Getting a free strlen() is NOT an advantage, by the way. In fact, that became a liability when UTF-8 arrived. With a library strlen() function, all you had to do was update the library, but when the compiler was hardcoded to just return the byte count, that wasn't an option. Sure, one could go to UTF-16 instead, but then there's a lot of wasted space.
You've made an error here. There are some Unicode codepoints that are 32-bits long in UTF-16 (surrogate pairs). So, no. Not even with UTF-16 can you get a hardcoded version. However, if you go to UTF-32, then you will have enough space to ensure that every codepoint is represented in the same size value, and thus it's a simple matter of (buffersize in bytes / 4) = number of codepoints. (Note: still not the total number of "characters" as some are combining... and thus an "a" + "combining accent" is one character)
"But who would waste all that space just to make calculations faster?" Actually, Perl represents all Unicode in UTF-32 (native byte-alignment) and renders out to UTF-8 when printed. Precisely because space is relatively cheap, and processing time is still often the current limiting factor.
I'm correcting myself here... apparently they weren't considering going with a 255-byte limit, but a 65535-byte limit, which would have increased the size overhead by one.
Which is worse? Having it be O(N) to get a string length and having inexperienced programmers get confused and make mistakes? Or capping your maximum string length at 0xffff?
I'll take the former, please. I do a lot of string manipulation in C and when you're used to it, it's actually not that bad to get right and still be efficient. And it provides a useful shibboleth to detect people who are no good at C.:-) Just think of how much harder it would be to interview a C programmer if you couldn't give them a crazy string manipulation problem.
I don't know why you have to bring sibolezes into this...
"...it would have had the insane limit of 255-byte strings, with no compatible way to have a string any longer."
Compatible with what? Seems to me they could have just used continuation bit for the size field, much the way UTF-8 works to store non-ASCII characters.
This would still make the strings incompatible, because you would only have a 127-byte string length before the "continuation bit" comes into play and you need to switch to a 15-bit string length. All the previous code written with longer-than-127-byte strings would be incompatible.
Not to mention the argument for "because space was at a premium" is specious, because either you had a 8-bit length prepended to the string, or you had an 8-bit special value appended to the end of the string. Both ways result in the same space usages.
From what I read in the summary, (didn't read TFA) this whole thing sounds like a propaganda piece supporting the idea that we should use length+string, by presenting it as "this should have been a no-brainer but the idiots making C screwed up."
As a nitpicky pedantic note though, if C had gone with length+string format, then other languages would have been written around the C standard, since most of them were written around the C standards to begin with to increase interoperability in the first place.
To be even more specific: you were given examples of Austrian theory demonstrably working, even when the pet theories of the people making these accusations against it equally demonstrably did not work.
And yet you deny real world examples and continue with the rhetoric instead. For what reason?
Because I'm not arguing that point, I'm arguing about if it is FALSIFIABLE. That means, fuck all evidence that it's "true", because it only matters how it could possibly be false. And you're dodging this entire question. What would prove Austrian school false?!
Austrian economists reject empirical statistical methods, natural experiments and constructed experiments as tools applicable to economics, saying that while it is appropriate in the natural sciences where factors can be isolated in laboratory conditions, the actions of human beings are too complex for this "numerical" treatment as passive non-adaptive subjects. Instead one should isolate the logical processes of human action. Von Mises called this discipline "praxeology" – a term he adapted from Alfred Espinas (but which had been in use by others).[44]
The Austrian praxeological method is based on the heavy use of logical deduction from what they assert to be undeniable, self-evident axioms or irrefutable facts about human existence. The primary axiom from which Austrian economists deduce further certain conclusions is the action axiom, which holds that humans take conscious action toward chosen goals.[45] Austrian economists focus on goal-directed action and say that it is undeniable because in order to deny action, one would have to employ action in the act of denial.
