You're actually making my case for me. I'm not talking about a single course that goes over the topics I discussed. I'm talking about the fact that this should be at the core of a good CS education.
Big-O isn't something you just 'have a course on' and then move on, nor is it something that you needs to explain to anyone who has a more traditional programming degree. Big-O is actually one of the most important aspects of computer programming (in video games) because it comes up on a daily basis.
Optimization isn't really a worthwhile course in schooling either, because the things they're teaching you how to optimize (pre-versus-post increment) either:
1) Rarely come up as actual performance problems 2) Are topics that the compiler will optimize for you.
The things that generally wind up needing optimization that the compiler cannot help you with are things such as algorithmic inefficiencies (yes, the constants actually do matter), data layout inefficiency (for example, cache misses), poor branch prediction, failure to provide early-out of complex computations, etc.
The fact that there is even a route through the school that involves simply throwing money at the problem pushes my point along even further. Go suggest that someone at MIT, Stanford, UC Berkeley, U Texas or UT Dallas get their education by continually throwing money at the problem. There are no brute force routes through those schools. You either hack it, you change majors, or you go home.
Not to be overly dismissive, but my experience (both in working and interviewing) with Digipen graduates is that they lack the fundamentals necessary to deal with the programming tasks that come up on a very regular basis in game development.
Long story short, I give them a polite 'thanks but no thanks' when the resume comes across my desk now.
To those considering a stint at Digipen (or any other 'gaming university' or even a degree that focuses on game problems): Do yourself a favor. Get a real CS education from a school that teaches the fundamentals (data structures, algorithms, algorithm analysis, 3-D math). I can teach you graphics programming if you get that stuff. However I will never be able to convince you why you should've used a list instead of an array if it hasn't been ground into your head for at least four years.
What you don't seem to realize is that releasing specs that would allow you to speak directly with the hardware would necessarily reveal trade secrets about how the hardware works.
Consider the overall architecture of the layers are gone through when you make a call to the hardware in OpenGL or D3D: Application -> Open/GL / D3D -> Hardware Abstraction Layer -> Device Driver -> Hardware
The device driver is translating from a hardware independent language into a hardware dependent one. This means that the driver not only speaks the language of the HAL, but it also speaks the language of the hardware. Embedded in that language are the techniques of how the low level hardware gets its job done. If they expose that, they might as well start giving their trade secrets to their competetion.
There *was* a beginning to the universe, it is widely accepted that the universe is 'the ultimate free lunch,' and was brought into existence by a quantum fluctuation. Not only that, but we have a pretty good picture of what happened during the first 3 minutes, as well as what happened after that.
As any astronomer would explain to you, there is a point which we could call the end of the universe, depending on who you believe about when the universe will end. I'm speaking of the two possible eventual outcomes for our universe: Heat Death or collapse. (Although at this point the majority of cosmologists agree that our universe is expanding at an accelerating pace, and thus the universe is doomed to become a very chilly place indeed).
"But how," you ask, "can you claim that heat death is the end of the universe, or of time?" Sure, time will go on after heat death, but there will be nothing around to MEASURE that time, because all activity in the universe will have stopped with the temperature at a very cold 0K.
I also found it a bit funny that you think we need a new Galileo or Copernicus, when we have someone even better. Weinberg is one of the most brilliant physicists that has ever lived, certainly the most influential of my life (if that gives you a clue to how old I am). Don't get me wrong, I'm not trying to belittle Copernicus, Galileo, Newton or even Einstein. But Weinberg came up with a theory to unify two of the four fundamental forces (weak and electromagnetic), as well as the Inflationary theory.
Now, as far as escaping our universe and getting to another universe, this is all highly theoretical stuff. But there are a fair number of scientists who believe that our universe is but one in a larger multiverse. There are numerous theories about how these other universes would be formed, as well as how they could be kept seperate from our own which I will not go into here (some involve black holes, others involve the period of unknown time immediately (picoseconds) after the Big Bang.
so you don't have to worry about setting two timers...
