>Bzzzt! Wrong... the Turing test says nothing about whether something is intelligent, merely whether it can fool a
>person. There are already some pretty good pieces of software out there about this, and they'll get better in the next
>few years. But they won't be intelligent.
Yes, but then there's the real Searle test:
'The ability to tell that a bullshit theory about the impossibility of doing something is full of it.' Hmm. Perhaps the real Searle test is a better test of whether someone or something is intelligent than the Turing test.
Actually, many of the things a human 'learns' in 17 years it already knew.
Mammals are all are born with a pretty good understanding of basic anatomy (e.g. horses and cats and dogs aren't bad at reading human body language- they know about faces, arms, legs- nobody taught them this stuff either), they all have built in emotions, and some appreciation of interactions between cause and effect, I poke/attack you, you usually go away that kind of thing; often this knowledge is semiunconscious, but its built in.
Computers are the ultimate tabulae rasa. They need this stuff that a few billion years of evolution knocked into our wiring.
As another example it seems pretty likely from the research on language that humans are born with the ability to learn ANY human grammar; we are genetically programmed for languages (if you challenge that, the only real difference between humans and sheep is genetic- lambs cannot learn language).
That's what CYC is. To wildly oversimplify, its a billion years of evolution plus a kindergarten education.
>A law is an observation of an occurence, such as the law of gravity. The law states that gravity occurs, but does not seek to explain why. There are innumerable THEORIES of gravity that seek to explain why.
Negatory. Newton's law of gravitation in fact is a theory. It is a theory that nature agrees with the law. In fact it has since been discovered that nature doesn't follow Newton's law, and does follow Einstein's general theory. Newton's law of gravity in fact has been disproven.
Unsurprisingly, modern physics now uses 'law' and 'theory' interchangeably.
>A scientific theory is NOT law, and it is NOT fact.
I have a theory. My theory is that I won't see anything in the cup in my hand when I look next. Gee, my cup really was empty. My theory is proven. So it is possible to prove some theories which can be explored via exhaustive search; or atleast come up with extremely persuasive evidence that amounts to turning the theory into a fact.
I know where you are coming from. But taking your argument to a reasonable conclusion, you are saying that there is no such thing as a fact ever.
So what is a fact in a scientific sense? It's the simplest theory that has good evidence to support it, and no known disproof.
By that standard, Newton's law isn't a fact, but Einstein's General Theory of relativity is.
>In science, a theory is an idea that has withstood the test of time and has not yet been disproven. You can NOT prove a theory. Never. EVER. You can only disprove it by experimentation.
An interesting scientific theory. Can you prove it?
>Theories are hypothesis with insufficient evidence to be laws... What do I mean?
>Theories have a few pieces of 'evidence' that support it, but not SUFFICIENT 'evidence' for it to be a law."
I'm glad your such an expert on the technical use of scientific language. Tell me, how many scientific papers have you actually read lately? Do you have any scientific qualifications? Or do you write dictionaries? Are you a lexicographer by profession? Who the hell are you to tell the scientific community what a word they use in a technical sense means?
'Cos I got news for you buddy- that ain't what a theory is. Thank you for playing.
No. You are saying these pieces don't fit, THEREFORE these theories don't fit. But you're missing the other 100,000 pieces that DO fit.
Unfortunately for you. Darwin is true, verified science. We can read the genome well enough to see the entire family tree of life. Everything is interrelated. You and every other person shares 50% of their genes with bananas, and that's because we have a common ancestor. There is no reasonable doubt that Darwinian evolution has happened. In your language evolution is a fact. Hard, hard evidence.
Another example. Wheat. Wheat is a mutation. We know this, because it's very recent and the mutation has been studied. A random mutation. Actually wheat is almost sterile. You have to winnow it to allow it to propogate. So it stuck for one reason and one reason only. A person walking past spotted it, and manually propogated it. Random mutation plus selection. Unnatural selection in that case...