Methodology is the one area where Austrian economists differ most significantly from other schools of economic thought. Mainstream schools such as the neoclassical economists, the Chicago school of economics, the Keynesians and New Keynesians, adopt empirical, mathematical and statistical methods, and focus on induction to construct and test theories—while Austrian economists reject this approach in favor of deduction and logically deduced inferences. According to Austrian economists, deduction is preferred, since if performed correctly, it leads to certain conclusions and inferences that must be true if the underlying assumptions are accurate. However Austrian economist Robert Murphy has stated that those using Austrian theories can still err in their interpretations of history, even if based on a theory formulated by deduction.[46] Caplan makes a similar point about quantitative significance, explaining that a theory, such as one which logically relates minimum wage and unemployment, tells nothing of the approximate quantity of change in unemployment one can expect upon minimum wage increases.
Either you're not practicing Austrian economics or I am right that they reject experimentation, rigorous models, and quantitative methodologies.
I'm not talking about proving that homeopathy works. You can't PROVE that a model works, because all models are flawed. What you can do is show how a model is falsifiable, which is by presenting some form of evidence that YOU as a proponent would need in order to stop believing in that theory.
I don't give a shit if you think that I would dismiss it as irrational, I want to know what evidence could possibly exist that would prove to you that Austrian school is false.
Of course it's vague, that's how people use it, and you know this. Otherwise, you wouldn't have been harping about the "historical meaning of 'capitalism'". Is there anything you can actually debate that doesn't involve mangling semantics?
Why does your brain have to "pop" before you can grasp very simple -- and long acknowledged -- rules of evidence? This is what is a mystery to me.
If something is testable, then it is NOT unfalsifiable. This is simple logic.
If a theory's ability to predict is tested -- and it turns out that it does in fact work -- then it is not unfalsifiable. Why are you having such trouble with this simple concept? You have been giving me peoples' OPINIONS, and I have been giving you verifiable facts. YOU weigh these things, and have decided that the opinions are worth more than demonstrated fact.
And yet *I* am supposed to be the stupid or weird one here? I'm not the one breaking all rules of logic, you are.
Nostradamus's prophesies are testable. Creationism is testable. However, both are not FALSIFIABLE, because even if you prove one part of them wrong, apologists can concoct a new set of rules that explain why the test failed.
"Creationism says the Earth is 6,000~10,000 years old!" We show that the world is significantly older, by many orders of magnitude. Response? "It was just designed to look that way!"
Because Austrian school doesn't use quantitative and mathematically models, we cannot actually lay out what the proper claims are and prove the theory itself false beyond all doubt. All of the methodology behind praxeology is to present things with dialectics and words and argument, rather than true rigorous models.
Let me ask you this. You declare that Austrian school is falsifiable. In order to be falsifiable, there must be evidence that would make you reject the theory and seek an alternative. What evidence can you purpose that would completely falsify Austrian school and make you reject it. (Note, the evidence need not exist, it must merely be possible. For instance, evolution could be shown to be false if fossil records showed no species-level variation of fossils, as well as by a randomized distribution of all fossils in the ground, rather than being layered according to the age of the medium. Neither piece of evidence actually exists, but if it were to exist, then it would disprove evolution. More drastically, we could falsify the existence of time in physics if everything stood still. Completely fanciful, but it is a logically possible condition.)
I await your answer to what kind of evidence could possibly disprove Austrian school.
No, CREDIBLE objection. You keep forgetting that part.
The sound of the goalposts scraping over the ground as you redefine "credible" is already jarring enough. I suspect that no matter what criticism I present to you, you're just going to declare it to be non-credible, and if I ask for "why?" You'll give me an answer, and once I produce criticism conforming to that "why", you will then just call it again non-credible.
Your theory is based on praxeology, and dialectics not rigorously defined quantitative models. You may as well be trying to sell me on homeopathy...
Then why does it say that in all the dictionaries and encyclopedias? Far be it from me to "appeal to authority" in the manner you have kept doing, but in a case like this "authority" is the only way to settle the matter. And the authorities disagree with you.
