Sorry, you must go to the pricipal's office now.
"Lightman. What a, uh, surprise."
(It was Matthew Broderick in _War Games_, not Matthew Broderick in _Ferris Bueller's Day Off_)
Both my brother and I had the same problem: Our wireless routers weren't fast enough (even with a wired connection) to process the external data.
As a test, connect your computer directly to the Cable/DSL modem. See if your speed is any different. Through the router, my speed (using Speakeasy's speed test) maxed out at somewhere around 3.5 to 3 Mbps. A direct connect got over 9.
I switched over to using my Vonage-supplied router as the first connection, and I now get the full bandwidth. It seems that the Netgear router wasn't designed to handle connection speeds in excess of 3 Mbps, and it isn't even a year old.
I just purchased a Microdrive JVC unit -- it uses a CompactFlash 2 based hard disk, 4 GB in size. Records about an hour of 6 Mb/s DVD-compliant MPEG2 video.
The whole thing is the same size as my old still camera. I love it!
Apparently basic aeronautics was not one of the things he has a curiosity about.
Sailing into the wind is actually a way to increase altitude, since airplanes have a little thing I like to call "lift", which increases the faster the air moves over the wings.
This is of course not really relevant to the point he was trying to make, but it is still better not to use an outright false statement for an analogy.
I think you got that wrong. If the book states that out of a pool of 100, you need to have a preview group size of 37, how does that number go down to 14 when your pool size is one million?
The stated value of 37 came from calculating N/e, not ln N.
You would have to date (about) 367879 people to generate your preview pool. For most of us, that is impractical.
Huh, I wonder if that was county-wide. Because I actually saw the ticker-tape on our machines when I voted. (The tape, not the contents.) Maybe we got the new machines or something. Or Bob lied to you.
As an aside, I was rather shocked this election to see several Not-So-Senior Citizens (tm) working the polling place. One was even cute!
Why does the poster assume there is no paper trail? From reading too many stories on Slashdot?
I am a Washington State voter, and my whole county (Snohomish) uses the same type of voting machine. Other counties are different. But here you can watch the little ticker-tape coming out of the back of each machine. I don't know how the votes are encoded, but there is definitely a paper audit trail.
I'm actually concerned about the accuracy of the recount, since it is likely to be hand-counted (required by law when the difference is below some threshold -- I don't know the specifics.) Despite any bugs in the electronic systems that may or may not affect the count, hand-counting pretty much guarantees a certain margin of error.
Anyone ever tried to accurately count a stack of ten thousand pieces of paper, dividing them into two separate piles in the process? I screw up occassionally just separating puzzle pieces into separate groups of edge and center pieces -- for small (100 piece) jigsaw puzzles!
I actually think they can get the best of both worlds by offering backwards-compatibility as an add-on. Find a source for cheap hard drives (always a problem, I know), and make a unit that plugs into the back that provides the HD and a ROM with the CPU and GPU emulation software.
That way, a customer who wants backwards compatibility pays $40 or whatever and can justify their purchase. For everbody else, the box is cheaper. Microsoft doesn't lose money on every box because of the HD and licensing fees.
As others have speculated, I don't believe that all the hardware has to be included. Just the things that can't be emulated. The only things I can think of are the hard drive and possibly the sound chip, since it had a couple of ultra-powerful DSPs on it that nobody used.
But the reason I chose this watch is because it receives the national atomic clock radio time signal thingy and sets itself every night. So I always know exactly what time it is.
As another poster pointed out, he left Microsoft in 2001. He may very well have been bound by a non-complete agreement, but it would have expired by now.
I'm sorry, but 22KHz is not "perfectly reproduced on a standard CD." Nyquist tells us only the maximum frequency that can be quantized. It does not guarantee that it is perfect. In fact another poster pointed out that the high frequencies are quite prone to aliasing.
Consider a wave of exactly 1/2 the sample frequency:
If the sample points fall on the peaks and valleys, you will be able to exactly reproduce the wave. If they happen to fall on the zero-crossover points, you will get silence. Anywhere in-between will be interpreted as a reduced amplitude wave.
That's why it is always better to go beyond doubling the largest frequency you wish to reproduce. I also think it was just plain shortsighted for CD's to use PCM encoding. But, I understand that DSP chips where just too expensive to put into consumer devices back when the format was introduced. Too bad they didn't think ahead and allow for growth.
Huh. I've never had a problem with at Blockbuster with this, but stopped renting at Hollywood for that exact same reason.
Sometimes they never check it in, but usually they do it late and then I have to bitch at them to get the bogus late fee removed from my account the next time I go in.
I never did decide if this is something they are purposefully sloppy on (knowing that many people either don't remember or often return movies late), or if their process genuinely sucks. It doesn't seem to me like it would be a difficult thing to get right.
Actually, it is not true that consoles usually only draw at 640x240. The default mode on the Xbox is to render a complete 640x480 screen, and the display hardware uses the extra scan lines to implement a flicker filter. Developers are encouraged to use this mode partially for the visual difference, but also because the game can support 480p without modification. Since the fill-rate is the same, the user gets progressive scan for free.
