The RDA can't do physics

[Clarke]

Originally Posted by Human No More

What relevance do they have to a difference in scale? An enzyme and associated proteins are huge compared to the scale you were until recently talking about.

Well, yes, they are huge. Yet they still work as nano-manipulators, which you insisted were impossible earlier. Considering how powerful biotechnology is, there's your answer: gene-modify something to build Unobtanium for you.

Exactly. Thank you for disproving your own point.

I wasn't the one who brought up the date.

No, a higher particle count per line would since the majority would not adopt the desired state. Stop being intentionally obtuse.

Then have fewer particles per line, and more lines. ...Though I wasn't aware that particles were grouped together at all.

It is still 3 bits/hour. That means the average latency is still 20 minutes even if you had 10,000,000 of them (which actually would be prohibitively expensive).

Ding! With 10,000,000 lines, we can send 10MB per 20 minutes... or be clever about it and send 22kb per second.

you're still multiplying the on-world presence needed, as well as, as I mentioned before, abandoning all pretence of following the contract.

How can we require more technicians than SpecOps?

No, the fact that there is no communication and therefore none of the cooperation that there was meant to be - you know, the entire reason avatars were started in the first place.

Then bring the Avatars. Cuts into your savings, but that's going to happen regardless.

...Since the avatar link is psionic and apparently uninterruptible, are we sure that it doesn't double as FTL? In that case, we save yet more fuel/money: we don't have to bring the avatar drivers.

Why else are they there then?
Also, great way to move the goalposts again
If humans ARE required. there is no conceivable reason not to use manual operations where possible on a world which impeded electronic operation and has an atmosphere that also causes increased maintenance requirements. Any kind of self-maintenance by robots would be limited in scope to known and programmed procedures, and without the marines, there would be no defence at all.

Automated defenses as well. That's actually a lot easier than self-maintaining robots.

Also, the "impeded electronics" thing is not real physics full stop. A magnetic field large enough to even noticeably impede low-velocity electronics will do really, really bizarre things to biological tissue. ...And shielding is still lighter than humans.

They did.
Now you've resorted to restating what actually happened in defence of some 'robots lololol' point?

No, they took the equipment and far, far too many slow, squishy humans. And I'm restating what actually happened as a solution to things you point out. Portions of the RDA's solutions are efficient, but the humans thing isn't.

...and reproducing the conditions during a the formation of a star and associated solar system on a macroscopic scale?

Temperatures up to 1GK can be generated in fusion reactors. (But that's not going to help to generate compounds. ) Magnetic fields up to 15T can be generated non-explosivly. Acceleration can't be generated in a chem. lab, beyond centrifuges, AFAIK, but that's not going to be true for much longer if unobtanium is discovered to require it. Anything else?

Then why were you going on about 'use nanotech'?

Because nanotech is guaranteed to work.

You seem to be ignoring the fact that knowledge of unobtainium's existence is less than 80 years old in any case, and that there are siple structures that today can not be reproduced which, while doing so would not prove economical, it would be a huge advancement in the related technology - no pure research turns a profit in the theoretical or development stage.

The 80 years is the time you have to wait to fuel your ISV, remember?

And there's your stopping point: "would not prove economical." Pure research will be funded if someone thinks it'll be economical, and unobtanium is basically the be-all and end-all of "Too expensive to gather naturally."

...so the maximum energy use is not going to increase, particularly not in a world without any usable farmland and an almost lack of natural light?

Go read that page again. It's a curve. The average now is nowhere near 890GJ/year, it's more like 100, if that. If everyone used 890GJ/year, then energy usage would increase dramatically, even without more-than-doubling the population.

[Human No More]

Originally Posted by Clarke

Well, yes, they are huge. Yet they still work as nano-manipulators, which you insisted were impossible earlier. Considering how powerful biotechnology is, there's your answer: gene-modify something to build Unobtanium for you.

Do you even know the difference in scale between a cell and a molecule of a compound?

Ding! With 10,000,000 lines, we can send 10MB per 20 minutes... or be clever about it and send 22kb per second.

If parallel data arrives at different times, it is only as fast as the last bit to arrive (error correcting code notwithstanding, but on the other hand, they only if a bit is actually received as all bits are needed to calculate if any are in error, so even that doesn't improve it).
Therefore there is always going to be a 20 minute lag time unless you wanted to send it serially multiple times, in which case you are limited to 3 bits/hour since there is no accurate enough method to maintain distributed state. Once again, considering the costs of $75,000/bit, you're even going against your original point of 'cheaper.

How can we require more technicians than SpecOps?

Ah... what?

Then bring the Avatars. Cuts into your savings, but that's going to happen regardless.

...so now you've made your ENTIRE argument redundant. Brilliant.

...Since the avatar link is psionic and apparently uninterruptible, are we sure that it doesn't double as FTL? In that case, we save yet more fuel/money: we don't have to bring the avatar drivers.

A theoretical maximum range of 11447km (diameter) is very different from one of 4.4ly, not to mention the required support infrastructure (technicians there, food, supplies, buildings).

Automated defenses as well. That's actually a lot easier than self-maintaining robots.

If I didn't know better, I'd think you are trolling. That fits in with minimising harm and impact how exactly?!

Also, the "impeded electronics" thing is not real physics full stop. A magnetic field large enough to even noticeably impede low-velocity electronics will do really, really bizarre things to biological tissue. ...And shielding is still lighter than humans.

Please don't tell me you're one of those 'Pandora's field would kill humans' people - just because JC made a physics error in an early interview, and the same one where he said Jake had no legs and got his name wrong, as well as calling Pandora a planet, does not prove your point. Electromagnetic interference DOES happen - we aren't talking about a total cease of operation, but lack of communication, primarily.

Temperatures up to 1GK can be generated in fusion reactors. (But that's not going to help to generate compounds. ) Magnetic fields up to 15T can be generated non-explosivly. Acceleration can't be generated in a chem. lab, beyond centrifuges, AFAIK, but that's not going to be true for much longer if unobtanium is discovered to require it. Anything else?

Radiation, gravity, planetary impacts, even the correct emission spectrum.

Because nanotech is guaranteed to work.

Yet IT DOES NOT EXIST. I genuinely do not understand why you are incapable of understanding this very simple point.