Oxford English Dictionary: "an economic and political system in which a country's trade and industry are controlled by private owners for profit, rather than by the state."
So, a country's trade... like I don't know... the stock market trade.
No, what he is saying is geometry is split between polygons and bump maps. Polygons provide absolute positing information in 3D space, and bump maps provide small scale texture information. Thus individual grains of sand on a piece of sandpaper could each create separate shadows even if the sandpapers position is represented by a single polygon.
No... the individual grains of sand done through a bump map do not cast a separate shadow. They reflect light as if they were not pointed towards the default normal of the surface that they are on. Sometimes that means they're darker than they would be on a purely flat texture, and sometimes they're a bit brighter than they would be on a purely flat texture. They are not however actually separate heights from the surface and all of them remain exactly in the same planar surface, as a result they cannot cast shadows on themselves anymore than a flat triangle can cast shadows on themselves.
That is what the bump map is for.. it contains the additional depth data.
A bump map does NOT contain additional depth data. It is a NORMAL map. Normally a triangle has only one normal (the vector projecting directly out from the surface) and so you get a flat triangle. However, if you keep an additional texture of vectors at each pixel in the triangle declaring what the deviation from the standard normal is at that point, then you can have a flat triangle that reflects light as if it had tiny facets that point in different directions. This yields the appearance of depth, say of like a grout line in a tile floor, but even if the normal reflects light properly, the point still sits level with the rest of the tile in geometry.
Thus there is no "additional depth data", there is an optical illusion that our minds generally are willing to accept as additional depth data, even when it acts geometrically contrary to what we think we see.
To be fair, not everyone here is an electrical engineer, and some may need a refresher of what a modulator is. The "language of the computer" part is stupid, but it is a direct quote, so I think we can excuse it.
Who cares if it's a direct quote, it's retarded to put it on slashdot. Would you post a summary that has a quote: "and Linux is a unix-type operating system, while Windows isn't". Direct quote or not, it's stupid to include.
I assumed he meant the company releases the changes the ex-employee made after the fork. Then the company could be sued.
This requires them to knowingly violate copyright law. Why don't we just assume that they hire someone to kill the ex-employee, and then we can just say that they go to jail for life? These assumptions are big assumptions to your original statement, and were not well laid out. The juxtaposition of simply moving from "once it's released, then sue!" implies that simply by open sourcing the fork, you can then go sue them for violating copyright law. Just adding a simple "once it's released, if they integrate your changes, then sue!" would have been enough to remain pedantically correct.
(This concludes the nitpicky bitch mode. If you are operating as a sensitive individual who can't listen to annoyingly pedantic criticism, you may now return to reading this comment.)
"Ah... so now we redefine "credible" to get out of the moral quandary of you being either a deceitful liar, or an ignorant moron."
Look, fool. One last time: "appeal to authority" does not trump verifiable facts.
You tried to pull the "appeal to authority" thing on ME before, yet don't recognize when you are doing the same.
I really shouldn't be answering you at all, because as I stated before, this is obviously a waste of time.
It is not appeal to authority to pull up text of another person's objections to a theory when you claim that I am the only person making such an objection.
If you say that I am the only person who says that the sun is purple, but I pull out an account from George Hisphaphaphaph and note that he says that the sun is purple, then obviously I am not the only person claiming that the sun is purple, and thus your claim that I am the only person making such a claim is false.
I don't get how you can't understand this. I'm not using these other sources as proof that Austrian Economics is unfalsifiable, I'm using these other sources as proof that I'm not the only one making that claim. Namely, their assertions could be as false as my example above that the sun is purple, because the only reason I'm quoting what they've said is to show that someone else said what I said. In such a case, quoting them shows first-order evidence that my claims are being made by others.