You are mostly right about the higher resolutions, though. The big issue is that few enough customers can take advantage of them, so it's not worth implementing them. The problem is not necessarily fill-rate (although that can be a factor), but that it affects gameplay. A 16:9 aspect ratio requires either displaying more stuff on the sides, or cutting off the top and bottom. In either case, it affects camera tuning, cutscene creation, and sometimes even level design (to accomodate the different visual presentation.)
A couple of years ago a colleague did some part-time teaching at a local video game school. He didn't do it again because it was ridiculous. The stuff they were teaching was outdated, and way too specific.
For instance, they spent many weeks learning about fixed-point math. Now, it has been a long time since most of us have relied on using fixed-point numbers instead of floats for speed. Sometimes they are still used for compression, but they are always converted to floating-point before anything is done with them.
Plus, anyone with even a reasonable mathematical background can learn everything there is to know about fixed-point numbers in an hour. If the students needed several weeks, they probably aren't the sharpest crayons in the box.
When I hire a programmer, they need to be flexible and smart. Just out of university, I want to see someone who can communicate well and has a solid grounding in computer science. They will learn the rest.
(Of course, if they are just out of school then they'll be really cheap because they don't really know anything yet. Ha! But it doesn't take long for the stars to shine through and get promoted.)
Having just shipped an Xbox title that uses RDTSC for timing, I'm sorry to say that you are misinformed. It takes nowhere near a couple of seconds to determine the timing, and it is even possible to do it while other initialization is processing -- thus costing you no time at all.
I also have to repectfully disagree about your CPU-bound comment. The Xbox GPU is far more powerful than the lame PIII CPU it is saddled with (although I was continually thankful that it's not a stall-prone P4!) In most cases, the GPU ate the triangles as fast as the CPU could throw them at it. We were constantly CPU bound, and no slackers on optimzation.
I will grant that some types of games, specifically ones with large triangles where fill-rate is the determining factor, will almost always be GPU bound. But that is not all titles. A game that features good character models and animations will be processing more bone animation and denser geometry meshes than one featuring expansive terrain (for instance).
One other thing to note: The Technical Certification Requirements include making sure the title runs with faster (or larger) hardware, including hard drives. I don't recall any specific requirement about CPU speed, but didn't you at least consider that when designing your timing code? If nothing else it makes porting and upgrading your engine a lot easier.
Actually, Glide was a very good thing for them at the time. DirectX was a horrible pig, and OpenGL was only vaguely supported on most consumer boards.
I added Glide support to the engine I was building, and the game ran significantly better on a 3DFX card than any other because of it. (And yes, I'm an actual game programmer working for an actual game company. Not some college kid who has an opinion because they read it on John Carmack's.plan file.)
Sure, it would be stupid nowadays. But back then (circa 1997) it wasn't. And did you notice that they eventually dropped Glide when the other APIs caught up?
Slashdot Mirror
Sorry, you must go to the pricipal's office now. "Lightman. What a, uh, surprise." (It was Matthew Broderick in _War Games_, not Matthew Broderick in _Ferris Bueller's Day Off_)
As a test, connect your computer directly to the Cable/DSL modem. See if your speed is any different. Through the router, my speed (using Speakeasy's speed test) maxed out at somewhere around 3.5 to 3 Mbps. A direct connect got over 9.
I switched over to using my Vonage-supplied router as the first connection, and I now get the full bandwidth. It seems that the Netgear router wasn't designed to handle connection speeds in excess of 3 Mbps, and it isn't even a year old.
The whole thing is the same size as my old still camera. I love it!
#define PI 3.14159265358979323f
Note the "f" on the end. You will probably only get the first six or seven decimal digits. The rest is gravy for the compiler to throw away.
Sailing into the wind is actually a way to increase altitude, since airplanes have a little thing I like to call "lift", which increases the faster the air moves over the wings.
This is of course not really relevant to the point he was trying to make, but it is still better not to use an outright false statement for an analogy.
When filing for an extension, you have to pay your estimated tax due (and there are penalties if you under-estimate by very much.)
So, of the two certain things in life (death and taxes), it looks like taxes still trumps death!
The stated value of 37 came from calculating N/e, not ln N.
You would have to date (about) 367879 people to generate your preview pool. For most of us, that is impractical.
As an aside, I was rather shocked this election to see several Not-So-Senior Citizens (tm) working the polling place. One was even cute!
I am a Washington State voter, and my whole county (Snohomish) uses the same type of voting machine. Other counties are different. But here you can watch the little ticker-tape coming out of the back of each machine. I don't know how the votes are encoded, but there is definitely a paper audit trail.
I'm actually concerned about the accuracy of the recount, since it is likely to be hand-counted (required by law when the difference is below some threshold -- I don't know the specifics.) Despite any bugs in the electronic systems that may or may not affect the count, hand-counting pretty much guarantees a certain margin of error.
Anyone ever tried to accurately count a stack of ten thousand pieces of paper, dividing them into two separate piles in the process? I screw up occassionally just separating puzzle pieces into separate groups of edge and center pieces -- for small (100 piece) jigsaw puzzles!