Gotta agree with you. Having the DVR capabilities integrated with the TV guide interface
What? You guys have clearly not used Tivo. There's no setting two timers. And Tivo comes with its own channel grid (TV Guide functionality) which has a pretty nice interface. When I want to record a show, I navigate to it (or search for it), and I can record it. I can set season passes to get all episodes, or first runs only. I can set how many episodes of the show to keep on the box for a particular show (so for example, 10 for the Simpsons and 5 for CSI). It uses fuzzy logic to suggest programs that it thinks I might like based on existing programs I record (if I want it to). Now it can even play music and show picture albums I store on my PC in the other room. Pretty much everytime I think 'gee, I wish they'd fix this about tivo,' I find the issue fixed within a matter of weeks with the regular software patches it receives.
And yes, Tivo can access all of the channels that your cable box can receive--it's a unit that runs between your cable box (should you be unfortunate enough to have one) and your TV.
I personally have to agree with the great-grandparent. The Tivo UI is really quite excellent, and one of the reasons that the box was worth $450 ($250 for the box, $200 for the lifetime subscription). Plus the fact that it runs linux. I thought you/. types would salivate over that.
Evolution and creationism are not mutually exclusive concepts.
To your average Christian who reads the bible and interprets it as an allegory you are correct. However, there are sects of Christians (in particular Southern Baptists), who read the bible and state that the bible should be interpreted literally.
For them, there cannot be an evolution because it challenges their beliefs directly. To accept evolution would mean that there was a flaw in the bible, but to accept a flaw in the bible (to them) would mean that the entire book was invalid.
Of course evolution is a theory. Anyone who has studied it knows it's a theory. But theory doesn't mean to scientists what it means to the lay person; a theory isn't a casual thought about how something might work. A theory is something that has been rigorously examined, studied, poked prodded, and still stands as 'the best explanation that we have.' But unlike religion, theory can be altered and adjusted for new evidence. Whereas (strict) religion chokes when presented with evidence that conflicts with the bible, theory thrives on it. We tweak the incorrect bits and represent the full picture.
And yes, there are theories that cannot be tested in a laboratory. This does not necessarily make them less valid. Everything we study in astronomy is also based on theory. We observe events, try to make our best guess about how they work, and we test it against other parts of the universe. Lather, rinse, repeat.
And yes, sometimes the evidence presented in a high school text book doesn't seem to make them out to be that credible. But why is that surprising? It's a high school text book. It's not supposed to be authoritative. How many peer-reviewed journals did you have to read for high school biology? How many things did you learn in college that demonstrated that what you were taught in high school was a dramatic over-simplification of the truth?
But the same is true for mathematics... If I believed that my high school algebra book contained the end-all be-all of mathematics, and refused to accept that math was capable of doing more wonderful things, and was capable of evolving, I certainly wouldn't be a graphics programmer today.
Whew! That's right! That's why programmers are not engineers! Did you take the PE? Then you're not an engineer.
On the other hand, part of what makes a good programmer is 'customer focus,' the ability to translate from what someone tells you they want into what they actually want.
I find it funny that they didn't consider 'fundamentally a worhtless idea' one of the top reasons that projects fail.
Seriosuly, how many of you have worked on a project (probably you were assigned by someone who wanted to get rid of you), where you thought from the beginning of the project 'this is a terrible product?'
Even if the product ships (which is unlikely, because your coworkers probably also think the product is terrible, thus morale plummets, thus productivity plummets, thus...), the customer won't want it, and won't buy it.
Madonna at least waited two weeks...
on
Sir Peter Molyneux?
·
· Score: 0, Redundant
Apparently now that he's been knighted, he has to speak like he's English: It's come completely out of the blue, I never would have guessed that I'd have that kind of honour.
No offense, but there's a few reasons we don't use ogl for developing video games that you're glossing over.
First, I like a standard that updates itself frequently. DirectX does this. In the time that it took for OGL to go from 1.2 to 2.0, DX went from roughly DX6 to DX9. What did you do if you were writing games in OGL during that time? Oh, you wrote ALL OF YOUR CODE TWICE, once for NVidia, and once for ATI, and hopefully you didn't care about other vendors. This is why when you look at Carmack's old.plans, he talks about the various codepaths.
Second, I like a standard that really pays attention to what I (the graphics programmer), has to say. Input is key. When I have a feature request, I notify MS, NVidia, and ATI. I convince each of them why it's a good idea. Generally, it makes it into the next release. Of course, I've tried this with ogl as well.. Not such a simple process. Of course, you can still bring it up with the vendors, but then you also have to bring it up with the standards comittee, and I can't just leave it at that even.