None of this specific line was about if Austrian Economics was falsifiable or not... rather the entire debate shifted to be exclusively about if I were the only person making such a claim. Showing you that other people are making this claim is thus a perfectly acceptable presentation of evidence, and not an appeal to authority.
You are REALLY bad at debate and logic, you know... you're so hyped up about defending your pet theory, that you've completely lost track of what I'm actually arguing about...
No, they don't use the raw point cloud models, they perform some processing first. And I doubt these guys use unprocessed point clouds for their engine either, so it's a ludicrous claim.
Their specific claim is that they are using point clouds. My thoughts are that if you strategically collapse points of the model (like a dynamic LOD sort of thing) that you could feasibly accomplish something similar to what they're doing.
I mean, seriously, they're talking about large point fields for grains of sand... if that were true, then why wouldn't they be able to use a raw point clouds from scanned objects?
5:45 - he makes the claim that real-world scanned objects can't be used in games because the resolution is too high. This is completely false. Game developers have scanned objects for a long time, and even more often, made extremely high resolution models on purpose. The models are then lowered in resolution down to a usable form, and the differences between the low-res and high-res models is compiled into a normal (bump) map. This is how almost all first person game textures are made these days. (The benefits of this process are mainly surrounding the better efficiency of textures in holding the depth data than polys, especially at varying distances where complex geometry results in extreme aliasing, and the fact that high-poly models cause serious issues with more advanced lighting schemes.) To make the claim this guy just did is highly suspect.
So, what you're saying is that people scan real world objects, but don't actually use those models in games... so... once one accounts for market speak "you can't use a scan of a real-world object in a game [without dropping enough detail so that you're not using the original scan]."
Slashdot Mirror
Getting a free strlen() is NOT an advantage, by the way. In fact, that became a liability when UTF-8 arrived. With a library strlen() function, all you had to do was update the library, but when the compiler was hardcoded to just return the byte count, that wasn't an option. Sure, one could go to UTF-16 instead, but then there's a lot of wasted space.
You've made an error here. There are some Unicode codepoints that are 32-bits long in UTF-16 (surrogate pairs). So, no. Not even with UTF-16 can you get a hardcoded version. However, if you go to UTF-32, then you will have enough space to ensure that every codepoint is represented in the same size value, and thus it's a simple matter of (buffersize in bytes / 4) = number of codepoints. (Note: still not the total number of "characters" as some are combining... and thus an "a" + "combining accent" is one character)
"But who would waste all that space just to make calculations faster?" Actually, Perl represents all Unicode in UTF-32 (native byte-alignment) and renders out to UTF-8 when printed. Precisely because space is relatively cheap, and processing time is still often the current limiting factor.
I'm correcting myself here... apparently they weren't considering going with a 255-byte limit, but a 65535-byte limit, which would have increased the size overhead by one.
Which is worse? Having it be O(N) to get a string length and having inexperienced programmers get confused and make mistakes? Or capping your maximum string length at 0xffff?
I'll take the former, please. I do a lot of string manipulation in C and when you're used to it, it's actually not that bad to get right and still be efficient. And it provides a useful shibboleth to detect people who are no good at C. :-) Just think of how much harder it would be to interview a C programmer if you couldn't give them a crazy string manipulation problem.
I don't know why you have to bring sibolezes into this...
Compatible with what? Seems to me they could have just used continuation bit for the size field, much the way UTF-8 works to store non-ASCII characters.
This would still make the strings incompatible, because you would only have a 127-byte string length before the "continuation bit" comes into play and you need to switch to a 15-bit string length. All the previous code written with longer-than-127-byte strings would be incompatible.
Not to mention the argument for "because space was at a premium" is specious, because either you had a 8-bit length prepended to the string, or you had an 8-bit special value appended to the end of the string. Both ways result in the same space usages.
From what I read in the summary, (didn't read TFA) this whole thing sounds like a propaganda piece supporting the idea that we should use length+string, by presenting it as "this should have been a no-brainer but the idiots making C screwed up."