Oh BTW. Theories. You've been had. Totally suckered. Formally, 'theory' is the scientific name for any law, concept or set of laws or concepts. It has absolutely nothing whatsoever to do with doubt or absense of trust. I can equally claim that you have a theory of God; you would call it 'belief'. And we'd mean exactly the same thing, (actually there's a lot more evidence for scientific theories!)
e.g. Einsteins Special Theory of Relativity is substantiated thousands of times per day in cyclotrons all around the world. Even GPS (a navigation system used by aircraft, cars ships and boats) is off by literally miles without corrections that can only be derived using the theory. There is no doubt but that this theory is true; its a law and a fact. As certain as water flowing down hill.
Actually I once read a sci-fi story. The basic premise was that the aliens went around taking samples of peoples DNA. Trouble was they were carnivores, but civilised. So, they grew a clone of you, and then... dinner time... everyone was happy. I think.
So, Hey Alien! Want a sample of MY DNA. I don't think so... You want some, you suck on this! Keep your filthy tentacles off me and my cloned buddies...
Science is the worlds biggest jigsaw puzzle. We've connected up hundreds of thousands of pieces, some of the sky and quite a lot of the bits of the earth, and it's like you're sitting there going: you've made these 6 mistakes! It doesn't fit!
Jigsaw puzzles don't fall on a few mistakes (if these are mistakes), jigsaw puzzles have their own internal consistency, and so does nature.
Take one "Origin of the species" and come back in the morning. Hopefully your religion will have cleared up by then. On the web at a browser near you.
Whilst it can be seen as amusing. This guy actually believes this stuff. He's a last fridayist. He doesn't deserve to be moderated up.
Check out his other postings, including the one about his wife's menstruation being a sin...
What a loon.
In this case I would be intrigued to know what he thinks about the frozen woolly mammoths that have been found in the Russian tundra... actually no, on second thoughts, I don't care.
The more deeply the optimiser is run the BIGGER the percentage speedup Dynamo gives. The reason is that the speedups are different speedups than the ones found by the compiler, and the less time spent in the rest of the code, the more significant the speedups are.
e.g. the compiler can't optimise into a DLL, but Dynamo can
e.g. Dynamo can profile the code and optimise virtual function calls in ways that the compiler can't
Xerox's idea is similar (although as far as I know they don't have color yet); but they (not unnaturally;-) think that a laser-printer type scheme is better.
That way it separates the paper from the drive circuitry. And the drive circuitry is expensive in fact because it needs relatively high voltages, and you need a lot of hardware around the edges of the paper to get all those pixels to do the right things.
Putting the drive circuitry inside a laser printer box means that the paper moves past it and hence the paper is much cheaper. However color is probably harder that way. Still, how often do you NEED color?
Where I work, we have a lot of B&W printouts of things that aren't needed to be kept for long; and the reusable paper would be ideal in that case.
I mean if it takes over a hundred years to discover that most stomach ulcers are caused by a bacteria; perhaps all these diseases are symptoms of a viruses that spread and only very occasionally cause these symptoms in some of the affected people.
After all, most diseases that we hear of are the nasty ones. But nasty diseases don't tend to last as well as the (mostly) innocuous ones.
Perhaps CTS is caused by a virus that has passed its epidemic! Just because they can't find a cause, doesn't mean that there isn't one. If there was a virus that weakens the nerves in hands or something, how would we know? Who would look for it?
Maybe there was a virus that caused low blood sugar in the 70s too.
I mean the easiest way to get a clue is to understand the issues by seeing them first hand. It would not surprise me if this is the start of a flood of laws to ban inappropriate email in various different jurisdictions. It doesn't take many court judgements to destroy the profitability of this and substantially moderate the frequency of junk emails. But there are always going to be some I suspect.
This is NOT a major money spinner on the internet; but there is money to be made right now. Regretably.
"I hope they are not going to continue to require rocket boosting to get fast enough for the scramjet engine to be operable. But then, how could they get up to that speed?"
So what if they do use rockets? Rockets are very efficient (more efficient than jet engines, and can be much less polluting); and can be made reliable.
