Disc medium: "UMD" 60mm optical secured ROM disc with cartridge (1.8GB)
60mm = 2.36 inches
GameCube disc = 80mm = 3.125 inches
Also, "with cartridge" suggests that the UMD will be encased in a cartridge, presumably like the minidisc.
GameCube discs store ~1.4-1.5 GB using one layer. Presumably UMD uses 2 layers to achieve greater capacity with its smaller size (either that or blue laser). This adds complexity to the drive mechanism.
I've used a halogen torchier lamp to desolder some connectors.
Didn't work as well as I would have liked. The heat was uneven and the middle of the connector started to melt a bit before the ends released. Would have worked fine for a shorter connector.
Would like to have used a blowtorch, but didn't have one handy.
A direct-mapped cache essentially strips off the high-order address bits (removes the upper digits of the number, leaving just the lower digits). This works okay in some cases, but can result in a lot of collisions (ie, addresses that map to the same location in the cache).
A collision just kicks out the old data and puts in the new. This can be a problem if you need that old data again real soon. Remember that main memory can be 100 times (or more) slower than the cache.
Adding associativity avoids this problem. For 2-way associative, you can have 2 addresses map to the same place before you've got problems. For 4-way, up to 4 addresses can map to the same location before you've got problems, etc.
Normally, each location in the cache keeps track of the upper address bits of the original location it's storing, plus the actual data. For a direct-mapped cache, you just need 1 set of hardware to compare the upper address bits against the appropriate location in the cache. For 2-way associative, there are up to 2 addresses (with the same lower address bits) that you need to compare (the upper address bits) against. So to do this fast (which is the whole point behind the cache), you need 2 sets of hardware to do the comparing at the same time. And similarly for 4-way and 8-way associative.
Standard disclaimer: This is all per memory. I could be completely wrong.
Unlike Sony and Nintendo, Microsoft _continues_ to lose money on hardware. Sony and Nintendo engineered (ie, cost-reduced) their hardware to eventually make money.
It's just like a P3, but it has half the amount of cache. So it probably is a P3 exactly, but the cache was probably modular to begin with, so they could just disable a defective half.
What were the other differences between a P3 and the comparable Celeron? Could they just all have been produced on the same assembly line, with features disabled to make a P3 into a Celeron?
> No proven model for network games? Surely Sony doesn't continue to produce expansions and sequels to Everquest just so they can lose more money.
You may have a point. But on the other hand, it will be interesting to see what happens to this market when it gets flooded with online, fee-based games.
> If memory serves me I believe the video settings for hi-def are in the Xbox dashboard and not in individual games which would lead me to believe that all games support it. I'm probably wrong on that though.
You are wrong. There is only one Xbox game that supports 1080i: Dragon's Lair 3D. There are a couple of games that support 720p (though not all HDTVs support 720p). There's a chart of this support at a gaming site somewhere.
> [hard drive comments]
Saving a game to HD instead of a memory card is a matter of convenience; it's not really a new ability. Faster loading time? Initial load speed is still limited to optical disk speed, since games can't install onto the HD. How many games actually make use of the HD to speed loading? And patches? Do you want your games to have patches? Add-ons can always come from an add-on optical disk.
> [Nintendo/Sony] costs are less due to missing features when compared to the Xbox.
That's exactly my point. They know how to make a product that's profitable. Microsoft doesn't.
The PS2 has an integrated DVD drive (ie, it's not a standard 5.25"-bay-size peripheral inside there like the Xbox has), and the expansion slot is the size of a 1" tall 3.5" hard drive.
I'm sure the first consoles were sold at a loss. Lots of engineering is devoted to cost reductions, and eventually the consoles are sold at a profit. Then there's a price-break, and the cycle may repeat a few times.
The CPU and graphics chip for the PS2 were initially produced in 0.25-micron, and were 240 mm^2 and 280 mm^2 initially. In case you don't know, this is f&cking HUGE! Read up on VLSI technology.
In short, $450 was probably pretty close to the mark, if not conservative; but it is irrelevant now, thanks to the ever advancing semiconductor progress train.
Market penetration of the console is only part of the picture that potential developers look at, I would hope. Actual game sales figures are another important factor. Thus, even if Microsoft ships an Xbox to every home in America, it won't help them much if everyone puts the box in the closet and buys no games. Sure, it may give them some leverage, but not nearly so much as strong game sales figures would.
> take a course in statistics... minimum sample size needed is 30 [...]
There's also a difference between statistics and data points. The previous poster didn't make any statistical comments; he was just providing his data points.
I've done a little research, and I think I've figured out the aiming:
The phase delay from each micro-antenna must be electronically adjustable. By carefully tuning the delay from each element, you can effectively aim the antenna.
This makes sense if you imagine the wave fronts as they intersect with the antenna. By carefully choosing what "points in time" are added up from each element, you can put together the wave fronts coming from the direction of your choice.
I still haven't figured out the polarization issue, but I imagine that they effectively have two antennas at right angles to each other, and that by doing some trignometric calculations on the two signals, they can figure out the original cross-polarized signals.
I know the very basics of a phased-array antenna: you've got a bunch of little antennas, and you combine their signals by matching the phase delay from the feed line for each one.