Third, did I mention that I don't want to write my code twice? I really hate doing that. It's bad enough that I have to check caps bits in the first place, and have to do minor if statements to deal with various hardware. But what I really want to avoid is having to write code to several seperate extensions.
OGL may become a powerful force in the (gaming) market in the next few years, time will tell. But it will be largely dependent on how quickly the standards comittee can get off their ass and update. If DX continues to release 2-3 major revs (and countless minor revs) in the same time it takes for OGL to update, then DX will remain the dominant platform to develop for, regardless of the 10% or so linux marketshare lost.
(Incidentally, in terms of 'borrowing' from OGL, that hasn't really been true since Dx8. Since then, OGL has really been trying to play catchup).
I purchased a flat 17" Samsung monitor when they first came out, at newegg.com. After receiving the monitor, I found that it had 4 dead pixels, in very conspicuous places (pretty much right at the center of the monitor).
I contacted NewEgg about this. At first, they said that they weren't going to do anything about it, and I had to take it up with Samsung. (They had said there was nothing they could do because it had been over 30 days since the purchase of the monitor, despite the fact that I had *received* the monitor that day, due to back order)
I contacted Samsung, and spoke with what sounded like an Italian mobster (no offense to any Italians, it's just what he sounded like.) He told me that Samsung's policy was not to do anything unless there were more then 7 dead pixels. After speaking to him for awhile, he told me that he could RMA me a REFURBISHED MONITOR. I kindly told him where he could stick his refurbished monitor and hung up.
I contacted NewEgg again, and told them that I had contacted the manufacturer and they weren't going to do anything about it. I told them that I had just received the monitor that day, and that it had dead pixels and this was utterly unacceptable. NewEgg maintained that it had been more then 30 days since they had billed me, so it must be the manufacturer's problem. I assured them that if I didn't have a return authorization code by the time I hung up the phone, I was going to contact my credit card company and issue a stop-payment on the entire order (about $2500 total), and I was going to ship the entire order back, with or without return authorization. They then decided that they could check the shipping information, and *lo-and-behold,* the monitor had only shipped four days before, and had arrived that day. I shipped back the monitor, and vowed never to purchase from Samsung again.
I purchased a Sony Multiscan E400, which I still use. Much better monitor, and a *way* better dead pixel policy (at the time it was 0 dead pixels, which would be replaced overnight with a new monitor).
And if you RTFA, you'd realize that no one is saying that chance will determine whether the rock slams into the planet or not.
They're saying that based on the current data they have, they've plotted out the 'cloud of possible locations' as the object passes by the earth, and some percentage of them (1 in 300) involves a direct collision with the third rock.
What's even funnier about this is the article at CNN.com about this (go ahead, I'll wait):
"This is not a problem for anyone and it shouldn't be a concern to anyone, but whenever we post one of these things and... somebody gets ahold of it, it just gets crazy," [Donald Yeomans, manager of the Near Earth Object Program at NASA's Jet Propulsion Laboratory in Pasadena] said."
In fact, now that they've noticed this one, they're going to pay closer attention to it and they will be able to predict more accurately where the asteroid is going to pass by the earth, likely ruling out a collision.
Re:Review: Review: Prime Obsession
on
Prime Obsession
·
· Score: 1
The problem is *where* you define atoms of the statement. In a sentence, words are atoms, and are seperated by spaces.
Byyoursuggestion,sentencesshouldbewrittenlikethi sa slesseyetrackingneeded.Whichisalwaystheprimeculpri tforhowlongittakestoreadsomething.
In English, humans parse on whitspace first, then a combination of the first and last few letters of a word. This is why that e-mail went around awhile back that suggseted taht wrods wtih letetrs swpaped aruond colud sitll be raed qucikly. That is, reading is primarily a game of pattern recognition once you have been reading for a significant amount of time. You cease to read every letter and begin to read words. There are of course, exceptions to this rule (particularly words that are often not used, such as antidisestablishmententarianism).
In math (and especially in programming), there is not such a strong familiarity with the 'words' used. It is unlikely that a piece of code being debugged has been around for 10-15 years. It is far *less* likely that *you* have maintained that particular piece of code and have spent any significant time in that code on a daily basis (and your respective deity help you if you have).