In an interview with the Chicago Sun-Times, Linux inventor Linus Torvalds says that Microsoft and closed source software companies are "good competition" because it will "force Linux to be innovative," but calls Microsoft "a cancer that attaches itself in an intellectual property sense to everything it touches." He also says that the inclusion of IE in windows has been "not great... for innovation in the software industry" (especially for Netscape) and that MS's new copy protections are just "bumps in the road" to "help customers understand when they are crossing the line . . . so they can't do the wrong thing." And he says a few more amusing things, also.
Yes, its difficult to get exact figures, subtly different questions give you very big differences in the figures (e.g. yearly launch pad costs are they included per launch or not?). Let's suffice it to say that depending on how you cut it, the space shuttle costs somewhere between $100 million and $1.5 billion; mostly towards the top end of the range.
Still, although I was trying to be fairly non controversial, in a fairly real sense, the comparison is unfair to the Russians as you can't buy a launch in the shuttle if you wanted to; but if you could it would seem reasonable for NASA to charge for profit ontop of the costs. Whereas the $85 million for a Proton V is all in.
"Maybe I'll see space tourism in my lifetime afterall."
I'm happy to tell you; you already have! You can die happy!;-)
Ok, on a very much less flippant note. There are reasonably good reasons to think that the underlying costs to get into space are just a few thousand per person (if you look at the fuel costs which are much, much less than most people would suppose.)
The Russians are only two orders of magnitude more than this, NASA is three.
The nice thing about space is that the costs are likely to go down as the rate of launch goes up- right now most things are handbuilt and so costs are huge.
Space launch is growing at about 15% a year. Wait 5 years and we will see an order of magnitude more stuff going up there. That's going to affect the price.
I think within 15 years (if the trend continues) we will see large scale tourism.
Actually Tito probably more or less paid for the whole flight. The entire Russian space budget runs to under $200 million; the cost of each rocket is about $4 million (source: Gary Hudson of Rotary Rocket fame). There will be launch pad costs on top of that, but I doubt they will add up to more than $8 million.
By comparison, the unit cost of a space shuttle launch is $100 million including pad maintenance; but not including development costs.
The rocket that Tito went up on is much less capable of course. The more capable rocket the Russians have Proton V, which can carry a similar payload mass to the shuttle costs about $5 million. The Russians charge $85 million per launch cf $500+ million for the space shuttle.
Yeah, probably can work. The downsides are that as people move around you can lose connectivity for moments or minutes (although cell phones have this problem anyway to some extent, but this would be worse). To counteract that it would probably be necessary to keep several connections up simultaneously; in the hopes that there is atleast one route to the far end. That will mean that the phone will take a lot more current and will either be heavier or will flatten its batteries much more quickly.
The other difficulty is routing in a highly dynamic link map- everyone is moving around all the time; links will be going up and down like mad things...
Actually, theoretically, the carbon doesn't get used up in the blast furnace; so you could import it to get started (Theoretically; in practice I suspect that carbides will form.)
But it does turn into carbon dioxide. It should be possible to use plants to turn the carbon dioxide back into carbon again; although that becomes a rate limiting step in the blast furnace.
So you only need enough carbon to top up the losses, with luck the plants might liberate enough for that as well, or it might need special processing like fractional distillation of the rock to extract that carbonates.
I haven't investigated the Calcium bit, perhaps that could be recycled somehow as well?
Carbon is needed for making steel and there ain't very much of it on the moon (ppm type abundancies).
Tritium isn't much use currently. Its supposed to be useful for nuclear fusion. But nobody knows how to make controlled nuclear fusion work, and personally I'm quite skeptical it will ever work economically. Still, you never know. Tokomaks might fly.
I like the Space Shuttle. The space shuttle is the most capable launch vehicle ever invented. It is a triumph of engineering. It is also the most expensive launch vehicle ever invented per pound of payload.
Do you have a citation for your statement that most of the space shuttle costs are fuel? I'm sorry but I suggest you check that out. The highest expense is the wages of the people that were needed to design and build the orbiters by a very, very long way.
p.s. a space shuttle launch actually is more like 100 million per launch. The figure of a 1 to 1.5 billion allows for the very high costs incurred during its development; they've been divided over the expected number of launches; and they don't expect to launch often enough to make it cheaper.
