien lifeforms would fail the Turing test, given their psychology would be dramatically different to humans' - yet we're going to assume that a silicon sentience would be identical to a humans'?
Hmm, that's true. But presumably if the thread is still running, then its doing something useful.
I feel that this also contrasts with C and C++, because here, you might iterate through an algorithm several times, losing memory each time, whereas with a thread, you could only lose the memory if you continuously created new threads without getting rid of your old ones - which seems poor programming practice to me. Then again, I am not, by any means, an expert in threaded programming, so I may be completely misunderstanding you!
I honestly couldn't say... but I know I got very, very frustrated with an incredibly poor development environment (if you made a syntax error in your code, at runtime the error message was "syntax error" - no line number info, and we didn't have keyword highlighting.)
Then of course there was the problem that you couldn't do anything complex, because you had no pointers and no structures.
I suppose you could say that I was more frustrated with the limitations of BASIC than any specific problems with BBC BASIC.
From a personal point of view, I find your comment very interesting, because I think I'm going through the same process:
I was taught Acorn BASIC (?!) during school. Hated it, so:
I learned C++, I thought this was pretty damn cool - I could do a lot of stuff you just couldn't do in BASIC, the STL was powerful, and objects were damn good.
Then I learnt Java - and no more memory leaks or GPFs (at least, none that were my fault!). But I miss my generic programming and the ability to use pointers more flexibly when I too... so now I'm looking at Lisp and going "hmmmmm".
Actually, engines do fall off planes. It happened a few years back, with UA I believe.
Basically, Boeing said: "Don't move the engines around with forklifts". UA said "OK" and promptly started shifting engines around on forklifts. This damaged the engine mounts and led to one of them falling off in flight. Naturally Boeing had to take the blame as they couldn't afford to endanger their future sales by being known as the company that actually forces the airlines to do proper maintenance...
If our national telecommunications infrastructure was built with a bunch of small cheap satellites, I doubt our phones and television would be as reliable.
This is one of those lovely generalisations that could well be - and probably is false - depending on how you like to add numbers. Here's my view:
We want 99.9% reliability in any one node of our communications system. We can have one satelite, at a cost of roughly 2 million or so (minimum), which is, on its' own, 99.9% reliable.
Or we can launch a fleet of smaller ones, with 1/10th the capacity, each costing $200,000 and with 90% reliability. For 5 million, we can launch 25 of them. More than 40 must fail to even cut in to our bandwidth. The chance of this happening is 0.1^15, or a-very-small-number.
Furthermore - and here is the absolute key advantage of using multiple smaller systems - with the big baby, if she goes down, that's it. Your communications bandwidth through this "node" is zero. However, lets say you lose 16 of your mini-satelites. Your losing "necessary" bandwidth - but you still have some. You still have some, in fact, right up until the last satelite fails. Giving you time to add more satelites to the node. Giving you even greater reliability.
It's been analogised this way: 10 country roads and your average super highway can each carry the same amount of traffic. One mega-spill can close the super-highway, but you need a minimum of at least 10 crashes to completely close the smaller roads - and as you add extra country roads, the number of minor and major mishaps you can suffer increases as well.
You can fiddle the numbers any which way you like, but you'd have to weight them pretty heavily against the little satelites for large satelites to be more economic and reliable. (Perhaps logarithmic cost-to-bandwidth increases would do it?)
are all the replies that quote some portion - or multiple portions - of either Gates' speech and then "reply" with something along the lines of: "Oh gosh Bill, you are so darned silly. You didn't invent OSS/the internet/the computer, XYZ did!
Are there that many people who don't realise the existence of the vast gulf between what is said and what is thought in the world of PR?
You want an example: "The Greening Earth Group" is funded by the various coal mining corporations.
Microsoft is no exception - whatever they are saying in public, you can bet a few hundred shares of MSFT that in their corporate strategy sessions they aren't standing around patting themselves on the back.
Slashdot Mirror
Actually, it'll work fine and is perfectly legal. There are at least two ways to do it:
1) Transfer the borrowed money into a company. Apply for a loan in company name - the bank will still ask for a personal guarantee, but the entity applying for the loan has a positive net worth.
