I only write this because it is likely to be missed.
I agree, but I think it is for different reasons.
on
Bad Day To Be Sony
·
· Score: 1
By buying the CD second hand you are decreasing the supply of cheap second hand disks, therefore making it more likely that someone will pay for a new copy when they can't find a used copy.
What is the distinction between doing that and doing calculus?
Suppose you are teaching a calculus class. One of your students never comes to class except on exam days, never bought the textbook in the first place, and has never been exposed to any calculus instruction whatsoever. However, this student manages to write down more than 90% of the answers correctly on every exam without writing down anything that looks like the process of doing calculus. Would you fail this student for not "knowing" calculus? So long as he can do it, does the way he does it factor into how you would grade him?
By the same approach, when our muscles move according to a very detailed kinematics plan, in what sense can you say that they aren't computing the solutions?
This makes every store in America a brick and mortar storefront for Amazon for anyone with a camera phone. Not to mention everyone's homes and workplaces.
Imagine this: You are at a friends house. They give you a glowing recommendation of a book they just read that is sitting on their coffee table. You take out your camera phone and take a picture of the barcode. A second later, pricing information from Amazon comes up, and you can order it right there with your phone.
This is really a pretty sweet technology.
True, but what we are evaluating here is programming languages, not methods of specification writing. If your programming language is good enough that translating from thorough specification to code is a trivial and nearly error proof process, then in some sense I'd say you have an ideal programming language.
Well if you have an incorrect specification, you are pretty much fubared anyways.
Let me formalize my original statement down to this. Functional programming reduces programming into a language that corresponds exactly to an inductive specification of an algorithm. The correctness of a functional program is as easy to prove as any exact specification of the algorithm used.
True, but with functional code, verifying correctness is no more difficult than verifying a proof of the correctness of your algorithm, because the code itself is structured as a proof.
Even the example code you posted is an induction proof of its own correctness, except it only proved its correctness for positive numbers. You can look at that code the same way you would look at induction with a base case of 0 and an induction case of f(n) dependant on f(n-1).
What SML did for me is made verifying my programs no more difficult than verifying an induction proof, and often considerably easier. The type safety and absence of side effects don't guarantee correctness of course, but they elminate the most common errors, and result in code that you can check as if it were a proof.
I'm being perfectly serious. I had a class that used SML of NJ, and for the last project or so, I didn't need to test it. It typechecked and I could see just by looking at it that it worked. So long as you are using the purely functional components of the language, the code is just an induction proof in slightly different symbols.
From a practicality standpoint, I don't think people are going to be using functional programming to develop the connective tissue of their applications, because it really doesn't make sense to think of a database or a user interface as stateless. However, I'd imagine that using SML for your core business logic, and using other languages to tie everything together and get the data where it needs to go, would be a great and extensible model.
Functional programming is awesome, and I'm thoroughly convinced that it will take over just about everything its feasible for it to take over. There is nothing like the feeling of writing a program, having it type check, and not having to test it because you can look at the code and tell that it proves its own correctness.
I wouldn't be too sure. I told them I've got asthma, which disqualifies me (according to the first recruiter who called) but three more called me over the next year.
If you are developing software in any capacity, personally I think nothing helps creativity more than a clean code base. When the biggest thing you associate with implementing a new interesting feature is the crap you have to go through to get it interacting with everything else, you aren't very likely to come up with good ideas and act on them often.
I take it that you have some significant experience in non-realtime rendering, but it doesn't sound like you have much experience with game engines and how level designers use them.
Precomputed light maps do indeed have to do with the choice of engine, because the engine takes care of computing the lightmaps for you. Halflife 2 for instance, supports normal mapped radiosity calculations, in which the diffuse lighting components are added along different vectors during compilation, and then dotted with the normal map during the rendering. "Level designers" don't store them in textures, the compile tools that are associated with the engine do, and the engine takes care of displaying them appropriately.
Having precomputed lightmaps in the doom 3 engine would break all the internal consistency of the lighting. Mobile lights in engines based on precomputed lighting are treated differently from static lights. Doom 3 doesn't have this distinction.