To pick a number at the other end of the spectrum, the Russian Proton V rocket costs them more like $5m per launch, but they charge $80m to help pay for their space program. Its not nearly as good as the Space Shuttle, but it has similarish payload.
If you do the maths as I have you find that the space shuttle is almost an order of magnitude more expensive per pound.
No offense, but I hope you failed your assignment.
Check out space and tech launch data. Soyuz has a 30:1 launch mass:payload ratio.
most of which is going to be fuel; and that's an old multistager and not at all cutting edge.
Besides, fuel is cheap (government tax notwithstanding.) Fuel costs are much less than 1% of rocketry costs right now. The rest goes into the armies of people needed to launch these things. Thing is, the armies don't get much bigger when you launch much more often - they are fixed costs and the unit costs are very much lower.
a) You're assuming you want to send the materials back to the earth. For a Mars mission you wouldn't want to do this.
b) the crust certainly isn't lacking metals- there's plenty of iron, aluminium, titanium etc; see Permanents list of minerals on the moon. And they are already oxides. For several reasons they are probably easier to separate in space (some techniques like fractional distillation of the oxides[!] are easier there due to ready availability of solar energy and the micro gravity.)
c) the moon is short of volatiles (especially hydrogen and carbon) however; there does seem to be a supply of hydrogen in some form at the poles, and there is speculation that it might contain carbon too.
d) Imagine it was back in the 15 century: "why would we want to go to the west indies? There's nothing there we haven't got plenty of...". In point of fact there's more stuff in space than down here on earth (particularly solar energy.)
Slashdot Mirror
>Blind adherence to rules is not intelligence.
.' Hmm. Perhaps the real Searle test is a better test of whether someone or something is intelligent than the Turing test.
That's a rule you blindly adhere to is it?
>Bzzzt! Wrong... the Turing test says nothing about whether something is intelligent, merely whether it can fool a
>person. There are already some pretty good pieces of software out there about this, and they'll get better in the next
>few years. But they won't be intelligent.
Yes, but then there's the real Searle test:
'The ability to tell that a bullshit theory about the impossibility of doing something is full of it
Actually, many of the things a human 'learns' in 17 years it already knew.
Mammals are all are born with a pretty good understanding of basic anatomy (e.g. horses and cats and dogs aren't bad at reading human body language- they know about faces, arms, legs- nobody taught them this stuff either), they all have built in emotions, and some appreciation of interactions between cause and effect, I poke/attack you, you usually go away that kind of thing; often this knowledge is semiunconscious, but its built in.
Computers are the ultimate tabulae rasa. They need this stuff that a few billion years of evolution knocked into our wiring.
As another example it seems pretty likely from the research on language that humans are born with the ability to learn ANY human grammar; we are genetically programmed for languages (if you challenge that, the only real difference between humans and sheep is genetic- lambs cannot learn language).
That's what CYC is. To wildly oversimplify, its a billion years of evolution plus a kindergarten education.
>A law is an observation of an occurence, such as the law of gravity. The law states that gravity occurs, but does not seek to explain why. There are innumerable THEORIES of gravity that seek to explain why.
Negatory. Newton's law of gravitation in fact is a theory. It is a theory that nature agrees with the law. In fact it has since been discovered that nature doesn't follow Newton's law, and does follow Einstein's general theory. Newton's law of gravity in fact has been disproven.
Unsurprisingly, modern physics now uses 'law' and 'theory' interchangeably.
>A scientific theory is NOT law, and it is NOT fact.
I have a theory. My theory is that I won't see anything in the cup in my hand when I look next. Gee, my cup really was empty. My theory is proven. So it is possible to prove some theories which can be explored via exhaustive search; or atleast come up with extremely persuasive evidence that amounts to turning the theory into a fact.
I know where you are coming from. But taking your argument to a reasonable conclusion, you are saying that there is no such thing as a fact ever.