The 80 years is the time you have to wait to fuel your ISV, remember?

Actually, no,. In your rush to bash Avatar, you seem to have forgotten what you said in your previous post, which I was replying to.
Here's a reminder:

The majority of high-value compounds were discovered less than 80 years ago, e.g. the entire plastics and semiconductor industries.

Certainly, such a process may be economical, but it does not yet exist at that time. I fail to understand why you are incapable of recognising this simple fact.

And there's your stopping point: "would not prove economical." Pure research will be funded if someone thinks it'll be economical, and unobtanium is basically the be-all and end-all of "Too expensive to gather naturally."

FOR THE THIRD TIME: It likely is under research, but that does not mean it is going to instantly become possible just because they have started it a few years ago. In the meantime, it is either this source or nothing. Indeed, if they hadn't gone there, they wouldn't know of its existence in the first place.

[Ashen Key]

Clarke, you are also forgetting that humans like going places themselves. Even if it was entirely practical for only robots on Pandora (which it's not - robots are slower than humans, and so far we don't have AIs who can think for themselves; what the robots on Mars have done in six months, a human could have done in a week), humans want to go and do things themselves. Hold the actual stones in our hands, put footsteps on the ground, look at the sky from somewhere else.

If it was the slightest bit feasible, humans would go.

[Clarke]

Originally Posted by Ashen Key

If it was the slightest bit feasible, humans would go.

But the RDA is doing this for profit. They're not going to invest a sizable portion of the world's power supply FOR SCIENCE. Possibly FOR PR, but not purely for science. (It would have been a far more interesting film if Aperture Science were the ones with the Avatars. )

Originally Posted by Human No More

Do you even know the difference in scale between a cell and a molecule of a compound?

Yes. For smaller molecules it's about 3 orders of magnitude. So? What is a cell, if not a nanomachine?

If parallel data arrives at different times, it is only as fast as the last bit to arrive (error correcting code notwithstanding, but on the other hand, they only if a bit is actually received as all bits are needed to calculate if any are in error, so even that doesn't improve it).
Therefore there is always going to be a 20 minute lag time unless you wanted to send it serially multiple times, in which case you are limited to 3 bits/hour since there is no accurate enough method to maintain distributed state. Once again, considering the costs of $75,000/bit, you're even going against your original point of 'cheaper.

I bet someone's already come up with an algorithm that lets you "break" as soon as you've received the bare minimum amount of data. If they haven't, it's not going to take long for someone to do so. Regardless, you've conjured an order of magnitude: it costs the RDA $7,500/bit.

And since the major cost there is going to be scarcity, bringing 10 m/billion of the same thing is going to cost less than $7.5k a bit.

Actually, I'm not even bother going to bring 10 billion particle-pairs. It'd be easy enough to entangle a whole mol of photons/electrons. Might take a little while, but we're waiting to generate the fuel for our spaceship anyway. Now, if I've got a 1/10000 error rate per channel, and I have 6x10^23 channels...

Ah... what?

"you're still multiplying the on-world presence needed, as well as, as I mentioned before, abandoning all pretence of following the contract."

The only way I'm multiplying anything is if I'm bringing more humans than the original plan involved bringing, which I'm not. I'm only bringing enough to maintain the robots.

...so now you've made your ENTIRE argument redundant. Brilliant.

So you're saying...
(mining machines + large mercenary squad + Avatars) < (mining machines + few technicians + Avatars)

...How does that work?

A theoretical maximum range of 11447km (diameter) is very different from one of 4.4ly, not to mention the required support infrastructure (technicians there, food, supplies, buildings).

Booster stations! ...No, I'm being silly. Fine, no remote-Avatars. (Though if it is FTL over those 11Mm, then you can still use them as a very short-duration time machine. )

If I didn't know better, I'd think you are trolling. That fits in with minimising harm and impact how exactly?!

...Because you program your defenses to only shoot at things actively shooting at you? I thought that was obvious.

Please don't tell me you're one of those 'Pandora's field would kill humans' people - just because JC made a physics error in an early interview, and the same one where he said Jake had no legs and got his name wrong, as well as calling Pandora a planet, does not prove your point. Electromagnetic interference DOES happen - we aren't talking about a total cease of operation, but lack of communication, primarily.

Last time I checked, radio interference was only produced by changing magnetic fields, and if your planetary magnetic field is changing significantly, you've got bigger problems than EM interference. When you control all the radio sources on the planet, the only thing that should get in your way is LOS. And again, shielded robots will still be cheaper than humans; you can build shielded robots on-site.

And if the mountains are held up by magnetic fields, then he's right; it will kill anyone who comes anywhere near it.

1) Radiation, 2) gravity, 3) planetary impacts, even the 4) correct emission spectrum.

1) Easy. I mean, alpha/beta/gamma wavicles aren't hard to generate in large quantity.
2) Indistinguishable from acceleration. See centrifuge comment earlier.
3) ...And what does that involve? Radiation, pressure, heat. We've got all those covered.
4) Again, easy.

Actually, no,. In your rush to bash Avatar, you seem to have forgotten what you said in your previous post, which I was replying to.
Here's a reminder:

Certainly, such a process may be economical, but it does not yet exist at that time. I fail to understand why you are incapable of recognising this simple fact.

They've known about unobtanium, what, 30 years? Longer?

And simply, the method doesn't exist because it would spoil Cameron's story. There's no in-universe explanation given or possible.

FOR THE THIRD TIME: It likely is under research, but that does not mean it is going to instantly become possible just because they have started it a few years ago. In the meantime, it is either this source or nothing. Indeed, if they hadn't gone there, they wouldn't know of its existence in the first place.

They know about it because they saw the emission spectrum from Pandora. (Which doesn't actually make sense, but nevermind.) And, economically, "nothing" is the logical choice. No sane business would even touch the ISV possibility until it was proven to be impossible to synthesize Unobtanium.

[Ashen Key]

Originally Posted by Clarke

But the RDA is doing this for profit. They're not going to invest a sizable portion of the world's power supply FOR SCIENCE. Possibly FOR PR, but not purely for science. (It would have been a far more interesting film if Aperture Science were the ones with the Avatars. )

They spent billions of dollars on the Avatar program. That's...for SCIENCE, not that practical if all they are interested in is a hole in the ground. You need to sell the shares for profit, after all.