As an example, here is a significantly more complex example:
(b*d-1)|(d+1)*(a+c)==(b*d-1)|(b+1)*(a+c)
Which can be more clearly written as: ((b * d - 1) | ((d + 1) * (a + c))) == ((b * d - 1) | ((b + 1) * (a + c)))
(and of course, this could be improved upon further by pulling out like terms and computing them in advance).
Then again, maybe we should do what you suggested. I imagine that the classic works would be much more enjoyable without whitespace:
Itwasthebestoftimes,itwastheworstoftimes...
Review: Review: Prime Obsession
on
Prime Obsession
·
· Score: 4, Insightful
... f(x)=x^2-2x+1... x=1 where f(x)=0
It's called white space. Look into it. Humans parse on it much faster then they parse on operators.
Yes, I'm not sure how he expects this to be easier for programmers (which is one of the things he lists as being a problem with the current calendar).
So, instead of a fairly simple (recursive) leap-year rule that I have to follow now, I instead have to check with a website if this year is a newton-year or not? What happens if I get a 404?
Everytime there is a format change most consumers go buy
Every time? There's only been a major format change once. In fact, it's exactly because consumers didn't buy new copies of the films they already owned that laserdisc didn't take off...
I still know quite a few people who don't own a DVD player, and use those large rectangular black plastic thingies to play movies.
All the "maintenance" jobs in the world can move overseas, but you still need people back at home making the decisions. I'll become one of those.
(Emphasis added)
When the primary workforce moves to Inda, China, South Korea, or somewhere else, middle management will move with it. What do you think, that the plant that put your last car together in Mexico left its middle management in Detroit?
And following similar logic, once you have very little middle-management here, director level posistions will migrate as well. The boards will remain here, of course, but the rest of the company will be in India.
I hope your MBA can get you to a boardroom in the next 5 years. Otherwise you might as well stop with the MBA now and start flipping burgers. After all, jobs are determined by the same market forces as everything else, and there is definitely not a shortage of obesity.
(Yeah, apparently I hit 'enter' a little quickly today. Full post follows...)
Because making the graphics card go the whole way back to the system is a lot slower than just using RAM right on the card?
That's actually not really true. The whole point of PCIe and the next generation of cards is that with a solid bus, access to system RAM can be within an order of magnitude of video RAM.
Yes it will still be slower then DRAM on the card, but the programmer time saved will allow for performance improvements in other places.
The point of PCIe is to add full-duplex fast bus access. This allows numerous things like (for example) using the GPU as a number of massively parallel multi-processors. In the past, this wasn't really feasible for games, because AGP was so painfully slow transferring backwards, and because if you *were* transferring backwards you couldn't be transferring forwards.
There are other problems with the ever spiralling amount of memory on video cards as well. For example, how much system memory do you need to run a video card with 512M of ram at full capacity? It's a lot more then you think. If you only have a gig of system RAM, windows is going to freak out whenever an application creates a device using the renderer, because one single 512M allocation just occurred, and windows gets pretty upset when half of memory is actively in use.
Because making the graphics card go the whole way back to the system is a lot slower than just using RAM right on the card?
That's actually not really true. The whole point of PCIe and the next generation of cards is that with a solid bus, access to system RAM can be within an order of magnitude of video RAM.
An open source gaming engine will eventually surpass a closed source one, however the issue right now is that there is so much more money to be had developing one closed source. But even that cannot delay the inevitable.
What is your reference for this statement? Why is it inevitable? Here's my list of closed-source engines. Let me know which of these will be surpassed in the short-term. Or, consider equivalently powerful engines for whatever period you would like to choose for the future.
It also might help if a certain powerful, ubiquitous church would change their ridiculuous stand on birth control.
I'm serious. I was brought up a catholic, and their stance on condom use (that it is a mortal sin) is utter lunacy. Yeah, that's great that you provide care for nearly 25% of AIDS patients. You know what would be better? 25% *fewer* AIDS patients. People not getting sick beats caring for sick people *every day of the week.*
It's really a minor flaw that most companies would never even come back to fix.
That's not really true at all. I try to fix rhythmic bugs (or periodic spikes) first, because they are: 1) Easy to identify (with the proper performance monitoring tools) 2) Very distracting to players and 3) Does tend to cause motion sickness, even in people otherwise unaffected by video games.