As a nitpicky pedantic note though, if C had gone with length+string format, then other languages would have been written around the C standard, since most of them were written around the C standards to begin with to increase interoperability in the first place.
To be even more specific: you were given examples of Austrian theory demonstrably working, even when the pet theories of the people making these accusations against it equally demonstrably did not work.
And yet you deny real world examples and continue with the rhetoric instead. For what reason?
Because I'm not arguing that point, I'm arguing about if it is FALSIFIABLE. That means, fuck all evidence that it's "true", because it only matters how it could possibly be false. And you're dodging this entire question. What would prove Austrian school false?!
Austrian economists reject empirical statistical methods, natural experiments and constructed experiments as tools applicable to economics, saying that while it is appropriate in the natural sciences where factors can be isolated in laboratory conditions, the actions of human beings are too complex for this "numerical" treatment as passive non-adaptive subjects. Instead one should isolate the logical processes of human action. Von Mises called this discipline "praxeology" – a term he adapted from Alfred Espinas (but which had been in use by others).[44]
The Austrian praxeological method is based on the heavy use of logical deduction from what they assert to be undeniable, self-evident axioms or irrefutable facts about human existence. The primary axiom from which Austrian economists deduce further certain conclusions is the action axiom, which holds that humans take conscious action toward chosen goals.[45] Austrian economists focus on goal-directed action and say that it is undeniable because in order to deny action, one would have to employ action in the act of denial.
Methodology is the one area where Austrian economists differ most significantly from other schools of economic thought. Mainstream schools such as the neoclassical economists, the Chicago school of economics, the Keynesians and New Keynesians, adopt empirical, mathematical and statistical methods, and focus on induction to construct and test theories—while Austrian economists reject this approach in favor of deduction and logically deduced inferences. According to Austrian economists, deduction is preferred, since if performed correctly, it leads to certain conclusions and inferences that must be true if the underlying assumptions are accurate. However Austrian economist Robert Murphy has stated that those using Austrian theories can still err in their interpretations of history, even if based on a theory formulated by deduction.[46] Caplan makes a similar point about quantitative significance, explaining that a theory, such as one which logically relates minimum wage and unemployment, tells nothing of the approximate quantity of change in unemployment one can expect upon minimum wage increases.
Either you're not practicing Austrian economics or I am right that they reject experimentation, rigorous models, and quantitative methodologies.
I'm not talking about proving that homeopathy works. You can't PROVE that a model works, because all models are flawed. What you can do is show how a model is falsifiable, which is by presenting some form of evidence that YOU as a proponent would need in order to stop believing in that theory.
I don't give a shit if you think that I would dismiss it as irrational, I want to know what evidence could possibly exist that would prove to you that Austrian school is false.
Of course it's vague, that's how people use it, and you know this. Otherwise, you wouldn't have been harping about the "historical meaning of 'capitalism'". Is there anything you can actually debate that doesn't involve mangling semantics?
Why does your brain have to "pop" before you can grasp very simple -- and long acknowledged -- rules of evidence? This is what is a mystery to me.
If something is testable, then it is NOT unfalsifiable. This is simple logic.
If a theory's ability to predict is tested -- and it turns out that it does in fact work -- then it is not unfalsifiable. Why are you having such trouble with this simple concept? You have been giving me peoples' OPINIONS, and I have been giving you verifiable facts. YOU weigh these things, and have decided that the opinions are worth more than demonstrated fact.
And yet * I * am supposed to be the stupid or weird one here? I'm not the one breaking all rules of logic, you are.
Nostradamus's prophesies are testable. Creationism is testable. However, both are not FALSIFIABLE, because even if you prove one part of them wrong, apologists can concoct a new set of rules that explain why the test failed.
"Creationism says the Earth is 6,000~10,000 years old!" We show that the world is significantly older, by many orders of magnitude. Response? "It was just designed to look that way!"