So what is a fact in a scientific sense? It's the simplest theory that has good evidence to support it, and no known disproof.
By that standard, Newton's law isn't a fact, but Einstein's General Theory of relativity is.
>In science, a theory is an idea that has withstood the test of time and has not yet been disproven. You can NOT prove a theory. Never. EVER. You can only disprove it by experimentation.
An interesting scientific theory. Can you prove it?
So you don't mind the idea of people sitting down to Bobo the Space Chimp steak then? Somebody sucking on one of your juicy eyeballs?
;-)
You sure?
How about you get to meet someone who had eaten your clone the night before and compliments you on your very good flavour?
These are rhetorical questions BTW
It raises some disturbing questions don't you think?
>Theories are hypothesis with insufficient evidence to be laws... What do I mean?
>Theories have a few pieces of 'evidence' that support it, but not SUFFICIENT 'evidence' for it to be a law."
I'm glad your such an expert on the technical use of scientific language. Tell me, how many scientific papers have you actually read lately? Do you have any scientific qualifications? Or do you write dictionaries? Are you a lexicographer by profession? Who the hell are you to tell the scientific community what a word they use in a technical sense means?
'Cos I got news for you buddy- that ain't what a theory is. Thank you for playing.
No. You are saying these pieces don't fit, THEREFORE these theories don't fit. But you're missing the other 100,000 pieces that DO fit.
Unfortunately for you. Darwin is true, verified science. We can read the genome well enough to see the entire family tree of life. Everything is interrelated. You and every other person shares 50% of their genes with bananas, and that's because we have a common ancestor. There is no reasonable doubt that Darwinian evolution has happened. In your language evolution is a fact. Hard, hard evidence.
Another example. Wheat. Wheat is a mutation. We know this, because it's very recent and the mutation has been studied. A random mutation. Actually wheat is almost sterile. You have to winnow it to allow it to propogate. So it stuck for one reason and one reason only. A person walking past spotted it, and manually propogated it. Random mutation plus selection. Unnatural selection in that case...
Oh BTW. Theories. You've been had. Totally suckered. Formally, 'theory' is the scientific name for any law, concept or set of laws or concepts. It has absolutely nothing whatsoever to do with doubt or absense of trust. I can equally claim that you have a theory of God; you would call it 'belief'. And we'd mean exactly the same thing, (actually there's a lot more evidence for scientific theories!)
e.g. Einsteins Special Theory of Relativity is substantiated thousands of times per day in cyclotrons all around the world. Even GPS (a navigation system used by aircraft, cars ships and boats) is off by literally miles without corrections that can only be derived using the theory. There is no doubt but that this theory is true; its a law and a fact. As certain as water flowing down hill.
Like evolution. Genes don't lie.
Actually I once read a sci-fi story. The basic premise was that the aliens went around taking samples of peoples DNA. Trouble was they were carnivores, but civilised. So, they grew a clone of you, and then... dinner time... everyone was happy. I think.
So, Hey Alien! Want a sample of MY DNA. I don't think so... You want some, you suck on this! Keep your filthy tentacles off me and my cloned buddies...
;-)
Science is the worlds biggest jigsaw puzzle. We've connected up hundreds of thousands of pieces, some of the sky and quite a lot of the bits of the earth, and it's like you're sitting there going: you've made these 6 mistakes! It doesn't fit!
Jigsaw puzzles don't fall on a few mistakes (if these are mistakes), jigsaw puzzles have their own internal consistency, and so does nature.
Take one "Origin of the species" and come back in the morning. Hopefully your religion will have cleared up by then. On the web at a browser near you.
Whilst it can be seen as amusing. This guy actually believes this stuff. He's a last fridayist. He doesn't deserve to be moderated up.
Check out his other postings, including the one about his wife's menstruation being a sin...
What a loon.
In this case I would be intrigued to know what he thinks about the frozen woolly mammoths that have been found in the Russian tundra... actually no, on second thoughts, I don't care.