They are also the most likely performance problems to require system rewrites, because they are generally indicative of something that you are doing wrong from time to time.
Will they be going with the original ending of AoD (where he gets back to S-mart), or the 'official' bootleg edition where he wakes up too late and every one is already dead?
That franchise has been fixed and tweaked and remade so many times that I haven't been this confused since the last time I saw a Kevin Costner movie...
Slashdot Mirror
You're actually making my case for me. I'm not talking about a single course that goes over the topics I discussed. I'm talking about the fact that this should be at the core of a good CS education.
Big-O isn't something you just 'have a course on' and then move on, nor is it something that you needs to explain to anyone who has a more traditional programming degree. Big-O is actually one of the most important aspects of computer programming (in video games) because it comes up on a daily basis.
Optimization isn't really a worthwhile course in schooling either, because the things they're teaching you how to optimize (pre-versus-post increment) either:
1) Rarely come up as actual performance problems
2) Are topics that the compiler will optimize for you.
The things that generally wind up needing optimization that the compiler cannot help you with are things such as algorithmic inefficiencies (yes, the constants actually do matter), data layout inefficiency (for example, cache misses), poor branch prediction, failure to provide early-out of complex computations, etc.
The fact that there is even a route through the school that involves simply throwing money at the problem pushes my point along even further. Go suggest that someone at MIT, Stanford, UC Berkeley, U Texas or UT Dallas get their education by continually throwing money at the problem. There are no brute force routes through those schools. You either hack it, you change majors, or you go home.
Not to be overly dismissive, but my experience (both in working and interviewing) with Digipen graduates is that they lack the fundamentals necessary to deal with the programming tasks that come up on a very regular basis in game development.
Long story short, I give them a polite 'thanks but no thanks' when the resume comes across my desk now.
To those considering a stint at Digipen (or any other 'gaming university' or even a degree that focuses on game problems): Do yourself a favor. Get a real CS education from a school that teaches the fundamentals (data structures, algorithms, algorithm analysis, 3-D math). I can teach you graphics programming if you get that stuff. However I will never be able to convince you why you should've used a list instead of an array if it hasn't been ground into your head for at least four years.
What you don't seem to realize is that releasing specs that would allow you to speak directly with the hardware would necessarily reveal trade secrets about how the hardware works.
Consider the overall architecture of the layers are gone through when you make a call to the hardware in OpenGL or D3D:
Application -> Open/GL / D3D -> Hardware Abstraction Layer -> Device Driver -> Hardware
The device driver is translating from a hardware independent language into a hardware dependent one. This means that the driver not only speaks the language of the HAL, but it also speaks the language of the hardware. Embedded in that language are the techniques of how the low level hardware gets its job done. If they expose that, they might as well start giving their trade secrets to their competetion.
No offense mate, but you're off your rocker.
There *was* a beginning to the universe, it is widely accepted that the universe is 'the ultimate free lunch,' and was brought into existence by a quantum fluctuation. Not only that, but we have a pretty good picture of what happened during the first 3 minutes, as well as what happened after that.
As any astronomer would explain to you, there is a point which we could call the end of the universe, depending on who you believe about when the universe will end. I'm speaking of the two possible eventual outcomes for our universe: Heat Death or collapse. (Although at this point the majority of cosmologists agree that our universe is expanding at an accelerating pace, and thus the universe is doomed to become a very chilly place indeed).
"But how," you ask, "can you claim that heat death is the end of the universe, or of time?" Sure, time will go on after heat death, but there will be nothing around to MEASURE that time, because all activity in the universe will have stopped with the temperature at a very cold 0K.
I also found it a bit funny that you think we need a new Galileo or Copernicus, when we have someone even better. Weinberg is one of the most brilliant physicists that has ever lived, certainly the most influential of my life (if that gives you a clue to how old I am). Don't get me wrong, I'm not trying to belittle Copernicus, Galileo, Newton or even Einstein. But Weinberg came up with a theory to unify two of the four fundamental forces (weak and electromagnetic), as well as the Inflationary theory.
Now, as far as escaping our universe and getting to another universe, this is all highly theoretical stuff. But there are a fair number of scientists who believe that our universe is but one in a larger multiverse. There are numerous theories about how these other universes would be formed, as well as how they could be kept seperate from our own which I will not go into here (some involve black holes, others involve the period of unknown time immediately (picoseconds) after the Big Bang.