2) Borrow the money and wait 6 months. After 6 months, the banks don't care where the deposit came from - remember, you're buying property, which serves as it's own collateral. The deposit is merely the banks safety margin, and they have first call on the property in the event you go under anyway.
How do I know this? Because I'm doing strategy number 1 - and I just got preapproval on two mortgages yesterday.
Not that I'd recommend doing it with Credit Cards! There are any number of non-bank institutions that will lend you the money at only a % or so more than the "standard" rate.
If you are actually trying to claim it's a hoax - I'm not sure - then think about the *difficulty* of pulling this off... what follows is quoted from another /. thread long ago. It covers most of the difficulties rather nicely I think.
#
For the interest of Slashdot readers, national governments, and any other interested organization, I am posting instructions on how to fake a moon landing and not get caught for 30 years.
Before the Landing
Put out a request for tenders for a contract to build the lunar hardware to major aerospace companies. It would be pretty obvious after the fact that no one had built your launchers and landers.
In the contracts, give a specification that would lead the 10,000 engineers who work on the project to reasonably believe that the equipment could be used to land on the moon. Engineers are smart people; they could easily spot holes in your assumptions if you make the requirements less stringent than they have to be. If it were obvious that the hardware couldn't land on the moon, you would be caught.
Have the hardward manufactured and delivered. Again easy to spot if this wasn't done, especially for a Saturn V-class rocket and related assemblies.
In summary: You would actually have to build stuff that would probably be able to land a man on the moon, with all the associated expenses.
During the missions:
You will actually have to launch the thing you contracted to build. You could launch something else -- but why bother? We've already established that you have to build a moon rocket, and you'd have to pay off everyone who was involved in its destruction and substitution. Besides, it would be big news, so news organizations would want to film the launch of the big rocket.
So, the capsule could be suborbital, or stay in orbit, and the rest of the mission could be faked, right? Wrong. Antennas around the world will be tracking the radio signals from the capsule, including the continuous telemetry feeds. Something would have to go to the moon, on a realistic lunar trajectory, or this would be immediately spotted by legions of radio astronomers and HAM radio amateurs around the world, many of whom have advanced signal processing available to them (like Doppler analysis, etc.). They would also be able to tell the difference between a lunar trajectory and a different orbit, like a geosynchronous orbit, because of the moon's particular position in the sky.
So, the capsule has to go to the moon. Does it have to orbit? Yes. The capsule must stay in the vicinity of the moon for several days (again checked by those with large radio antennas). The only foolproof way to do that is to orbit.
So, the capsule has to orbit. Does it have to land? Yes. While in orbit, the capsule can't communicate with Earth from the far side of the moon. Yet a lander must be able to send continuous telemetry to the Earth. It would be pretty obvious fakery to have the "lander"'s telemetry fade out at the same time as the capsule's.
Does it have to come back? Yes; for the same trajectory reason. The return trajectory could be tracked.
Does the capsule/lander have to be manned? Not necessarily, but there would be many complications if it weren't. You would have to be able to carry on ground/capsule communications in a realistic manner even though the communications from the capsule would have to be recorded and beamed back (because your radio is being monitored). The "astronauts" would be unable to perform any diagnostic tasks aboard the spacecraft (because they're not aboard it), so the entire flight control team would have to be in on the hoax (dozens, even hundreds, of people to pay off).
In summary: You would have to actually send something to the moon, which may as well be manned.
After the Landing
Bring back tons of "moon rocks" and other materials for analysis by independent scientists around the world. These rocks could not be obviously of terrestrial origin, implying some exotic materials science (or creative geology). Either that or pay off anyone who comes in contact with the "lunar samples".
And if you're NASA - do this seven times, with one of the seven attempts turning into a remarkably realistic failure.
The upshot: It's equally easy and expensive to actually land a man on the moon than fake it convincingly. Furthermore, the evidence for fakery would not be found in trivial forms of evidence, like photographs, but in more obvious places, like contracts, accounting, radio monitoring, and the lunar samples themselves.
What like structured programming? Or - God forbid! - Object Oriented Programming! That crap's all just too damn slow to run and to damn difficult to code!