The doom 3 approach allows lights to be much more dynamic, but when a light is that dynamic, you can't have precomputed light maps. You wouldn't have any way of updating them to reflect changing light conditions. Every time an imp warps in, all the lights dim. This couldn't be done realistically with precomputed light maps.
Adding precomputed light maps would require redoing all the internal assumptions about lights in the engine, and you would be basically writing your own.
My Mom has signed onto it too as part of their trial service. She was previously a zoomtown customer. I've had some issues with it dropping packets, and it seems to be pretty sensitive to where the "modem" (or whatever they call it) is plugged into your wall.
At first we just plugged it in wherever in the house we were using the computer near an outlet, and that worked ok. It was like having a home network with just a little added inconvenience. People apparantly liked the mobility of that, but didn't like dropping packets. They hooked my mom up with a wireless router, and tested all the outlets in the house to find the most stable one to hook it up to. The device is about 3"x4"x2".
I'm pretty sure they have to install something in the power system to enable it for an area. (At the transformer level maybe?)
I'm not sure Cincinnati was such a great choice for where to try this out first. We aren't the most tech savvy of places, and a lot of homes around here are pretty old and probably haven't had their wiring replaced in 50 years. It seems like a pretty adverse environment to be testing it in.
All in all, its pretty cool, although it gives me the willies plugging my ethernet cable directly into the wall socket some times. (http://www.fiftythree.org/etherkiller/) I wonder what happens when the device starts to go bad?
Cincinnati's got Graeter's, which is the best icecream in the world flat out.
Cincinnati's also got a excellent art museum, one of the best zoos in the country, great public libraries, and that wonderful heartwarming Skyline Chili.
Slashdot Mirror
I only write this because it is likely to be missed.
By buying the CD second hand you are decreasing the supply of cheap second hand disks, therefore making it more likely that someone will pay for a new copy when they can't find a used copy.
What is the distinction between doing that and doing calculus?
Suppose you are teaching a calculus class. One of your students never comes to class except on exam days, never bought the textbook in the first place, and has never been exposed to any calculus instruction whatsoever. However, this student manages to write down more than 90% of the answers correctly on every exam without writing down anything that looks like the process of doing calculus. Would you fail this student for not "knowing" calculus? So long as he can do it, does the way he does it factor into how you would grade him?
By the same approach, when our muscles move according to a very detailed kinematics plan, in what sense can you say that they aren't computing the solutions?
I wish I had the points right now. GJ.
Cleartype for one.
It does show that climate can change rapidly on a global scale without the help of man . . . on mars.
This makes every store in America a brick and mortar storefront for Amazon for anyone with a camera phone. Not to mention everyone's homes and workplaces. Imagine this: You are at a friends house. They give you a glowing recommendation of a book they just read that is sitting on their coffee table. You take out your camera phone and take a picture of the barcode. A second later, pricing information from Amazon comes up, and you can order it right there with your phone. This is really a pretty sweet technology.
Perfectly on target assessment I think.
... at least you can sleep at night.
You track them with your mouse. In order for moving the mouse around to look realistic, you need motion blur.
True, but what we are evaluating here is programming languages, not methods of specification writing. If your programming language is good enough that translating from thorough specification to code is a trivial and nearly error proof process, then in some sense I'd say you have an ideal programming language.
Well if you have an incorrect specification, you are pretty much fubared anyways.
Let me formalize my original statement down to this. Functional programming reduces programming into a language that corresponds exactly to an inductive specification of an algorithm. The correctness of a functional program is as easy to prove as any exact specification of the algorithm used.
True, but with functional code, verifying correctness is no more difficult than verifying a proof of the correctness of your algorithm, because the code itself is structured as a proof.
Even the example code you posted is an induction proof of its own correctness, except it only proved its correctness for positive numbers. You can look at that code the same way you would look at induction with a base case of 0 and an induction case of f(n) dependant on f(n-1). What SML did for me is made verifying my programs no more difficult than verifying an induction proof, and often considerably easier. The type safety and absence of side effects don't guarantee correctness of course, but they elminate the most common errors, and result in code that you can check as if it were a proof.