What? You guys have clearly not used Tivo. There's no setting two timers. And Tivo comes with its own channel grid (TV Guide functionality) which has a pretty nice interface. When I want to record a show, I navigate to it (or search for it), and I can record it. I can set season passes to get all episodes, or first runs only. I can set how many episodes of the show to keep on the box for a particular show (so for example, 10 for the Simpsons and 5 for CSI). It uses fuzzy logic to suggest programs that it thinks I might like based on existing programs I record (if I want it to). Now it can even play music and show picture albums I store on my PC in the other room. Pretty much everytime I think 'gee, I wish they'd fix this about tivo,' I find the issue fixed within a matter of weeks with the regular software patches it receives.
And yes, Tivo can access all of the channels that your cable box can receive--it's a unit that runs between your cable box (should you be unfortunate enough to have one) and your TV.
I personally have to agree with the great-grandparent. The Tivo UI is really quite excellent, and one of the reasons that the box was worth $450 ($250 for the box, $200 for the lifetime subscription). Plus the fact that it runs linux. I thought you
Evolution and creationism are not mutually exclusive concepts.
To your average Christian who reads the bible and interprets it as an allegory you are correct. However, there are sects of Christians (in particular Southern Baptists), who read the bible and state that the bible should be interpreted literally.
For them, there cannot be an evolution because it challenges their beliefs directly. To accept evolution would mean that there was a flaw in the bible, but to accept a flaw in the bible (to them) would mean that the entire book was invalid.
Of course evolution is a theory. Anyone who has studied it knows it's a theory. But theory doesn't mean to scientists what it means to the lay person; a theory isn't a casual thought about how something might work. A theory is something that has been rigorously examined, studied, poked prodded, and still stands as 'the best explanation that we have.' But unlike religion, theory can be altered and adjusted for new evidence. Whereas (strict) religion chokes when presented with evidence that conflicts with the bible, theory thrives on it. We tweak the incorrect bits and represent the full picture.
And yes, there are theories that cannot be tested in a laboratory. This does not necessarily make them less valid. Everything we study in astronomy is also based on theory. We observe events, try to make our best guess about how they work, and we test it against other parts of the universe. Lather, rinse, repeat.
And yes, sometimes the evidence presented in a high school text book doesn't seem to make them out to be that credible. But why is that surprising? It's a high school text book. It's not supposed to be authoritative. How many peer-reviewed journals did you have to read for high school biology? How many things did you learn in college that demonstrated that what you were taught in high school was a dramatic over-simplification of the truth?
But the same is true for mathematics... If I believed that my high school algebra book contained the end-all be-all of mathematics, and refused to accept that math was capable of doing more wonderful things, and was capable of evolving, I certainly wouldn't be a graphics programmer today.
It's not the job of engineers goddamit !
Whew! That's right! That's why programmers are not engineers! Did you take the PE? Then you're not an engineer.
On the other hand, part of what makes a good programmer is 'customer focus,' the ability to translate from what someone tells you they want into what they actually want.
And the bloody glove.
I find it funny that they didn't consider 'fundamentally a worhtless idea' one of the top reasons that projects fail.
...), the customer won't want it, and won't buy it.
Seriosuly, how many of you have worked on a project (probably you were assigned by someone who wanted to get rid of you), where you thought from the beginning of the project 'this is a terrible product?'
Even if the product ships (which is unlikely, because your coworkers probably also think the product is terrible, thus morale plummets, thus productivity plummets, thus
Apparently now that he's been knighted, he has to speak like he's English: It's come completely out of the blue, I never would have guessed that I'd have that kind of honour.
No offense, but there's a few reasons we don't use ogl for developing video games that you're glossing over.
.plans, he talks about the various codepaths.
First, I like a standard that updates itself frequently. DirectX does this. In the time that it took for OGL to go from 1.2 to 2.0, DX went from roughly DX6 to DX9. What did you do if you were writing games in OGL during that time? Oh, you wrote ALL OF YOUR CODE TWICE, once for NVidia, and once for ATI, and hopefully you didn't care about other vendors. This is why when you look at Carmack's old
Second, I like a standard that really pays attention to what I (the graphics programmer), has to say. Input is key. When I have a feature request, I notify MS, NVidia, and ATI. I convince each of them why it's a good idea. Generally, it makes it into the next release. Of course, I've tried this with ogl as well.. Not such a simple process. Of course, you can still bring it up with the vendors, but then you also have to bring it up with the standards comittee, and I can't just leave it at that even.