Sure, LP's not perfect now... but if you bother to learn only what's currently in fashion, by the time you've learnt it, the Next Bit Thing has arrived...
That said... you're perfectly correct when you say that no tool - be it OOP, LP, or any other silly acronym - can prevent bad programmers writing bad code - it can only help good programmers write better code.
But you'll notice that a large number of "normal" religions would not... for example:
The organization may not intervene in political campaigns, and
Large segments of the Christian world fail this test... c.f. abortion and the whole Religious Right bull.
No substantial part of its activity may be attempting to influence legislation,
See above.
Net earnings may not inure to the benefit of any private individual or shareholder,
And I guess the Catholic church fails this one too... you can't tell me that the officials in the Catholic church haven't become inured to their benefits... all those US20 million mansions the Cardinals live in... the golf courses and convention centres...
Ahh hell, who am I kidding. Your average American is, and always will be, a religious hypocrite. You'd have more luck reasoning with a rock. (NB: This probably doesn't include the vast majority of /. readers, or geeks in general).
You've got 7 hours during which you can't be interrupted. Seven hours free from your boss. Seven whole hours of entirely interruption free bliss. And you want to use it to play games? Dammit man! Code, or work, or do something productive... those 7 interruption-free hours are worth a week of "normal" work!
And your asking /. for advice?! Come on! If you want an opinion that's not based on someone else's opinion, for the love of God don't do an AskSlashdot!
please, dear god, offer RING BOUND versions of your books! Hell Yes! And if you can't ring bind them, at least bind them properly - I've had two books (albeit probably not yours) fall apart on me already just *this* year - and we're only one month in!
Yep, and it runs essentially something like this: If it walks like a duck, and quacks like a duck, then it is a duck. That is, to say, while neither the books, nor the person are individually a sentient mind that "knows" Chinese, together they comprise a (sentient) system that does.
As an analogy, lets say I can seperate the neurons that receive visual information, and those that process it into mental pictures. Arguably, on their own, neither set of nerves can "see" in a useful way. Searle's argument would have you believe that when you combine these two sets, what you get is not actually vision, but merely a convincing simulation of it. But if it looks like vision, and works like vision, then I'm afraid I'm gonna call it vision.
For them the mind is completely separate from the brain, an idea that has been pretty much discarded by the thinking public
This is a strawman argument. Dualism argues that the mind is seperate, partially or completely, from the physical world. I'd be interested who "them" are, since this is not a view held by many (any??) strong AI believers - after all, strong AI and dualism do contradict...
I think you'll find that what your actually talking about is a form of materialism, namely the variety that is characterised by the quote "the mind is software that runs on the brain". From this point of view, the mind and the brain are different, but very tightly related, in the same way that a compiled program and a CPU are.
If you want to say that this is in fact dualism, then be my guest. But redefining terms in this arbitrary manner is rather pointless if you actually want to try and debate an issue.
Umm... if I can't charge it to credit or a debit card (via EFTPOS or the like) I get annoyed. Hell, I haven't been forced to pay cash in New Zealand for at least 2 years... Occasionally I still do get cash out and end up with small change - but only as a result of the goddamned buses here. (And yes... only the buses. Even taxis have mobile EFTPOS terminals).
This is very much a cultural thing, and frankly, given how much more techie Taiwan, Japan, Korea etc are than even NZ, I wouldn't be surprised at all if you could be as physically cashless as you liked.
Of course the US might be different :)
Actually, this is only a partial solution. Because of little movements in the read/write heads, you actually have to a one, then a zero, then a one... and so on, depending on how securely you want to wipe out the data. It's the flipping of the polarisation of the little bits of oxide back and forth that actually wipes it out... anything else, will as you say, leave a residual magnetic signature which is recoverable with an oscilloscope and very fine motor control (still not easy though!).
For more information, see: Secure Deletion
- The earth has a radius of about 6300 km.
