Sorry to reply here, but the discussion we where having here:
http://ask.slashdot.org/comments.pl?sid=113251&c id =9594082
doesn't allow comments to be posted anymore.
Anyway, in answer to the question you asked right at the end, the answer is of course yes. It's called CNC (computer numeric control) and basically you program the millhead's (or lathe, or rapid prototyping unit's) movements, insert the material and there you go: tolerances up to and including ten thousanths of a centimeter, every time (or at least as long as the toolhead remains pristine).
And yes, sometimes you really need that kind of accuracy.
This isn't a dig at you, but the one reason I can't replace my PDA with a phone is data entry. Because I have quite a few appointments and a lot made (well) in advance, I have to specify them with qualifiers (14.15 meeting room 2.15 isn't enough). It just takes too long when you have to press buttons twice, thrice or even four times for a single letter (and T9 doesn't work for that kind of thing).
And that's why I need and probably won't replace a PDA. Plus, try reading books on a cellphone:) Or doing some quick matrix calculations with graphed solutions:(
You know, I can very much understand that point of view. But I think that there is something to be said for the sheer culture and technology of a good timepiece. It represents the hight, in some ways, of the Industrial Revolution, just like a well crafted lighter represents man's conquest of fire.
That's purely an emotional view, of course, but then again I am a tech fetishist I guess. Then again, I bet most of the people who visit/. are too.
I'm prez. of the student management for my flat, which means I get all the keys. So on sunny days I every so often sit on top of the building (20 stories high), looking over the city. My room is on the top floor, so I sometimes look around the net on my Tungsten 3 connected to my PC via a USB bluetooth dongle slung ou the window:)
Two things: nothing ever has 100% accuracy in physics...there is always at the very least observer error. And what you say about higher order shapes could well be correct, but is alreasy accounted for by the Pauli principle: no two particles can ever have the exact same position and energy levels. Two objects passing through each other might look as if they're occupying the same space...but the particles involved will never have the same energy function and position. Not that you coulod actually tell, thanks to Heisenberg, but hey, c'est la vie.
He's right though: you can teach physics using maple, mathcad or whatever, and you would get physicists who could function correctly most of the time.
But actually doing and knowing the math serves two purposes: you get a better understanding of how stuff works, and what effects certain things will have on others, plus you can troubleshoot equations/algorithms/theories much more efficiently.
From what I understood, it is he himself who tied together the final theory of why he's wrong; that's also the reason why he's the one making the presentation. Sure, it builds on previous work, but he came up with his own undoing, I thought.
Anyway, needing enough and correct evidence before taking up a theory strikes me as the correct scientific way.
The speed difference in getting to the info you're looking for is quite staggering. I had this exact same discussion with a friend of mine at work, so we set up a task list we had to run through, timed; call it a speed trial (yeah, we're geeks...guess what the byline of this website is.). And to make sure that there wasn't an advantage for one person due to quicker reflexes or whatever, after that runthrough we switched devices, and after familiarising ourselves with 'em, did the runthrough on each other's device.
PalmOS won both times. Duh.
Also, I have to ask: does PocketPC still not really close applications and get bogged down when you've opened (and closed) a couple of apps in a row? Do you think that's normal?
See, the main difference can be found in your post: you only talk hardware. And it is true...PocketPC hardware has been a bit ahead of PalmOS hardware in the past...but so what? If you actually use a PDA, you'd know the only thing that makes it usefull (as oposed to being just a toy) is the software...and PalmOS is so much more usefull.
And finally, have a look at the number of PocketPC's for second hand sale on internet compared to the number of Treo's. Don't there seem to be many more PocketPC's up for sale? I think that's because those with Treo's like 'em so much they dont wanna sell 'em:)
Not really true. Especially the kind of trivia on these shows: you might know enough to be able to say that e=mc^2 should actually be e=m*gamma(u)*c^2, but that doesn't neccessarily mean you can do lorentztransformations or calculate schrodinger functions.
Intelligence means being able to use a collection of facts to do something. Knowing a lot of facts just means you know a lot of facts.