Third, did I mention that I don't want to write my code twice? I really hate doing that. It's bad enough that I have to check caps bits in the first place, and have to do minor if statements to deal with various hardware. But what I really want to avoid is having to write code to several seperate extensions.
OGL may become a powerful force in the (gaming) market in the next few years, time will tell. But it will be largely dependent on how quickly the standards comittee can get off their ass and update. If DX continues to release 2-3 major revs (and countless minor revs) in the same time it takes for OGL to update, then DX will remain the dominant platform to develop for, regardless of the 10% or so linux marketshare lost.
(Incidentally, in terms of 'borrowing' from OGL, that hasn't really been true since Dx8. Since then, OGL has really been trying to play catchup).
I have to disagree.
I purchased a flat 17" Samsung monitor when they first came out, at newegg.com. After receiving the monitor, I found that it had 4 dead pixels, in very conspicuous places (pretty much right at the center of the monitor).
I contacted NewEgg about this. At first, they said that they weren't going to do anything about it, and I had to take it up with Samsung. (They had said there was nothing they could do because it had been over 30 days since the purchase of the monitor, despite the fact that I had *received* the monitor that day, due to back order)
I contacted Samsung, and spoke with what sounded like an Italian mobster (no offense to any Italians, it's just what he sounded like.) He told me that Samsung's policy was not to do anything unless there were more then 7 dead pixels. After speaking to him for awhile, he told me that he could RMA me a REFURBISHED MONITOR. I kindly told him where he could stick his refurbished monitor and hung up.
I contacted NewEgg again, and told them that I had contacted the manufacturer and they weren't going to do anything about it. I told them that I had just received the monitor that day, and that it had dead pixels and this was utterly unacceptable. NewEgg maintained that it had been more then 30 days since they had billed me, so it must be the manufacturer's problem. I assured them that if I didn't have a return authorization code by the time I hung up the phone, I was going to contact my credit card company and issue a stop-payment on the entire order (about $2500 total), and I was going to ship the entire order back, with or without return authorization. They then decided that they could check the shipping information, and *lo-and-behold,* the monitor had only shipped four days before, and had arrived that day. I shipped back the monitor, and vowed never to purchase from Samsung again.
I purchased a Sony Multiscan E400, which I still use. Much better monitor, and a *way* better dead pixel policy (at the time it was 0 dead pixels, which would be replaced overnight with a new monitor).
They're saying that based on the current data they have, they've plotted out the 'cloud of possible locations' as the object passes by the earth, and some percentage of them (1 in 300) involves a direct collision with the third rock.
What's even funnier about this is the article at CNN.com about this (go ahead, I'll wait):
In fact, now that they've noticed this one, they're going to pay closer attention to it and they will be able to predict more accurately where the asteroid is going to pass by the earth, likely ruling out a collision.
The problem is *where* you define atoms of the statement. In a sentence, words are atoms, and are seperated by spaces.
i sa slesseyetrackingneeded.Whichisalwaystheprimeculpri tforhowlongittakestoreadsomething.
Byyoursuggestion,sentencesshouldbewrittenliketh
In English, humans parse on whitspace first, then a combination of the first and last few letters of a word. This is why that e-mail went around awhile back that suggseted taht wrods wtih letetrs swpaped aruond colud sitll be raed qucikly. That is, reading is primarily a game of pattern recognition once you have been reading for a significant amount of time. You cease to read every letter and begin to read words. There are of course, exceptions to this rule (particularly words that are often not used, such as antidisestablishmententarianism).
In math (and especially in programming), there is not such a strong familiarity with the 'words' used. It is unlikely that a piece of code being debugged has been around for 10-15 years. It is far *less* likely that *you* have maintained that particular piece of code and have spent any significant time in that code on a daily basis (and your respective deity help you if you have).
As an example, here is a significantly more complex example:
(b*d-1)|(d+1)*(a+c)==(b*d-1)|(b+1)*(a+c)
Which can be more clearly written as:
((b * d - 1) | ((d + 1) * (a + c))) ==
((b * d - 1) | ((b + 1) * (a + c)))
(and of course, this could be improved upon further by pulling out like terms and computing them in advance).