The more deeply the optimiser is run the BIGGER the percentage speedup Dynamo gives. The reason is that the speedups are different speedups than the ones found by the compiler, and the less time spent in the rest of the code, the more significant the speedups are.
e.g. the compiler can't optimise into a DLL, but Dynamo can
e.g. Dynamo can profile the code and optimise virtual function calls in ways that the compiler can't
Xerox's idea is similar (although as far as I know they don't have color yet); but they (not unnaturally ;-) think that a laser-printer type scheme is better.
That way it separates the paper from the drive circuitry. And the drive circuitry is expensive in fact because it needs relatively high voltages, and you need a lot of hardware around the edges of the paper to get all those pixels to do the right things.
Putting the drive circuitry inside a laser printer box means that the paper moves past it and hence the paper is much cheaper. However color is probably harder that way. Still, how often do you NEED color?
Where I work, we have a lot of B&W printouts of things that aren't needed to be kept for long; and the reusable paper would be ideal in that case.
I mean if it takes over a hundred years to discover that most stomach ulcers are caused by a bacteria; perhaps all these diseases are symptoms of a viruses that spread and only very occasionally cause these symptoms in some of the affected people.
;-)
After all, most diseases that we hear of are the nasty ones. But nasty diseases don't tend to last as well as the (mostly) innocuous ones.
Perhaps CTS is caused by a virus that has passed its epidemic! Just because they can't find a cause, doesn't mean that there isn't one. If there was a virus that weakens the nerves in hands or something, how would we know? Who would look for it?
Maybe there was a virus that caused low blood sugar in the 70s too.
Or perhaps not...
... but say they are? ;-)
Or servers that say they aren't in Washington; but are? Make money fast that way!
(FYI there is a standard for identifying physical location for servers, but it isn't well implemented.)
... judges get spammed isn't it? ;-)
I mean the easiest way to get a clue is to understand the issues by seeing them first hand. It would not surprise me if this is the start of a flood of laws to ban inappropriate email in various different jurisdictions. It doesn't take many court judgements to destroy the profitability of this and substantially moderate the frequency of junk emails. But there are always going to be some I suspect.
This is NOT a major money spinner on the internet; but there is money to be made right now. Regretably.
"I hope they are not going to continue to require rocket boosting to get fast enough for the scramjet engine to be operable. But then, how could they get up to that speed?"
So what if they do use rockets? Rockets are very efficient (more efficient than jet engines, and can be much less polluting); and can be made reliable.
In an interview with the Chicago Sun-Times, Linux inventor Linus Torvalds says that Microsoft and closed source software companies are "good competition" because it will "force Linux to be innovative," but calls Microsoft "a cancer that attaches itself in an intellectual property sense to everything it touches." He also says that the inclusion of IE in windows has been "not great ... for innovation in the software industry" (especially for Netscape) and that MS's new copy protections are just "bumps in the road" to "help customers understand when they are crossing the line . . . so they can't do the wrong thing." And he says a few more amusing things, also.
Yes, its difficult to get exact figures, subtly different questions give you very big differences in the figures (e.g. yearly launch pad costs are they included per launch or not?). Let's suffice it to say that depending on how you cut it, the space shuttle costs somewhere between $100 million and $1.5 billion; mostly towards the top end of the range.
Still, although I was trying to be fairly non controversial, in a fairly real sense, the comparison is unfair to the Russians as you can't buy a launch in the shuttle if you wanted to; but if you could it would seem reasonable for NASA to charge for profit ontop of the costs. Whereas the $85 million for a Proton V is all in.
"Maybe I'll see space tourism in my lifetime afterall."
;-)
I'm happy to tell you; you already have! You can die happy!
Ok, on a very much less flippant note. There are reasonably good reasons to think that the underlying costs to get into space are just a few thousand per person (if you look at the fuel costs which are much, much less than most people would suppose.)
The Russians are only two orders of magnitude more than this, NASA is three.
The nice thing about space is that the costs are likely to go down as the rate of launch goes up- right now most things are handbuilt and so costs are huge.