Because Austrian school doesn't use quantitative and mathematically models, we cannot actually lay out what the proper claims are and prove the theory itself false beyond all doubt. All of the methodology behind praxeology is to present things with dialectics and words and argument, rather than true rigorous models.
Let me ask you this. You declare that Austrian school is falsifiable. In order to be falsifiable, there must be evidence that would make you reject the theory and seek an alternative. What evidence can you purpose that would completely falsify Austrian school and make you reject it. (Note, the evidence need not exist, it must merely be possible. For instance, evolution could be shown to be false if fossil records showed no species-level variation of fossils, as well as by a randomized distribution of all fossils in the ground, rather than being layered according to the age of the medium. Neither piece of evidence actually exists, but if it were to exist, then it would disprove evolution. More drastically, we could falsify the existence of time in physics if everything stood still. Completely fanciful, but it is a logically possible condition.)
I await your answer to what kind of evidence could possibly disprove Austrian school.
No, CREDIBLE objection. You keep forgetting that part.
The sound of the goalposts scraping over the ground as you redefine "credible" is already jarring enough. I suspect that no matter what criticism I present to you, you're just going to declare it to be non-credible, and if I ask for "why?" You'll give me an answer, and once I produce criticism conforming to that "why", you will then just call it again non-credible.
Your theory is based on praxeology, and dialectics not rigorously defined quantitative models. You may as well be trying to sell me on homeopathy...
Then why does it say that in all the dictionaries and encyclopedias? Far be it from me to "appeal to authority" in the manner you have kept doing, but in a case like this "authority" is the only way to settle the matter. And the authorities disagree with you.
Oxford English Dictionary: "an economic and political system in which a country's trade and industry are controlled by private owners for profit, rather than by the state."
So, a country's trade... like I don't know... the stock market trade.
No, what he is saying is geometry is split between polygons and bump maps. Polygons provide absolute positing information in 3D space, and bump maps provide small scale texture information. Thus individual grains of sand on a piece of sandpaper could each create separate shadows even if the sandpapers position is represented by a single polygon.
No... the individual grains of sand done through a bump map do not cast a separate shadow. They reflect light as if they were not pointed towards the default normal of the surface that they are on. Sometimes that means they're darker than they would be on a purely flat texture, and sometimes they're a bit brighter than they would be on a purely flat texture. They are not however actually separate heights from the surface and all of them remain exactly in the same planar surface, as a result they cannot cast shadows on themselves anymore than a flat triangle can cast shadows on themselves.
That is what the bump map is for.. it contains the additional depth data.
A bump map does NOT contain additional depth data. It is a NORMAL map. Normally a triangle has only one normal (the vector projecting directly out from the surface) and so you get a flat triangle. However, if you keep an additional texture of vectors at each pixel in the triangle declaring what the deviation from the standard normal is at that point, then you can have a flat triangle that reflects light as if it had tiny facets that point in different directions. This yields the appearance of depth, say of like a grout line in a tile floor, but even if the normal reflects light properly, the point still sits level with the rest of the tile in geometry.
Thus there is no "additional depth data", there is an optical illusion that our minds generally are willing to accept as additional depth data, even when it acts geometrically contrary to what we think we see.
To be fair, not everyone here is an electrical engineer, and some may need a refresher of what a modulator is. The "language of the computer" part is stupid, but it is a direct quote, so I think we can excuse it.
Who cares if it's a direct quote, it's retarded to put it on slashdot. Would you post a summary that has a quote: "and Linux is a unix-type operating system, while Windows isn't". Direct quote or not, it's stupid to include.
Seriously, does anyone here on Slashdot need their summary dumbed down that far?
I assumed he meant the company releases the changes the ex-employee made after the fork. Then the company could be sued.