Then again, maybe we should do what you suggested. I imagine that the classic works would be much more enjoyable without whitespace:
Itwasthebestoftimes,itwastheworstoftimes...
It's called white space. Look into it. Humans parse on it much faster then they parse on operators.
Yes, I'm not sure how he expects this to be easier for programmers (which is one of the things he lists as being a problem with the current calendar).
So, instead of a fairly simple (recursive) leap-year rule that I have to follow now, I instead have to check with a website if this year is a newton-year or not? What happens if I get a 404?
Every time? There's only been a major format change once. In fact, it's exactly because consumers didn't buy new copies of the films they already owned that laserdisc didn't take off...
I still know quite a few people who don't own a DVD player, and use those large rectangular black plastic thingies to play movies.
When the primary workforce moves to Inda, China, South Korea, or somewhere else, middle management will move with it. What do you think, that the plant that put your last car together in Mexico left its middle management in Detroit?
And following similar logic, once you have very little middle-management here, director level posistions will migrate as well. The boards will remain here, of course, but the rest of the company will be in India.
I hope your MBA can get you to a boardroom in the next 5 years. Otherwise you might as well stop with the MBA now and start flipping burgers. After all, jobs are determined by the same market forces as everything else, and there is definitely not a shortage of obesity.
That's actually not really true. The whole point of PCIe and the next generation of cards is that with a solid bus, access to system RAM can be within an order of magnitude of video RAM.
Yes it will still be slower then DRAM on the card, but the programmer time saved will allow for performance improvements in other places.
The point of PCIe is to add full-duplex fast bus access. This allows numerous things like (for example) using the GPU as a number of massively parallel multi-processors. In the past, this wasn't really feasible for games, because AGP was so painfully slow transferring backwards, and because if you *were* transferring backwards you couldn't be transferring forwards.
There are other problems with the ever spiralling amount of memory on video cards as well. For example, how much system memory do you need to run a video card with 512M of ram at full capacity? It's a lot more then you think. If you only have a gig of system RAM, windows is going to freak out whenever an application creates a device using the renderer, because one single 512M allocation just occurred, and windows gets pretty upset when half of memory is actively in use.
That's actually not really true. The whole point of PCIe and the next generation of cards is that with a solid bus, access to system RAM can be within an order of magnitude of video RAM.
Additionally, the
Instead of bitching about how broken the USPTO is, and how the patents they grant are obvious...
Get a job at the USPTO as a patent master.
What is your reference for this statement? Why is it inevitable? Here's my list of closed-source engines. Let me know which of these will be surpassed in the short-term. Or, consider equivalently powerful engines for whatever period you would like to choose for the future.
Source
DOOM 3
Unreal *
Renderware
Gamebryo
Quake *
Farcry
Versus:
Irrlicht
XEngine
That terrible "cube" renderer that supports multitexturing
OSS prides itself on being really quick to fix things, but frankly, it's really sluggish when it comes to bleeding edge technology.
PS: Sadly, I maintain my own OSS rendering technology, which is also a bit laggy compared to existing technologies.
Fortunately, I also maintain a closed source engine which lets me work on all the bleeding edge problems I want.
It also might help if a certain powerful, ubiquitous church would change their ridiculuous stand on birth control.
I'm serious. I was brought up a catholic, and their stance on condom use (that it is a mortal sin) is utter lunacy. Yeah, that's great that you provide care for nearly 25% of AIDS patients. You know what would be better? 25% *fewer* AIDS patients. People not getting sick beats caring for sick people *every day of the week.*
That's not really true at all. I try to fix rhythmic bugs (or periodic spikes) first, because they are:
1) Easy to identify (with the proper performance monitoring tools)
2) Very distracting to players
and
3) Does tend to cause motion sickness, even in people otherwise unaffected by video games.
They are also the most likely performance problems to require system rewrites, because they are generally indicative of something that you are doing wrong from time to time.
(AoD Spoiler alert)
Will they be going with the original ending of AoD (where he gets back to S-mart), or the 'official' bootleg edition where he wakes up too late and every one is already dead?
That franchise has been fixed and tweaked and remade so many times that I haven't been this confused since the last time I saw a Kevin Costner movie...