[Clarke]

Originally Posted by Ashen Key

They spent billions of dollars on the Avatar program. That's...for SCIENCE, not that practical if all they are interested in is a hole in the ground. You need to sell the shares for profit, after all.

Yeah, you see shares with Public Relations. I thought Selfridge was fairly clear that the avatars were only there to appease the environmentalists/there as a statement of "We really do care, honest guv!"

[Human No More]

Originally Posted by Clarke

Yes. For smaller molecules it's about 3 orders of magnitude. So? What is a cell, if not a nanomachine?

Actually, it's not. Cells still can not arbitrarily manipulate elements as you believe.

I bet someone's already come up with an algorithm that lets you "break" as soon as you've received the bare minimum amount of data. If they haven't, it's not going to take long for someone to do so. Regardless, you've conjured an order of magnitude: it costs the RDA $7,500/bit.

Again, that isn't possible. ECCs still require a value for every bit in order to calculate, and the larger the number of error bits, the far more likely it is that the entire packet has to be thrown out.

And since the major cost there is going to be scarcity, bringing 10 m/billion of the same thing is going to cost less than $7.5k a bit.

Yet it is directly stated to cost 7,500/bit. You can not avoid that - do you even have any idea how much manufacturing x billion of something so complex would cost anyway?

Actually, I'm not even bother going to bring 10 billion particle-pairs. It'd be easy enough to entangle a whole mol of photons/electrons. Might take a little while, but we're waiting to generate the fuel for our spaceship anyway. Now, if I've got a 1/10000 error rate per channel, and I have 6x10^23 channels...

I have already pointed out why this will not work, more than once. The success rate of getting the particle to adopt the desired state is so low that the higher the particle count, the more you are boosting interference - it would not uniformly apply or fail on every single particle at once. Dealing with multiple particle sets in a single encoding also adds huge amounts of complexity in terms of reading the resulting particle from each.

The only way I'm multiplying anything is if I'm bringing more humans than the original plan involved bringing, which I'm not. I'm only bringing enough to maintain the robots.

...not to mention the supplies needed, the support staff for those humans (everything from maintenance to food), the marines you completely ignored, some kind of administration presence.

So you're saying...
(mining machines + large mercenary squad + Avatars) < (mining machines + few technicians + Avatars)

...How does that work?

You're ignoring my point that without any defence, any mining will be easily countered. The avatars were for the Na'vi, but are not going to do anything with an 'angtsik.

...Because you program your defenses to only shoot at things actively shooting at you? I thought that was obvious.

I really don't think you understand the logic of complexity here.

Last time I checked, radio interference was only produced by changing magnetic fields, and if your planetary magnetic field is changing significantly, you've got bigger problems than EM interference.

Also, via moving through one.

And again, shielded robots will still be cheaper than humans; you can build shielded robots on-site.

So now you need large quantities of rare-earths as well. What makes you believe there even might be any there?

1) Easy. I mean, alpha/beta/gamma wavicles aren't hard to generate in large quantity.
2) Indistinguishable from acceleration. See centrifuge comment earlier.
3) ...And what does that involve? Radiation, pressure, heat. We've got all those covered.
4) Again, easy.

Emphasis on a macroscopic scale.

They've known about unobtanium, what, 30 years? Longer?

EXACTLY. Thank you. Based on your 'it will always take 80 years to synthesise anything' claim, there is still ~40-50 years to go even if that was true.
I'm finding it hard to argue with someone as logically inconsistent as you.
You go from "they should be able to produce it" with no justification to "it will take 80 years form discovery" to "they have known of it for at least 30 years so they should be producing it". All three premises are flawed and you are jumping between them as you see fit, backing further into a logical contradiction each time I counter the fallacious reasoning behind each other than the second.

And simply, the method doesn't exist because it would spoil Cameron's story. There's no in-universe explanation given or possible.

No, you just want one to exist because it would create the story you want to see.

They know about it because they saw the emission spectrum from Pandora.

Source?

And, economically, "nothing" is the logical choice. No sane business would even touch the ISV possibility until it was proven to be impossible to synthesize Unobtanium.

...which is has so far in that all attempts, if they have even reached a practical attempt - which is extremely unlikely - have failed.

[Clarke]

Originally Posted by Human No More

Actually, it's not. Cells still can not arbitrarily manipulate elements as you believe.

Well, yes, no existing cell can. Why can't cells in general?

Again, that isn't possible. ECCs still require a value for every bit in order to calculate, and the larger the number of error bits, the far more likely it is that the entire packet has to be thrown out.

I mean that you set up your algorithm to abort as soon as possible.

Yet it is directly stated to cost 7,500/bit. You can not avoid that - do you even have any idea how much manufacturing x billion of something so complex would cost anyway?

The idea is that you build one/two devices, and just multiply your medium. (Which is easy, since it's just trapped electrons.) In the same way, you don't have 1 billion RAM chips; you have one chip that holds 1 billion switches.

I have already pointed out why this will not work, more than once. The success rate of getting the particle to adopt the desired state is so low that the higher the particle count, the more you are boosting interference - it would not uniformly apply or fail on every single particle at once. Dealing with multiple particle sets in a single encoding also adds huge amounts of complexity in terms of reading the resulting particle from each.

...What? The state influence is done by electric currents. (which is what an oscillating magnetic field is) That would not be affected by there being more particles in the field at once, unless physics specifically abhors FTL machinery. It'd be, "1/1000 particles are affected", not, "Exactly one particle is affected, and we just have 10,000 available."

...not to mention the supplies needed, the support staff for those humans (everything from maintenance to food), the marines you completely ignored, some kind of administration presence.

...What marines? For that matter, what administrative presence? How's that going to help anything? You don't want to send anyone up there who isn't generating either money or PR.

You're ignoring my point that without any defence, any mining will be easily countered. The avatars were for the Na'vi, but are not going to do anything with an 'angtsik.

And the marines help with this over semi-automated defenses... how?

I really don't think you understand the logic of complexity here.

Computer programs have zero mass. Also, they can be updated from Earth.

If you're talking about hardware, how would that be more complex than, say, the Avatar linker machinery?

Also, via moving through one.

True, but that effect is incredibly minor when your speed is under 100m/s.

So now you need large quantities of rare-earths as well. What makes you believe there even might be any there?