(Not that there's anything wrong with that...)
Yes. I have moderator points, and I went to go mod you up, but couldn't because it was already at 5.
And just so you don't feel unusual, my wife came down on the Aragorn side in the infamous "Logolas vs. Aragorn" debate.
That way, a customer who wants backwards compatibility pays $40 or whatever and can justify their purchase. For everbody else, the box is cheaper. Microsoft doesn't lose money on every box because of the HD and licensing fees.
As others have speculated, I don't believe that all the hardware has to be included. Just the things that can't be emulated. The only things I can think of are the hard drive and possibly the sound chip, since it had a couple of ultra-powerful DSPs on it that nobody used.
But the reason I chose this watch is because it receives the national atomic clock radio time signal thingy and sets itself every night. So I always know exactly what time it is.
As another poster pointed out, he left Microsoft in 2001. He may very well have been bound by a non-complete agreement, but it would have expired by now.
The game that a previous poster's dad wrote would might get a zero because it was never published.
Except that Gamespot wouldn't review it, since it wasn't published.
Thus the scoring system starting at "1". This is what they were thinking when they designed the scale. I'm sure of it!
Consider a wave of exactly 1/2 the sample frequency:
If the sample points fall on the peaks and valleys, you will be able to exactly reproduce the wave. If they happen to fall on the zero-crossover points, you will get silence. Anywhere in-between will be interpreted as a reduced amplitude wave.
That's why it is always better to go beyond doubling the largest frequency you wish to reproduce. I also think it was just plain shortsighted for CD's to use PCM encoding. But, I understand that DSP chips where just too expensive to put into consumer devices back when the format was introduced. Too bad they didn't think ahead and allow for growth.
Expected Release December 2000
I'm not saying that Chris won't eventually release another RTS (I know he wants to make one), I'm just saying don't hold your breath for TA2
Another clue should have been all the stuff about hardware acceleration. <sarcasm>An RTS with a free camera? How forward looking!</sarcasm>
Sometimes they never check it in, but usually they do it late and then I have to bitch at them to get the bogus late fee removed from my account the next time I go in.
I never did decide if this is something they are purposefully sloppy on (knowing that many people either don't remember or often return movies late), or if their process genuinely sucks. It doesn't seem to me like it would be a difficult thing to get right.
You are mostly right about the higher resolutions, though. The big issue is that few enough customers can take advantage of them, so it's not worth implementing them. The problem is not necessarily fill-rate (although that can be a factor), but that it affects gameplay. A 16:9 aspect ratio requires either displaying more stuff on the sides, or cutting off the top and bottom. In either case, it affects camera tuning, cutscene creation, and sometimes even level design (to accomodate the different visual presentation.)
Good for you! But it must be a lot of work to maintain a mistress. Most of us with kids don't have that kind of energy.
A couple of years ago a colleague did some part-time teaching at a local video game school. He didn't do it again because it was ridiculous. The stuff they were teaching was outdated, and way too specific.
For instance, they spent many weeks learning about fixed-point math. Now, it has been a long time since most of us have relied on using fixed-point numbers instead of floats for speed. Sometimes they are still used for compression, but they are always converted to floating-point before anything is done with them.
Plus, anyone with even a reasonable mathematical background can learn everything there is to know about fixed-point numbers in an hour. If the students needed several weeks, they probably aren't the sharpest crayons in the box.
When I hire a programmer, they need to be flexible and smart. Just out of university, I want to see someone who can communicate well and has a solid grounding in computer science. They will learn the rest.
(Of course, if they are just out of school then they'll be really cheap because they don't really know anything yet. Ha! But it doesn't take long for the stars to shine through and get promoted.)
Canada has a warship?
(Sorry, had to be done. I love our neighbors to the north!)
Although rumor has it that they are now porting it back to the PC. Go figure.
I also have to repectfully disagree about your CPU-bound comment. The Xbox GPU is far more powerful than the lame PIII CPU it is saddled with (although I was continually thankful that it's not a stall-prone P4!) In most cases, the GPU ate the triangles as fast as the CPU could throw them at it. We were constantly CPU bound, and no slackers on optimzation.
I will grant that some types of games, specifically ones with large triangles where fill-rate is the determining factor, will almost always be GPU bound. But that is not all titles. A game that features good character models and animations will be processing more bone animation and denser geometry meshes than one featuring expansive terrain (for instance).
One other thing to note: The Technical Certification Requirements include making sure the title runs with faster (or larger) hardware, including hard drives. I don't recall any specific requirement about CPU speed, but didn't you at least consider that when designing your timing code? If nothing else it makes porting and upgrading your engine a lot easier.
I added Glide support to the engine I was building, and the game ran significantly better on a 3DFX card than any other because of it. (And yes, I'm an actual game programmer working for an actual game company. Not some college kid who has an opinion because they read it on John Carmack's .plan file.)
Sure, it would be stupid nowadays. But back then (circa 1997) it wasn't. And did you notice that they eventually dropped Glide when the other APIs caught up?