So? Turning around an aircraft carrier in midocean so the Pres. can say "the war is over and all hostilities are at an end" costs a lot of money too, and is used for election purposes aswell.
Dude, go to university, study the subject and then say you don't beleive in the HIV/AIDS link. You won't: the link is proven, just like the greenhouse effect. The only questions left (in both cases) is what the mechanism is exactly...not if there is a mechanism.
Lemme guess: you work in production. Once you go into large scale design or research and development there just is no way you can do your job without a computer. Unless you're a secretary or a manager, of course, where pen and paper is enough. Engineering has advanced, and it's not just at the cutting edge. Anyone who dares to say that he can do his work without computational crunching power just hasn't done either complex enough work or large scale enough work. That simple. So even though I was still trained in paper drafting (last class to have been that), the work I did and what I do now mean that no: I cannot work effectively without a computer. But then, no-one there and here can: well, theoretically we could, it'd just take centuries to finish the work we have to do.
Point is: most above-median-level engineering jobs require computational skills and can't work without computers. Now if you're a welder or a pipe fitter, sure, you can function without them. Design a particle accelerator, an efficient engine or a radar system and you can't.
Finite element analysis was only first proposed in 1943 and it wasn't until 1956 that something was done with that proposition. FEA is one of those things which can only be done using computers (or many, many man-years of expensive engineering time).
I might be a young grasshopper, but I am also a mechanical engineer who knows what tasks are so computationally intensive that anyone who considers doing them by hand can't possibly call him- or herself an engineer. And it's not just about decimal points (although they do help in certain calculations)...it's about volume of work.
Now, sure, if all you need to design is a wrench, then you can do that by hand...but if you want to overengineer for example a soda can so that you can shave off 0.005 cents per can for material costs (which gets multiplied by a couple of million over the production run), or maybe simulate waves crashing onto the pilons of your oilrig, I'm very sorry, but you need computers.
Now go get a sliderule, young grasshopper, and tell me when you're sick of using that thing to simulate the flow in a combustion chamber. Maybe then I'll let you have go on that shiny box called a computer
But you're effectively saying (together with all the rest off the 'praise NASA, without them this wouldn't happen' crowd) is: "praise the dead horse which has become rotten for something it did decades ago and which has since bloated and died!"
You've never had to make LODs for a game model, have you? An efficient algorithm that does this well (ie without producing odd looking things) will save time and money.
As Zorba says: poly reduction is especially usefull for LOD. A model right at the range of rendering sight shouldn't have the full amount of polys. But to do that by hand is very time consuming (especially if you want a full set of 5 or so LODs), so a program which can do that automagically and do it well (which is a mayor problem with the functions found in 3dsmax and Maya; simply put, they stink) is quite a time saver. And thus a money saver...and thus Something Good(TM).
"Lastly, quake and others generated meshes from smooth NURBS objects."
If by quake you're refering to the computer game, you're very, very wrong. In no way do they employ NURBS. Nor does any computergame (well, no succesfull mainstream one I know of) due to the fact that 3d accelerators only support triangles (well, the TNT did, but no-one used it...one might even make the case that Carmack killed off the use of parametric curves in 3d hardware acceleration).
As for high poly models, you are right...or at least, were. Zbrush 2 has made it very easy for artists to build highpoly models from scratch. Thanks to that piece of software, I can see scanning to get a base mesh (or pointcloud) diminishing in use by a significant factor.
I'd say the easiest thing to do is use your job desciption and ask Demron for a sample patch of the material for field testing. They'll be happy to comply, and you'll have prooof either way of their claim.
As for Demron not bothering to answer your request: duh. He's already written many scientific papers on it, and he's probably bound by NDA to his company. You really think he's gonna tell you how it works? Hell, unless you're a physicist who's phd'd in that area, I doubt you'd understand the explanation he'd give you:)
Slashdot Mirror
Sorry to reply here, but the discussion we where having here:
c id =9594082
http://ask.slashdot.org/comments.pl?sid=113251&
doesn't allow comments to be posted anymore.