Faraday cages are built out of any conductor.

Emphasis on a macroscopic scale.

...Yes? I don't see how anything about what I said falls down on a macro (i.e. visible) scale.

EXACTLY. Thank you. Based on your 'it will always take 80 years to synthesise anything' claim, there is still ~40-50 years to go even if that was true.

I said, it'll take 80 years at most. Since there is major business interest behind it, it'll be less.

I'm finding it hard to argue with someone as logically inconsistent as you.
You go from "they should be able to produce it" with no justification to "it will take 80 years form discovery" to "they have known of it for at least 30 years so they should be producing it". All three premises are flawed and you are jumping between them as you see fit, backing further into a logical contradiction each time I counter the fallacious reasoning behind each other than the second.

Just to clear this up:
Regardless of timeframe, there would be no point in sending a starship unless it'd been proven impossible to synthesize Unobtanium. Any starship you do send would be purely for science, not for profit.
If Unobtanium was discovered more than 30 years ago, they probably would have a synthesis process, because of the vast business interest behind it.

The second premise was never an argument in the first place.

No, you just want one to exist because it would create the story you want to see.

...It existing wouldn't spoil Cameron's story? Also, it's nothing to do with the story I want to see, beyond the fact that I want to see logically coherent plots.

Source?

I read wrong, apparently.

...which is has so far in that all attempts, if they have even reached a practical attempt - which is extremely unlikely - have failed.

Then the ISVs don't exist. Think about that business proposal for a minute.

Imagine the ISV is actually built, ready, and we have the infrastructure to fuel it in 6 years. Then, it takes 5.5 years to actually get to Pandora. Then there's... what, 3 years, at minimum? to actually mine a ship's load of Unobtanium, which then takes another 5.5 years to ship back. That adds up to 20 years, for one shipment, which costs the majority of Earth's power for all of those 6 years.

Marketing can't tell you anything about the price of Unobtanium 20 years in the future. Why on earth would you accept that as a business idea?

[Moco Loco]

Originally Posted by Moco Loco

OHOHO, wait until HNM sees THIS.

I called it. Didn't I call it?

[Human No More]

Originally Posted by Clarke

Well, yes, no existing cell can. Why can't cells in general?

Cells are too large, primarily.

I mean that you set up your algorithm to abort as soon as possible.

I that case, you are still facing 20 minute latency, if you are using it in a similar manner to Bluetooth - transmitting the data over many channels and majority vote for error correction. In the other case, there still needs to be at least one complete packet received, with a per-byte latency of 20 minutes (meaning slightly under 3 weeks for an average-sized packet, in a single direction - double that for return).

The idea is that you build one/two devices, and just multiply your medium. (Which is easy, since it's just trapped electrons.) In the same way, you don't have 1 billion RAM chips; you have one chip that holds 1 billion switches.

I have explained this at least four times.
The higher the particle count, the more particles you have that do not adopt the desired state (0.1% success rate), which, if they are used to encode the same bit, cause a huge degradation in signal to noise ratio.

...What? The state influence is done by electric currents. (which is what an oscillating magnetic field is) That would not be affected by there being more particles in the field at once, unless physics specifically abhors FTL machinery. It'd be, "1/1000 particles are affected", not, "Exactly one particle is affected, and we just have 10,000 available."

No, the success rate is 0.1%, not one out of every 10,000 particles is capable of being encoded. If the latter was true, there would not be such a limitation on bitrate. Face it, trying to argue network mechanics is going to get you nowhere here.
If you want me to give you a simplified explanation, I can do that.

...What marines? For that matter, what administrative presence? How's that going to help anything? You don't want to send anyone up there who isn't generating either money or PR.

1. Defence, as mentioned (and not, do not just say 'hurr durr robots' again.
2. It won't run itself - even if it's a single person, some decisionmaking authority is needed there if there are to be humans.

And the marines help with this over semi-automated defenses... how?

They are actually there and can make judgements, not be programmed to shoot at anything and everything ("semi-" automated now? with a 3 byte/hour control bandwidth which, when used, blocks all other FTL communications? Brilliant... I really could not make this stuff up if I was trying to come up with impractical ideas.

Computer programs have zero mass. Also, they can be updated from Earth.

Wither either a 4.4 year light speed latency, or 3 bit/h FTL. They would have to rely on 100% automated decisionmaking. Yes, complexity refers to software, to decisionmaking and control, and to telemetry.

Faraday cages are built out of any conductor.

So you're adding an even greater mining effort? Also, a Faraday cage would also block any communication for remote control or interlinking between units.

...Yes? I don't see how anything about what I said falls down on a macro (i.e. visible) scale.

OK, I'll rephrase it - it is theoretically possible that:
1. If the conditions were actually known
2. Large investment was found
3. The structure proved replicable
then it could potentially be synthesised - but doing so will take time, and in them meantime there is a source there which does prove economical, even if more marginal than a potential domestically produced source - but since it is producing a profit, they use it for the meantime - once the unobtainium is in place, it continually produces profit/savings without further investment, in terms of playing the long game.

I said, it'll take 80 years at most. Since there is major business interest behind it, it'll be less.

...and it hasn't been anywhere more than 40 at the very most.
It is potentially possible it may eventually be done, but it has not at this point. See above.

Just to clear this up:
Regardless of timeframe, there would be no point in sending a starship unless it'd been proven impossible to synthesize Unobtanium. Any starship you do send would be purely for science, not for profit.
If Unobtanium was discovered more than 30 years ago, they probably would have a synthesis process, because of the vast business interest behind it.

What about the vast business interests today that lack such processes after over 100 years?

If there is a source that proves recoverable ad profitable, even marginal in the short term, remember, it will increase exponentially over the long term once in place due to the actual applications. IF (not when) they do ever find a way to synthesise it, it may replace the offworld source if it proves more economical.

The second premise was never an argument in the first place.

...It existing wouldn't spoil Cameron's story? Also, it's nothing to do with the story I want to see, beyond the fact that I want to see logically coherent plots.

Again, since you do not seem to understand the definition, it is logically coherent. There is no contradiction in universe that would render it impossible or goes against other facts/figures. You're right, adding your own pet plot elements would ruin the story, but there is no obligation to add them. You are not the storyline dictator of the world, and it remains plausible without them (indeed, many of yours are actually highly implausible or both the timeframe and discovery period).