I'm being perfectly serious. I had a class that used SML of NJ, and for the last project or so, I didn't need to test it. It typechecked and I could see just by looking at it that it worked. So long as you are using the purely functional components of the language, the code is just an induction proof in slightly different symbols.
From a practicality standpoint, I don't think people are going to be using functional programming to develop the connective tissue of their applications, because it really doesn't make sense to think of a database or a user interface as stateless. However, I'd imagine that using SML for your core business logic, and using other languages to tie everything together and get the data where it needs to go, would be a great and extensible model.
Functional programming is awesome, and I'm thoroughly convinced that it will take over just about everything its feasible for it to take over. There is nothing like the feeling of writing a program, having it type check, and not having to test it because you can look at the code and tell that it proves its own correctness.
I certainly did. Provided a little chuckle.
I wouldn't be too sure. I told them I've got asthma, which disqualifies me (according to the first recruiter who called) but three more called me over the next year.
That is the most unintelligible argument I have ever seen professionally published.
If you are developing software in any capacity, personally I think nothing helps creativity more than a clean code base. When the biggest thing you associate with implementing a new interesting feature is the crap you have to go through to get it interacting with everything else, you aren't very likely to come up with good ideas and act on them often.
I take it that you have some significant experience in non-realtime rendering, but it doesn't sound like you have much experience with game engines and how level designers use them.
Precomputed light maps do indeed have to do with the choice of engine, because the engine takes care of computing the lightmaps for you. Halflife 2 for instance, supports normal mapped radiosity calculations, in which the diffuse lighting components are added along different vectors during compilation, and then dotted with the normal map during the rendering. "Level designers" don't store them in textures, the compile tools that are associated with the engine do, and the engine takes care of displaying them appropriately.
Having precomputed lightmaps in the doom 3 engine would break all the internal consistency of the lighting. Mobile lights in engines based on precomputed lighting are treated differently from static lights. Doom 3 doesn't have this distinction.
The doom 3 approach allows lights to be much more dynamic, but when a light is that dynamic, you can't have precomputed light maps. You wouldn't have any way of updating them to reflect changing light conditions. Every time an imp warps in, all the lights dim. This couldn't be done realistically with precomputed light maps.
Adding precomputed light maps would require redoing all the internal assumptions about lights in the engine, and you would be basically writing your own.
Naturally, I was just saying that your "I can't see it happening" assertion was a little after the fact.
Um, it's already happening . . . My mom is using their service. She likes it.
My Mom has signed onto it too as part of their trial service. She was previously a zoomtown customer. I've had some issues with it dropping packets, and it seems to be pretty sensitive to where the "modem" (or whatever they call it) is plugged into your wall.
At first we just plugged it in wherever in the house we were using the computer near an outlet, and that worked ok. It was like having a home network with just a little added inconvenience. People apparantly liked the mobility of that, but didn't like dropping packets. They hooked my mom up with a wireless router, and tested all the outlets in the house to find the most stable one to hook it up to. The device is about 3"x4"x2".
I'm pretty sure they have to install something in the power system to enable it for an area. (At the transformer level maybe?)
I'm not sure Cincinnati was such a great choice for where to try this out first. We aren't the most tech savvy of places, and a lot of homes around here are pretty old and probably haven't had their wiring replaced in 50 years. It seems like a pretty adverse environment to be testing it in.
All in all, its pretty cool, although it gives me the willies plugging my ethernet cable directly into the wall socket some times. (http://www.fiftythree.org/etherkiller/) I wonder what happens when the device starts to go bad?
Cincinnati's got Graeter's, which is the best icecream in the world flat out.
Cincinnati's also got a excellent art museum, one of the best zoos in the country, great public libraries, and that wonderful heartwarming Skyline Chili.
Dayton on the other hand . . . hmm.