- This gives a volume of about 10^12 km^3
- This asteroid was at the closest about 830000 km from earth
- A sphere with a radius of 830000 has a volume of about 2.5*10^18
If we divide these numbers, we find that an object will be this close to earth on the average something a bit more than 2 million times as often as it actually hits the earth.Yup, you are simplifying... as I will be shortly. However you are actually being too tricky... the question is not whether the asteroid will intersect the sphere which makes up the earth, but whether it's trajectory will intersect the disk that makes up the earth (you've got one dimension to many I'm afraid!). Hence the chances are more likely to be in the region calculated as follows:
- Radius of Earth=6300 km
- Therefore the "target area" the earth presents is approximately pi*6300^2 = 125,000,000 sq. km
- Closest point of asteroid approach = 830,000 km
- Therefore the size of the disk that the asteroid actually hit is approximately = 2,164,000,000,000 sq. km
This gives odds nearer to 17000 to 1 against. Gravitational attraction and the fact that most asteroids lie in or near a plane which intersects the Earth will slightly increase the odds of a collision, but probably not significantly.And I do agree... these are still long odds. But odds worth doing something about :)
Oh but of course. And that's why Excel wasn't successful until it was both forwards and backwards compatible with Lotus 123?
Read this.
Read Joel's other stuff too.
And as much as I hate to say it, this also provides ammunition to the people who claim that open source is very good at copying other projects' work, but terrible at innovating. Honestly, of all the high profile open source projects, how many of them are a significant innovation, and how many are merely an attempt to produce an equivalent of feature Z of Windows or Unix or Mac OS on Linux?
Yeah? And what proportion of "high profile" closed-source projects really innovate?
Hmmm... very sexy. I will have to look at that... I can see it being incredibly useful for a project I am planning at the moment.
While you may be correct, I can't help but wonder how you are meant to implement any form of generalisation or inheritance in an RDBMS, without a huge mess of tables and complex relationships.
I mention this, because it is something I find myself wanting to do all the time, for example, when storing data that originates in OO programs. Being able to store it in an RDBMS has heaps of advantages for me (primarily that it is easier and less buggy to load and save data) - but I can't easily store the different info of different derived classes.
If you have any suggestions on how to solve this, I'd definitely appreciate it!
I agree with all of what you say. Talking someone through a decision tree is not at all complex, and sometimes expert systems are not the right solution.
The problem comes in constructing the decision tree itself - this is the type of expert knowledge which someone can know, but not enunciate. The situation is also made more complex when you have decision trees with feedback loops, more than one parallel tree, etc. I probably should have been more careful in my previous post, and done a better job of defining this kind of implicit knowledge.
I think in large it depends on how we create the silicon-based intelligence. If we program it directly and then just "tweak" it, then yes, I suspect you might be right. If on the other hand we evolve an intelligence on silicon, then I guess the result would be an intellectual architecture that is much more alien, and much more suited to the environment in a computer. But this is all very hand-wavy hypothesising :)
Futhermore, when I refer to the Turing test, I'm perhaps abusing the term a bit to refer to a vague, hand-wavy notion of "Does this program appear to be sentient to the end-user?" rather than the gimmicky implementation of the Loebner test. I know it's poor science, but when it comes to a sentient program, my criteria would be "I'll know it when I see it."
Indeed, defining sentience is not easy! I think a part of the problem arises intelligence and sentience are confused - I don't think they are the same thing. Solving calculus problems requires, in some shape or form, intelligence. However it doesn't require sentience. Perhaps you could define sentience along the lines of an ability to introspect (cf one definition of consciousness: "mental states that monitor mental states") and self-modify. From this it would seem that sentience requires intelligence, but not vice versa? Or have I simply argued circularly around the topic? I think my brain just blew a tube.
All of this is trivial for anyone with even rudimentry programming experience to implement, it's not especially profound, and it'll never pass the Turing test, but it is a legitimate part of the AI field.
Err-hmm... ummm no. It's not trivial. Writing the code is trivial. Implementing an expert system is a pain in the arse. Why? Because you have to get the expert knowledge out of the experts. This can be hard when the experts don't know what they know - to them it feels like they are going on hunches and "feelings" and not using rules, which they most often are. So you have a long, drawn out interview process which goes something like this:
"Well, if X, Y, Z, A, B, C, D, E, and F are present, well it's gotta be Q."
"I see... do you think you could break that down into steps for me. For example, if you consider just X, Y, and Z, does that mean anything?"