Then the ISVs don't exist. Think about that business proposal for a minute.

"We're researching synthesising it, with extremely uncertain results and time for return, costing us billions in research money with no immediate return, but in the mean time, there is a convenient if rather more marginal offworld source"
Seems like it would work to me. Plus, there was ALREADY an ISV in order to get there in the first place - they had likely committed to more well before the fact (you never order just one of anything), and all unobtainium actually did was improve their efficiency.

Originally Posted by Moco Loco

I called it. Didn't I call it?

Very true. Ironically, I didn't even have an issue with the OP so much.

[Clarke]

Originally Posted by Human No More

Cells are too large, primarily.

Yet they can build DNA just fine.

I that case, you are still facing 20 minute latency, if you are using it in a similar manner to Bluetooth - transmitting the data over many channels and majority vote for error correction. In the other case, there still needs to be at least one complete packet received, with a per-byte latency of 20 minutes (meaning slightly under 3 weeks for an average-sized packet, in a single direction - double that for return).

We're not going to be using standard TCP over an FTL connection if that's inefficient for whatever reason.

I have explained this at least four times.
The higher the particle count, the more particles you have that do not adopt the desired state (0.1% success rate), which, if they are used to encode the same bit, cause a huge degradation in signal to noise ratio.


No, the success rate is 0.1%, not one out of every 10,000 particles is capable of being encoded. If the latter was true, there would not be such a limitation on bitrate. Face it, trying to argue network mechanics is going to get you nowhere here.
If you want me to give you a simplified explanation, I can do that.

I want the full explanation, since what you're saying isn't making a lot of sense and/or appears to have obvious workarounds.

1. Defence, as mentioned (and not, do not just say 'hurr durr robots' again.
2. It won't run itself - even if it's a single person, some decisionmaking authority is needed there if there are to be humans.

I'll come back to the hurr durr robots.
...Yes it will? Your technicians are skilled, aren't they? You don't trust them to make decisions themselves? If not, are you making the administrative leader pay his 1.8TW fuel cost?

They are actually there and can make judgements, not be programmed to shoot at anything and everything ("semi-" automated now? with a 3 byte/hour control bandwidth which, when used, blocks all other FTL communications? Brilliant... I really could not make this stuff up if I was trying to come up with impractical ideas.

...I said this, just a few posts ago. You don't program your computers to shoot everyone. You program them with whatever type of fidelity you want, with them even having the option to consult the humans on base, so they won't shoot things you don't want them to shoot. The defense computers have nigh infinite memory, processing power and sensory data available to them compared to the slow, squishy, uncoordinated, vulnerable marines. The computer is superior to a human in almost all respects here, particularly since it would be an order of magnitude cheaper to ship there, since you can build most of it on-site!

Wither either a 4.4 year light speed latency, or 3 bit/h FTL. They would have to rely on 100% automated decisionmaking. Yes, complexity refers to software, to decisionmaking and control, and to telemetry.

All of which is not only massless, but can be tested beforehand?

...And they don't have to be 100% automated. They can ask the humans on base.

So you're adding an even greater mining effort? Also, a Faraday cage would also block any communication for remote control or interlinking between units.

Fake-psionics? (i.e. exploit whatever's powering the Avatars.) LOS optics? (Not the most practical, but it'd be completely unaffected by EM.) You can also use optic fibre to break the cage, which means you only have to shield your radio transmitter from EM interference, rather than everything.

OK, I'll rephrase it - it is theoretically possible that:
1. If the conditions were actually known
2. Large investment was found
3. The structure proved replicable

1) These can be calculated. We can do quantum electrodynamics now, (if slowly) let alone in 140 years.
2) Guaranteed, for all the reasons you suggest.
3) Guaranteed again, because you can use nanotech if absolutely everything else fails.

then it could potentially be synthesised - but doing so will take time, and in them meantime there is a source there which does prove economical, even if more marginal than a potential domestically produced source - but since it is producing a profit, they use it for the meantime - once the unobtainium is in place, it continually produces profit/savings without further investment, in terms of playing the long game.

If you can't guarantee Unobtanium will not drop in price even a smidge, let alone be synthesized in any quantity, within 20 years, then it is a catastrophic business proposal. It is simply not good business sense to gamble that amount of resources on something you know very little about.
Or are you forgetting that this involves multiples of the entire planet's power supply?

...and it hasn't been anywhere more than 40 at the very most.
It is potentially possible it may eventually be done, but it has not at this point. See above.

See above.

What about the vast business interests today that lack such processes after over 100 years?

Such as? I don't know of one. Also, is the substance in question worth $20m/kg? Because gold, AFAIK one of the most expensive substances that can't be synthesized, is worth about 1/350th of that.

If there is a source that proves recoverable ad profitable, even marginal in the short term, remember, it will increase exponentially over the long term once in place due to the actual applications. IF (not when) they do ever find a way to synthesise it, it may replace the offworld source if it proves more economical.

'Scuse me, money saved increases linearly with time installed. Incidentally, a synthesis process will prove more economical, energy-wise. Basic physics.

Again, since you do not seem to understand the definition, it is logically coherent. There is no contradiction in universe that would render it impossible or goes against other facts/figures. You're right, adding your own pet plot elements would ruin the story, but there is no obligation to add them. You are not the storyline dictator of the world, and it remains plausible without them (indeed, many of yours are actually highly implausible or both the timeframe and discovery period).

• They are there for profit, i.e. some economist decided that spending the fuel of the Venture Star in return for Unobtanium would be profitable.
• This is almost completely impossible, due to the canon engineering involved.

There's your contradiction. You can resolve it by bringing in elements (i.e. Dyson shell fragments) that are not only not alluded to, but that Cameron has never featured before and so probably would never think of if he even wanted to be rigidly consistent.

"We're researching synthesising it, with extremely uncertain results and time for return, costing us billions in research money with no immediate return, but in the mean time, there is a convenient if rather more marginal offworld source"
Seems like it would work to me. Plus, there was ALREADY an ISV in order to get there in the first place - they had likely committed to more well before the fact (you never order just one of anything), and all unobtainium actually did was improve their efficiency.