This requires them to knowingly violate copyright law. Why don't we just assume that they hire someone to kill the ex-employee, and then we can just say that they go to jail for life? These assumptions are big assumptions to your original statement, and were not well laid out. The juxtaposition of simply moving from "once it's released, then sue!" implies that simply by open sourcing the fork, you can then go sue them for violating copyright law. Just adding a simple "once it's released, if they integrate your changes, then sue!" would have been enough to remain pedantically correct.
(This concludes the nitpicky bitch mode. If you are operating as a sensitive individual who can't listen to annoyingly pedantic criticism, you may now return to reading this comment.)
urp.
Replace "fractal" with "shit tons of data", and yes, you pretty much have what I suspect that they're doing.
"Ah... so now we redefine "credible" to get out of the moral quandary of you being either a deceitful liar, or an ignorant moron."
Look, fool. One last time: "appeal to authority" does not trump verifiable facts.
You tried to pull the "appeal to authority" thing on ME before, yet don't recognize when you are doing the same.
I really shouldn't be answering you at all, because as I stated before, this is obviously a waste of time.
It is not appeal to authority to pull up text of another person's objections to a theory when you claim that I am the only person making such an objection.
If you say that I am the only person who says that the sun is purple, but I pull out an account from George Hisphaphaphaph and note that he says that the sun is purple, then obviously I am not the only person claiming that the sun is purple, and thus your claim that I am the only person making such a claim is false.
I don't get how you can't understand this. I'm not using these other sources as proof that Austrian Economics is unfalsifiable, I'm using these other sources as proof that I'm not the only one making that claim. Namely, their assertions could be as false as my example above that the sun is purple, because the only reason I'm quoting what they've said is to show that someone else said what I said. In such a case, quoting them shows first-order evidence that my claims are being made by others.
None of this specific line was about if Austrian Economics was falsifiable or not... rather the entire debate shifted to be exclusively about if I were the only person making such a claim. Showing you that other people are making this claim is thus a perfectly acceptable presentation of evidence, and not an appeal to authority.
You are REALLY bad at debate and logic, you know... you're so hyped up about defending your pet theory, that you've completely lost track of what I'm actually arguing about...
"I've simply objected to your definition being a good definition."
Well, it has only been pretty much exactly the same definition for about 300 years. You go ahead and argue with it, if it makes you feel better.
Funny that your definition is no longer in use by pretty much anyone, except in strict academic speech registers...
Ok, you just made my brain pop from your stupid... I'm leaving with my sanity intact...
No, they don't use the raw point cloud models, they perform some processing first. And I doubt these guys use unprocessed point clouds for their engine either, so it's a ludicrous claim.
Their specific claim is that they are using point clouds. My thoughts are that if you strategically collapse points of the model (like a dynamic LOD sort of thing) that you could feasibly accomplish something similar to what they're doing.
I mean, seriously, they're talking about large point fields for grains of sand... if that were true, then why wouldn't they be able to use a raw point clouds from scanned objects?
The idea you've had is the standard traversal of an octree. Congratulations on reinventing a standard tool in CGI. :-)
Yet another person falls to Saunt Lora's Proposition.
5:45 - he makes the claim that real-world scanned objects can't be used in games because the resolution is too high. This is completely false. Game developers have scanned objects for a long time, and even more often, made extremely high resolution models on purpose. The models are then lowered in resolution down to a usable form, and the differences between the low-res and high-res models is compiled into a normal (bump) map. This is how almost all first person game textures are made these days. (The benefits of this process are mainly surrounding the better efficiency of textures in holding the depth data than polys, especially at varying distances where complex geometry results in extreme aliasing, and the fact that high-poly models cause serious issues with more advanced lighting schemes.) To make the claim this guy just did is highly suspect.
So, what you're saying is that people scan real world objects, but don't actually use those models in games... so... once one accounts for market speak "you can't use a scan of a real-world object in a game [without dropping enough detail so that you're not using the original scan]."
Why is that easier?
Smaller numbers, less intermediate calculations.