Space launch is growing at about 15% a year. Wait 5 years and we will see an order of magnitude more stuff going up there. That's going to affect the price.
I think within 15 years (if the trend continues) we will see large scale tourism.
Actually Tito probably more or less paid for the whole flight. The entire Russian space budget runs to under $200 million; the cost of each rocket is about $4 million (source: Gary Hudson of Rotary Rocket fame). There will be launch pad costs on top of that, but I doubt they will add up to more than $8 million.
By comparison, the unit cost of a space shuttle launch is $100 million including pad maintenance; but not including development costs.
The rocket that Tito went up on is much less capable of course. The more capable rocket the Russians have Proton V, which can carry a similar payload mass to the shuttle costs about $5 million. The Russians charge $85 million per launch cf $500+ million for the space shuttle.
P2P cell phones?
Yeah, probably can work. The downsides are that as people move around you can lose connectivity for moments or minutes (although cell phones have this problem anyway to some extent, but this would be worse). To counteract that it would probably be necessary to keep several connections up simultaneously; in the hopes that there is atleast one route to the far end. That will mean that the phone will take a lot more current and will either be heavier or will flatten its batteries much more quickly.
The other difficulty is routing in a highly dynamic link map- everyone is moving around all the time; links will be going up and down like mad things...
Actually, theoretically, the carbon doesn't get used up in the blast furnace; so you could import it to get started (Theoretically; in practice I suspect that carbides will form.)
But it does turn into carbon dioxide. It should be possible to use plants to turn the carbon dioxide back into carbon again; although that becomes a rate limiting step in the blast furnace.
So you only need enough carbon to top up the losses, with luck the plants might liberate enough for that as well, or it might need special processing like fractional distillation of the rock to extract that carbonates.
I haven't investigated the Calcium bit, perhaps that could be recycled somehow as well?
Carbon is needed for making steel and there ain't very much of it on the moon (ppm type abundancies).
Tritium isn't much use currently. Its supposed to be useful for nuclear fusion. But nobody knows how to make controlled nuclear fusion work, and personally I'm quite skeptical it will ever work economically. Still, you never know. Tokomaks might fly.
Do you have a citation for your statement that most of the space shuttle costs are fuel? I'm sorry but I suggest you check that out. The highest expense is the wages of the people that were needed to design and build the orbiters by a very, very long way.
p.s. a space shuttle launch actually is more like 100 million per launch. The figure of a 1 to 1.5 billion allows for the very high costs incurred during its development; they've been divided over the expected number of launches; and they don't expect to launch often enough to make it cheaper.
To pick a number at the other end of the spectrum, the Russian Proton V rocket costs them more like $5m per launch, but they charge $80m to help pay for their space program. Its not nearly as good as the Space Shuttle, but it has similarish payload.
If you do the maths as I have you find that the space shuttle is almost an order of magnitude more expensive per pound.
Check out space and tech launch data. Soyuz has a 30:1 launch mass:payload ratio. most of which is going to be fuel; and that's an old multistager and not at all cutting edge.
Besides, fuel is cheap (government tax notwithstanding.) Fuel costs are much less than 1% of rocketry costs right now. The rest goes into the armies of people needed to launch these things. Thing is, the armies don't get much bigger when you launch much more often - they are fixed costs and the unit costs are very much lower.
a) You're assuming you want to send the materials back to the earth. For a Mars mission you wouldn't want to do this.
b) the crust certainly isn't lacking metals- there's plenty of iron, aluminium, titanium etc; see Permanents list of minerals on the moon. And they are already oxides. For several reasons they are probably easier to separate in space (some techniques like fractional distillation of the oxides[!] are easier there due to ready availability of solar energy and the micro gravity.)
c) the moon is short of volatiles (especially hydrogen and carbon) however; there does seem to be a supply of hydrogen in some form at the poles, and there is speculation that it might contain carbon too.
d) Imagine it was back in the 15 century: "why would we want to go to the west indies? There's nothing there we haven't got plenty of...". In point of fact there's more stuff in space than down here on earth (particularly solar energy.)