"More marginal?" How about 250 million times more expensive, given impossibly perfect machinery? You don't have impossibly perfect machinery? Tough, it's going to cost you even more! Billions of dollars is pittance compared to the energy expenditure of fueling antimatter-driven spacecraft.

Also, Avatar-wiki suggests it was an unmanned probe, not a manned ISV. This could be made cheap enough to be worth it, especially if it was one-way.

Re: ordering more than one ISV; all my arguments above assumed that the ISV, a multi-km long, 100-ton spacecraft was free. That's how ridiculous the economics involved are.

Very true. Ironically, I didn't even have an issue with the OP so much.

I'm surprised, since you're suggesting a business plan that takes 80 years for a one-way trip is somehow good business.

[Human No More]

Originally Posted by Clarke

We're not going to be using standard TCP over an FTL connection if that's inefficient for whatever reason.

Hence the comparison for an average packet - that is 1500 bytes, which is still an extremely small amount of data even if sent using a custom protocol or even plaintext only (and addressing would still be needed if there was anything more than one destination for it). Remember that a bit is 12.5% of a single character.

I want the full explanation, since what you're saying isn't making a lot of sense and/or appears to have obvious workarounds.

From background: A state can be induced in a particle. There is a 0.1% success rate of the state being correctly induced, so if there were two particles being used in the same system at each end, the chance of both being induced would be (1x10^-6)%. If you use multiple particles and they are not all induced, you are reducing the probability that the particle with the desired state can be read (indeed, if it was even possible to determine which particle was the one with the desired state.

If you build multiple communication systems to get around this then you are adding to the cost at $7500/bit, which presumably includes actual production costs as well as running. In addition, if you had 100, for example, and you were transmitting in parallel, each packet would need a fragment offset to state where it is in the datastream, increasing overhead. Since the bitrate is stated as 3 per hour, that makes an average latency of 20 minutes even if you transmitted every single bit in parallel, and every bit would be required in order to assemble the packet - even with a highly redundant ECC scheme (which would add large amounts of overhead compared to a simple checksum, possibly actually reducing overall throughput), it can only compensate for a very small number, and generally only error bits - not missing bits entirely, so the receiver must wait until the entire packet is received. All bits are used to checksum specific other bits - as such, an error bit will cause a checksum conflict for other bits, which, using the correct pattern, allows the error bit to be determined. If there is no value for the error bit at all (not to be confused with 0, which is a value), it can not be calculated.

...I said this, just a few posts ago. You don't program your computers to shoot everyone. You program them with whatever type of fidelity you want, with them even having the option to consult the humans on base, so they won't shoot things you don't want them to shoot.

Until recently, you said there should be no humans there. You're moving the goalposts again.

The defense computers have nigh infinite memory, processing power and sensory data available to them compared to the slow, squishy, uncoordinated, vulnerable marines. The computer is superior to a human in almost all respects here, particularly since it would be an order of magnitude cheaper to ship there, since you can build most of it on-site!

So you're implying there will be true AI? Computers NOW have more memory and processing power in raw floating point operations than humans, but can not make value judgements.

All of which is not only massless, but can be tested beforehand?

...And they don't have to be 100% automated. They can ask the humans on base.

Again, you've now switched to there being humans there. Indeed, doing so would require some kind of decisionmaking authority anyway.

Fake-psionics? (i.e. exploit whatever's powering the Avatars.) LOS optics? (Not the most practical, but it'd be completely unaffected by EM.) You can also use optic fibre to break the cage, which means you only have to shield your radio transmitter from EM interference, rather than everything.

The avatar link is intentionally vague, and clearly closely matched with the operator (the entire reason Jake was hired), with absolutely zero indication it works outside of wetware. LOS optics are certainly impractical - the point is that, from their view, they are using technology that does just about work there, but with problems, without having to develop anything new.

1) These can be calculated. We can do quantum electrodynamics now, (if slowly) let alone in 140 years.
2) Guaranteed, for all the reasons you suggest.
3) Guaranteed again, because you can use nanotech if absolutely everything else fails.

Yes - as I said, it's theoretically possible, but such provisions are not immediate.

If you can't guarantee Unobtanium will not drop in price even a smidge, let alone be synthesized in any quantity, within 20 years, then it is a catastrophic business proposal. It is simply not good business sense to gamble that amount of resources on something you know very little about.
Or are you forgetting that this involves multiples of the entire planet's power supply?

...according to your own figures, which are not consistent with the ones in canon. I will remind you, there is never any specific figure given in canon, so it can be assumed to fit whatever else is stated.

Again, saying "we may or may not be able to synthesise it - it may take 80 years, it may take too years, but in the meantime there is a more marginal supply we can reach here" will work, just as it does on Earth. Certainly, there are higher costs, but also a higher return.

Such as? I don't know of one. Also, is the substance in question worth $20m/kg? Because gold, AFAIK one of the most expensive substances that can't be synthesized, is worth about 1/350th of that.

Again - not a lack of synthesis at all, but even a lack of economical processes to do so. Producing a small quantity in perfect lab conditions with huge funding is different form producing large quantities with an eye to a profit.

'Scuse me, money saved increases linearly with time installed. Incidentally, a synthesis process will prove more economical, energy-wise. Basic physics.

It's exponential if conditions remain constant - using superconductors for energy distribution, for example, all but removes any upper limit on transmission from losses.

"More marginal?" How about 250 million times more expensive, given impossibly perfect machinery?

...so now you havean exact idea of the costs of synthesis?

You don't have impossibly perfect machinery? Tough, it's going to cost you even more! Billions of dollars is pittance compared to the energy expenditure of fueling antimatter-driven spacecraft.

Throwing money at research will not cause it to gain instant results.

Also, Avatar-wiki suggests it was an unmanned probe, not a manned ISV. This could be made cheap enough to be worth it, especially if it was one-way.

Fair enough, it actually does, although there is no source. Either way, if so, the fact that they were actually looking specifically for extrasolar resources implies that even if the first ISV was not already under construction, it was on the drawing board.

[quote]Re: ordering more than one ISV; all my arguments above assumed that the ISV, a multi-km long, 100-ton spacecraft was free. That's how ridiculous the economics involved are. [quote]
It isn't as expensive as you think it might be. This is not the Enterprise - the most expensive parts would be getting the components into orbit (not that expensive with the clear availability of SSTO craft) and the engine superconductors.