And you get to do this for hours on a topic that your often struggling to understand yourself. It might be easy for the genuine polymaths out there... but I think you'll find they are pretty damn rare!
Finally, to call the Turing Test a test of AI is, well, it's stupid. It may be the best test we have (I don't know), but frankly, the ability of an AI to imitate a human is not a good indication of sentience. After all, you would have to assume that the vast majority of alien lifeforms would fail the Turing test, given their psychology would be dramatically different to humans' - yet we're going to assume that a silicon sentience would be identical to a humans'?
Hmm, that's true. But presumably if the thread is still running, then its doing something useful.
I feel that this also contrasts with C and C++, because here, you might iterate through an algorithm several times, losing memory each time, whereas with a thread, you could only lose the memory if you continuously created new threads without getting rid of your old ones - which seems poor programming practice to me. Then again, I am not, by any means, an expert in threaded programming, so I may be completely misunderstanding you!
I honestly couldn't say... but I know I got very, very frustrated with an incredibly poor development environment (if you made a syntax error in your code, at runtime the error message was "syntax error" - no line number info, and we didn't have keyword highlighting.)
Then of course there was the problem that you couldn't do anything complex, because you had no pointers and no structures.
I suppose you could say that I was more frustrated with the limitations of BASIC than any specific problems with BBC BASIC.
From a personal point of view, I find your comment very interesting, because I think I'm going through the same process:
I was taught Acorn BASIC (?!) during school. Hated it, so:
I learned C++, I thought this was pretty damn cool - I could do a lot of stuff you just couldn't do in BASIC, the STL was powerful, and objects were damn good.
Then I learnt Java - and no more memory leaks or GPFs (at least, none that were my fault!). But I miss my generic programming and the ability to use pointers more flexibly when I too... so now I'm looking at Lisp and going "hmmmmm".
I ain't gonna try arguing with you there! :)
Since when has 1770 feet been a kilometre?! It works out to being only just over 530m (3 feet to a yard, 0.9 metres to a yard).
At least try for some basic arithmetical accuracy before launching those flames.
Basically, Boeing said: "Don't move the engines around with forklifts". UA said "OK" and promptly started shifting engines around on forklifts. This damaged the engine mounts and led to one of them falling off in flight. Naturally Boeing had to take the blame as they couldn't afford to endanger their future sales by being known as the company that actually forces the airlines to do proper maintenance...
This is one of those lovely generalisations that could well be - and probably is false - depending on how you like to add numbers. Here's my view:
We want 99.9% reliability in any one node of our communications system. We can have one satelite, at a cost of roughly 2 million or so (minimum), which is, on its' own, 99.9% reliable.
Or we can launch a fleet of smaller ones, with 1/10th the capacity, each costing $200,000 and with 90% reliability. For 5 million, we can launch 25 of them. More than 40 must fail to even cut in to our bandwidth. The chance of this happening is 0.1^15, or a-very-small-number.
Furthermore - and here is the absolute key advantage of using multiple smaller systems - with the big baby, if she goes down, that's it. Your communications bandwidth through this "node" is zero. However, lets say you lose 16 of your mini-satelites. Your losing "necessary" bandwidth - but you still have some. You still have some, in fact, right up until the last satelite fails. Giving you time to add more satelites to the node. Giving you even greater reliability.
It's been analogised this way: 10 country roads and your average super highway can each carry the same amount of traffic. One mega-spill can close the super-highway, but you need a minimum of at least 10 crashes to completely close the smaller roads - and as you add extra country roads, the number of minor and major mishaps you can suffer increases as well.
You can fiddle the numbers any which way you like, but you'd have to weight them pretty heavily against the little satelites for large satelites to be more economic and reliable. (Perhaps logarithmic cost-to-bandwidth increases would do it?)
"Oh gosh Bill, you are so darned silly. You didn't invent OSS/the internet/the computer, XYZ did!
Are there that many people who don't realise the existence of the vast gulf between what is said and what is thought in the world of PR?
You want an example: "The Greening Earth Group" is funded by the various coal mining corporations.
Microsoft is no exception - whatever they are saying in public, you can bet a few hundred shares of MSFT that in their corporate strategy sessions they aren't standing around patting themselves on the back.