Anyway, in answer to the question you asked right at the end, the answer is of course yes. It's called CNC (computer numeric control) and basically you program the millhead's (or lathe, or rapid prototyping unit's) movements, insert the material and there you go: tolerances up to and including ten thousanths of a centimeter, every time (or at least as long as the toolhead remains pristine).
And yes, sometimes you really need that kind of accuracy.
D'ya have a nice sleep after?
Talk about an expensive sundial!
Oh, come on! Be consequent and use 006 for the remainder :)
This isn't a dig at you, but the one reason I can't replace my PDA with a phone is data entry. Because I have quite a few appointments and a lot made (well) in advance, I have to specify them with qualifiers (14.15 meeting room 2.15 isn't enough). It just takes too long when you have to press buttons twice, thrice or even four times for a single letter (and T9 doesn't work for that kind of thing).
:) Or doing some quick matrix calculations with graphed solutions :(
And that's why I need and probably won't replace a PDA. Plus, try reading books on a cellphone
You know, I can very much understand that point of view. But I think that there is something to be said for the sheer culture and technology of a good timepiece. It represents the hight, in some ways, of the Industrial Revolution, just like a well crafted lighter represents man's conquest of fire.
/. are too.
That's purely an emotional view, of course, but then again I am a tech fetishist I guess. Then again, I bet most of the people who visit
I'm prez. of the student management for my flat, which means I get all the keys. So on sunny days I every so often sit on top of the building (20 stories high), looking over the city. My room is on the top floor, so I sometimes look around the net on my Tungsten 3 connected to my PC via a USB bluetooth dongle slung ou the window :)
/I/ think it's pretty neat :)
Hey,
What about the baby Bell's?
Two things: nothing ever has 100% accuracy in physics...there is always at the very least observer error. And what you say about higher order shapes could well be correct, but is alreasy accounted for by the Pauli principle: no two particles can ever have the exact same position and energy levels. Two objects passing through each other might look as if they're occupying the same space...but the particles involved will never have the same energy function and position. Not that you coulod actually tell, thanks to Heisenberg, but hey, c'est la vie.
He's right though: you can teach physics using maple, mathcad or whatever, and you would get physicists who could function correctly most of the time.
But actually doing and knowing the math serves two purposes: you get a better understanding of how stuff works, and what effects certain things will have on others, plus you can troubleshoot equations/algorithms/theories much more efficiently.
From what I understood, it is he himself who tied together the final theory of why he's wrong; that's also the reason why he's the one making the presentation. Sure, it builds on previous work, but he came up with his own undoing, I thought.
Anyway, needing enough and correct evidence before taking up a theory strikes me as the correct scientific way.
Maybe because it runs on PalmOS?
:)
The speed difference in getting to the info you're looking for is quite staggering. I had this exact same discussion with a friend of mine at work, so we set up a task list we had to run through, timed; call it a speed trial (yeah, we're geeks...guess what the byline of this website is.). And to make sure that there wasn't an advantage for one person due to quicker reflexes or whatever, after that runthrough we switched devices, and after familiarising ourselves with 'em, did the runthrough on each other's device.
PalmOS won both times. Duh.
Also, I have to ask: does PocketPC still not really close applications and get bogged down when you've opened (and closed) a couple of apps in a row? Do you think that's normal?
See, the main difference can be found in your post: you only talk hardware. And it is true...PocketPC hardware has been a bit ahead of PalmOS hardware in the past...but so what? If you actually use a PDA, you'd know the only thing that makes it usefull (as oposed to being just a toy) is the software...and PalmOS is so much more usefull.
And finally, have a look at the number of PocketPC's for second hand sale on internet compared to the number of Treo's. Don't there seem to be many more PocketPC's up for sale? I think that's because those with Treo's like 'em so much they dont wanna sell 'em
Not really true. Especially the kind of trivia on these shows: you might know enough to be able to say that e=mc^2 should actually be e=m*gamma(u)*c^2, but that doesn't neccessarily mean you can do lorentztransformations or calculate schrodinger functions.
Intelligence means being able to use a collection of facts to do something. Knowing a lot of facts just means you know a lot of facts.
For the lazy people amongst you:
you want information?