I'm surprised, since you're suggesting a business plan that takes 80 years for a one-way trip is somehow good business.

You mean the whole 'synthesis' thing again? No, that was your argument.

[Sempu]

FYI, Vinton Cerf worked with NASA to develop long-latency network protocols for a kind of "TCP for the solar system" to link the spacecraft out there. Should be easy to find with some searching.

[Clarke]

Originally Posted by Human No More

Hence the comparison for an average packet - that is 1500 bytes, which is still an extremely small amount of data even if sent using a custom protocol or even plaintext only (and addressing would still be needed if there was anything more than one destination for it). Remember that a bit is 12.5% of a single character.

Why would there be more than one destination for it? Anyway, the networking only becomes important if the rate is 3b/hr, which is silly. See next paragraph.

From background: A state can be induced in a particle. There is a 0.1% success rate of the state being correctly induced, so if there were two particles being used in the same system at each end, the chance of both being induced would be (1x10^-6)%. If you use multiple particles and they are not all induced, you are reducing the probability that the particle with the desired state can be read (indeed, if it was even possible to determine which particle was the one with the desired state.

Ah, the solution to that is quite simple: induce the state again, and then measure it, and then try again if the induction failed. Try this a few thousand times, and suddenly your success rate rises to arbitrarily high levels. Ideally, you will want to treat all particles like this in parallel, so you don't disturb any particle you've already successfully set.

Because this takes non-zero time, you will have to decrease your "clock" frequency at the other end, perhaps to as high as 1 second. However, after that 1 second, >99% of the bits will be accurate. If you have enough particles, you can send entire packets in that one second. (Though I'm fairly confident a 1 second clock frequency is far too long, but that depends on how accurate you want to be, and how fast your electromagnets are.)

Oh, just to throw a spanner in the works: did you know FTL comms are functionally equivalent to a time machine, and so the RDA is not going to even bother with unobtanium if they have one? The stock market will make them rich enough.

If you build multiple communication systems to get around this then you are adding to the cost at $7500/bit, which presumably includes actual production costs as well as running.

Why would that be? We don't usually include startup cost in that sort of figure.

In addition, if you had 100, for example, and you were transmitting in parallel, each packet would need a fragment offset to state where it is in the datastream, increasing overhead.

Your offsets would be pre-arranged? That honestly seems obvious.

[error correction - snipped because of char limits]

Some EC schemes let you infer missing or corrupt bits.

Until recently, you said there should be no humans there. You're moving the goalposts again.

I don't think I've said there were no humans there, only less than the RDA were putting there. You'd need perhaps 5-10, rather than the hundreds shown.

So you're implying there will be true AI? Computers NOW have more memory and processing power in raw floating point operations than humans, but can not make value judgements.

Computers now can play Jeopardy, and drive a car over 1,000 miles. I'm not going to even try to guess what they can do 140 years from now, since the concept of AI has only been around half that.

And they can make whatever value judgements you want them to make, so long as you're specific. As mentioned, they can drive cars. (True AI would be overkill. You don't want it to talk to you, you want it to defend you.)

Again, you've now switched to there being humans there. Indeed, doing so would require some kind of decisionmaking authority anyway.

So assign one or two as leaders. Having someone who does nothing but leading is a tremendous waste of resources.

The avatar link is intentionally vague, and clearly closely matched with the operator (the entire reason Jake was hired), with absolutely zero indication it works outside of wetware.

Then synthesize wetware at both ends. You have genetic engineering cracked, a few colonies of neurons should be fairly trivial.

LOS optics are certainly impractical - the point is that, from their view, they are using technology that does just about work there, but with problems, without having to develop anything new.

But why would you skimp on a few months of R&D when it saves you astronomical quantities of energy, and thus money?

Yes - as I said, it's theoretically possible, but such provisions are not immediate.

They are immediate on the scale of interstellar flight. It takes over 11 years for any investment in ISVs to make a return. Businesses live and die in 11 years. Entire new technological paradigms appear in 11 years. The original iPod did not exist 11 years ago. The internet was not in widespread use 11 years ago.

And that rate of technological change is increasing. It is not good business to gamble unless you're relatively sure that it will pay off. You can't possibly be sure that an ISV will pay off, hence no business would endorse you.

...according to your own figures, which are not consistent with the ones in canon. I will remind you, there is never any specific figure given in canon, so it can be assumed to fit whatever else is stated.

Yes, but in the case of the energy I'm talking being available, there are far more useful things to spend it on than unobtanium.

Again, saying "we may or may not be able to synthesise it - it may take 80 years, it may take too years, but in the meantime there is a more marginal supply we can reach here" will work, just as it does on Earth. Certainly, there are higher costs, but also a higher return.

It doesn't work here on Earth, when the costs are high enough. If it did, the asteroid belt wouldn't be around any more. There's gold in themthar rocks, it just takes ludicrous funding to get to it.

Again - not a lack of synthesis at all, but even a lack of economical processes to do so. Producing a small quantity in perfect lab conditions with huge funding is different form producing large quantities with an eye to a profit.

Yes; the former suggests the latter is possible.

It's exponential if conditions remain constant - using superconductors for energy distribution, for example, all but removes any upper limit on transmission from losses.

I think there is an upper limit on the DC current you can put through a superconductor before it stops superconducting, but I can't find anything on that. However, there's definitely an upper limit on AC current; superconductors fail when exposed to a powerful enough magnetic field.

And it's still linear if conditions remain constant. The difference in running costs between superconducting and non-superconducting infrastructure doesn't depend on how long that structure has been there for.

...so now you havean exact idea of the costs of synthesis?

Synthesizing a molecule requires arranging electrons. This has an energy cost of maybe a few hundred to a few thousand electron-volts per atom. Manufacturing antimatter requires constructing particles out of the vacuum; this requires hundreds of millions of electron-volts per atom. For every atom involved in the construction of unobtanium, you need either 1) 900,000keV of antimatter or 2) 2-3keV of electron rearranging.

This is a hard limit produced by the laws of physics, and cannot be avoided. In what possible universe is option 1 more economical than option 2?

Throwing money at research will not cause it to gain instant results.

True enough. So be patient. It's going to save, oh I don't know, a few billion watts, pretty much indefinitely. (With savings per second probably measured in thousands of dollars.) Your shareholders will thank you, if nothing else.

Fair enough, it actually does, although there is no source. Either way, if so, the fact that they were actually looking specifically for extrasolar resources implies that even if the first ISV was not already under construction, it was on the drawing board.

Well, yes, but that doesn't mitigate the completely ludicrous energy requirements the ISV needs.

It isn't as expensive as you think it might be. This is not the Enterprise - the most expensive parts would be getting the components into orbit (not that expensive with the clear availability of SSTO craft) and the engine superconductors.

It's expense is irrelevant. It doesn't matter if you have a replicator that can produce the thing for free; it's fuel cannot be produced with any sort of business sense.

You mean the whole 'synthesis' thing again? No, that was your argument.

No, I mean the figure quoted in the OP as the time you have to wait to even launch your spacecraft. I possibly need to redo that calculation, though. How is the ISV getting back again?

[Human No More]

Originally Posted by Clarke

Why would there be more than one destination for it? Anyway, the networking only becomes important if the rate is 3b/hr, which is silly. See next paragraph.

You need if it you want your 'hurrrrrr remote control the robots' argument.

Ah, the solution to that is quite simple: induce the state again, and then measure it, and then try again if the induction failed. Try this a few thousand times, and suddenly your success rate rises to arbitrarily high levels. Ideally, you will want to treat all particles like this in parallel, so you don't disturb any particle you've already successfully set.

Yes, in parallel - requiring a different multi-billion dollar system for each one. I thought you were the one complaining about cost, now you've switched track to fantasising about the film you wanted to see?

Oh, just to throw a spanner in the works: did you know FTL comms are functionally equivalent to a time machine, and so the RDA is not going to even bother with unobtanium if they have one? The stock market will make them rich enough.

Without distances on an astronomical scale, the time change is less than a single clock cycle. You can not do anything on a single clock cycle other than the very basic instructions. If they do attempt to use it as such, then the data needs to return, with a 4.4 year light speed lag.

Why would that be? We don't usually include startup cost in that sort of figure.

I said potentially. If it isn't, that just digs your hole even deeper.

Your offsets would be pre-arranged? That honestly seems obvious.

Only if each bit had a defined exact time to transmit or each transmitter was only used for a single bit per packet.

Some EC schemes let you infer missing or corrupt bits.

I know. Don't profess to greater knowledge here than me - each error bit reduces the chance of detection of others, and increases overhead - in usual applications, bandwidth is far less of a concern than noise itself, as the more complicated an ECC implementation, the more data overhead it creates.

I don't think I've said there were no humans there, only less than the RDA were putting there. You'd need perhaps 5-10, rather than the hundreds shown.

That seems marginally more realistic an assumption, but it still doesn't show that other humans should not be there.

And they can make whatever value judgements you want them to make, so long as you're specific. As mentioned, they can drive cars. (True AI would be overkill. You don't want it to talk to you, you want it to defend you.)

Depends on what's at stake. Determining the difference between lifeforms and making a judgement that could have such huge repercussions is very different from following defined traffic laws (and assuming that others will as well).

Then synthesize wetware at both ends. You have genetic engineering cracked, a few colonies of neurons should be fairly trivial.

That may potentially work, but there has to be some reason it wasn't used - there could still be a reason it doesn't work - even if it's signal strength or light speed lag.

But why would you skimp on a few months of R&D when it saves you astronomical quantities of energy, and thus money?

Hang on, I thought we were talking about communication within the magnetic field here. Stop jumping form one topic to another for the same argument.

They are immediate on the scale of interstellar flight. It takes over 11 years for any investment in ISVs to make a return. Businesses live and die in 11 years. Entire new technological paradigms appear in 11 years. The original iPod did not exist 11 years ago. The internet was not in widespread use 11 years ago.

I'd contest the second, certainly, but that's beside the point. A rate at which things are developed does not mean any research will instantly complete, ever.

And that rate of technological change is increasing. It is not good business to gamble unless you're relatively sure that it will pay off. You can't possibly be sure that an ISV will pay off, hence no business would endorse you.

...unless they were looking for minerals available outside Earth's solar system, exactly as they were. They most likely already are utilising resources found elsewhere in Earth's system including asteroids and potentially comets, the next logical step is further. No resource in the solar system lasts forever even if humanity WAS planning to only ever live there.

Yes, but in the case of the energy I'm talking being available, there are far more useful things to spend it on than unobtanium.

Useful? Possibly (yet you fail to give any examples) - but profitable?
Surprisingly, there isn't unlimited use for energy anyway.

It doesn't work here on Earth, when the costs are high enough. If it did, the asteroid belt wouldn't be around any more. There's gold in themthar rocks, it just takes ludicrous funding to get to it.

...and when it is reachable?

Yes; the former suggests the latter is possible.

Yet it doesn't instantly make it possible. As far as we can see, the former has not happened yet, and while the latter MAY be possible, it's still a long way off if so.

I think there is an upper limit on the DC current you can put through a superconductor before it stops superconducting, but I can't find anything on that. However, there's definitely an upper limit on AC current; superconductors fail when exposed to a powerful enough magnetic field.

That level also varies by superconductor, with some available today requiring >100T to do so.

And it's still linear if conditions remain constant. The difference in running costs between superconducting and non-superconducting infrastructure doesn't depend on how long that structure has been there for.

No, but if the amount of said infrastructure is growing, the savings are too.

Synthesizing a molecule requires arranging electrons. This has an energy cost of maybe a few hundred to a few thousand electron-volts per atom. Manufacturing antimatter requires constructing particles out of the vacuum; this requires hundreds of millions of electron-volts per atom. For every atom involved in the construction of unobtanium, you need either 1) 900,000keV of antimatter or 2) 2-3keV of electron rearranging.

This is a hard limit produced by the laws of physics, and cannot be avoided. In what possible universe is option 1 more economical than option 2?

For about the fifth time, option 2 is not an option at this point in time.

It may well be in the future, but with the current state of understanding at that time, it is either flat out not yet possible, or possible but not on a macroscopic scale. You are clearly underestimating energy availability via the simple fact that they are there (and for that matter, who is to say they haven't found a natural source of antimatter much like the recent belt around Earth, albeit far larger? - their understanding of physics will clearly be far more advanced than a 2011 one).