So? Turning around an aircraft carrier in midocean so the Pres. can say "the war is over and all hostilities are at an end" costs a lot of money too, and is used for election purposes aswell.
Dude, go to university, study the subject and then say you don't beleive in the HIV/AIDS link. You won't: the link is proven, just like the greenhouse effect. The only questions left (in both cases) is what the mechanism is exactly...not if there is a mechanism.
Lemme guess: you work in production. Once you go into large scale design or research and development there just is no way you can do your job without a computer. Unless you're a secretary or a manager, of course, where pen and paper is enough.
Engineering has advanced, and it's not just at the cutting edge. Anyone who dares to say that he can do his work without computational crunching power just hasn't done either complex enough work or large scale enough work. That simple.
So even though I was still trained in paper drafting (last class to have been that), the work I did and what I do now mean that no: I cannot work effectively without a computer. But then, no-one there and here can: well, theoretically we could, it'd just take centuries to finish the work we have to do.
Point is: most above-median-level engineering jobs require computational skills and can't work without computers. Now if you're a welder or a pipe fitter, sure, you can function without them. Design a particle accelerator, an efficient engine or a radar system and you can't.
Finite element analysis was only first proposed in 1943 and it wasn't until 1956 that something was done with that proposition. FEA is one of those things which can only be done using computers (or many, many man-years of expensive engineering time).
I might be a young grasshopper, but I am also a mechanical engineer who knows what tasks are so computationally intensive that anyone who considers doing them by hand can't possibly call him- or herself an engineer. And it's not just about decimal points (although they do help in certain calculations)...it's about volume of work.
Now, sure, if all you need to design is a wrench, then you can do that by hand...but if you want to overengineer for example a soda can so that you can shave off 0.005 cents per can for material costs (which gets multiplied by a couple of million over the production run), or maybe simulate waves crashing onto the pilons of your oilrig, I'm very sorry, but you need computers.
Now go get a sliderule, young grasshopper, and tell me when you're sick of using that thing to simulate the flow in a combustion chamber. Maybe then I'll let you have go on that shiny box called a computer
What, you do Finite Element Analysis and/or complex fluid dynamics simulation by hand? Wow.
But you're effectively saying (together with all the rest off the 'praise NASA, without them this wouldn't happen' crowd) is: "praise the dead horse which has become rotten for something it did decades ago and which has since bloated and died!"
Seems to me the answer is simple on Win machines: keep tht damned 'autorun on insert' function for your cd/dvd-rom off!
I haven't tried this, but I can imagine that this is yet another one of those progs which slips in using that flaw.
You've never had to make LODs for a game model, have you? An efficient algorithm that does this well (ie without producing odd looking things) will save time and money.
As Zorba says: poly reduction is especially usefull for LOD. A model right at the range of rendering sight shouldn't have the full amount of polys. But to do that by hand is very time consuming (especially if you want a full set of 5 or so LODs), so a program which can do that automagically and do it well (which is a mayor problem with the functions found in 3dsmax and Maya; simply put, they stink) is quite a time saver. And thus a money saver...and thus Something Good(TM).
"Lastly, quake and others generated meshes from smooth NURBS objects."
If by quake you're refering to the computer game, you're very, very wrong. In no way do they employ NURBS. Nor does any computergame (well, no succesfull mainstream one I know of) due to the fact that 3d accelerators only support triangles (well, the TNT did, but no-one used it...one might even make the case that Carmack killed off the use of parametric curves in 3d hardware acceleration).
As for high poly models, you are right...or at least, were. Zbrush 2 has made it very easy for artists to build highpoly models from scratch. Thanks to that piece of software, I can see scanning to get a base mesh (or pointcloud) diminishing in use by a significant factor.
I'd say the easiest thing to do is use your job desciption and ask Demron for a sample patch of the material for field testing. They'll be happy to comply, and you'll have prooof either way of their claim.
:)
As for Demron not bothering to answer your request: duh. He's already written many scientific papers on it, and he's probably bound by NDA to his company. You really think he's gonna tell you how it works? Hell, unless you're a physicist who's phd'd in that area, I doubt you'd understand the explanation he'd give you