But how do you aim them in the case of a mobile unit like this? The only other phased-array DBS antenna I've seen requires you to point the whole thing.
Also, DirecWay uses linear polarization (vs. circular polarization for DBS). How is that handled? (ie, you've got to deal with not only azimuth and elevation, but also skew angle.)
Slashdot Mirror
To prevent earthquakes from killing people, we adapt our building codes so that the buildings don't fall down when earthquakes hit.
To prevent volcanoes from killing people, it would probably be wise to keep people from living next to them.
To prevent flooding rivers from killing people, is our only recourse to "kill" the river? Couldn't we build flood-proof housing?
Where does breathable air fit in?
From the press release:
Disc medium: "UMD" 60mm optical secured ROM disc with cartridge (1.8GB)
This does not sound like a GameCube disc, which is 80mm and no cartridge and only ~1.4-1.5 GB.
According to the Sony press release:
Disc medium: "UMD" 60mm optical secured ROM disc with cartridge (1.8GB)
60mm = 2.36 inches
GameCube disc = 80mm = 3.125 inches
Also, "with cartridge" suggests that the UMD will be encased in a cartridge, presumably like the minidisc.
GameCube discs store ~1.4-1.5 GB using one layer. Presumably UMD uses 2 layers to achieve greater capacity with its smaller size (either that or blue laser). This adds complexity to the drive mechanism.
Battery Life (according to Game Gear): that thing which is to be eliminated at all costs
No, don't download the series.
Download a couple of episodes. If you like it, you can buy the series, legally.
I've used a halogen torchier lamp to desolder some connectors.
Didn't work as well as I would have liked. The heat was uneven and the middle of the connector started to melt a bit before the ends released. Would have worked fine for a shorter connector.
Would like to have used a blowtorch, but didn't have one handy.
Yes, but if you believe everything Gord says, you will be rather misinformed.
A direct-mapped cache essentially strips off the high-order address bits (removes the upper digits of the number, leaving just the lower digits). This works okay in some cases, but can result in a lot of collisions (ie, addresses that map to the same location in the cache).
A collision just kicks out the old data and puts in the new. This can be a problem if you need that old data again real soon. Remember that main memory can be 100 times (or more) slower than the cache.
Adding associativity avoids this problem. For 2-way associative, you can have 2 addresses map to the same place before you've got problems. For 4-way, up to 4 addresses can map to the same location before you've got problems, etc.
Normally, each location in the cache keeps track of the upper address bits of the original location it's storing, plus the actual data. For a direct-mapped cache, you just need 1 set of hardware to compare the upper address bits against the appropriate location in the cache. For 2-way associative, there are up to 2 addresses (with the same lower address bits) that you need to compare (the upper address bits) against. So to do this fast (which is the whole point behind the cache), you need 2 sets of hardware to do the comparing at the same time. And similarly for 4-way and 8-way associative.
Standard disclaimer: This is all per memory. I could be completely wrong.
Unlike Sony and Nintendo, Microsoft _continues_ to lose money on hardware. Sony and Nintendo engineered (ie, cost-reduced) their hardware to eventually make money.
It's just like a P3, but it has half the amount of cache. So it probably is a P3 exactly, but the cache was probably modular to begin with, so they could just disable a defective half.
What were the other differences between a P3 and the comparable Celeron? Could they just all have been produced on the same assembly line, with features disabled to make a P3 into a Celeron?
> No proven model for network games? Surely Sony doesn't continue to produce expansions and sequels to Everquest just so they can lose more money.
You may have a point. But on the other hand, it will be interesting to see what happens to this market when it gets flooded with online, fee-based games.
> If memory serves me I believe the video settings for hi-def are in the Xbox dashboard and not in individual games which would lead me to believe that all games support it. I'm probably wrong on that though.
You are wrong. There is only one Xbox game that supports 1080i: Dragon's Lair 3D. There are a couple of games that support 720p (though not all HDTVs support 720p). There's a chart of this support at a gaming site somewhere.
> [hard drive comments]
Saving a game to HD instead of a memory card is a matter of convenience; it's not really a new ability. Faster loading time? Initial load speed is still limited to optical disk speed, since games can't install onto the HD. How many games actually make use of the HD to speed loading? And patches? Do you want your games to have patches? Add-ons can always come from an add-on optical disk.
> [Nintendo/Sony] costs are less due to missing features when compared to the Xbox.
That's exactly my point. They know how to make a product that's profitable. Microsoft doesn't.
The PS2 has an integrated DVD drive (ie, it's not a standard 5.25"-bay-size peripheral inside there like the Xbox has), and the expansion slot is the size of a 1" tall 3.5" hard drive.
Again, it's never a static picture.
I'm sure the first consoles were sold at a loss. Lots of engineering is devoted to cost reductions, and eventually the consoles are sold at a profit. Then there's a price-break, and the cycle may repeat a few times.
You don't know very far.
The CPU and graphics chip for the PS2 were initially produced in 0.25-micron, and were 240 mm^2 and 280 mm^2 initially. In case you don't know, this is f&cking HUGE! Read up on VLSI technology.
In short, $450 was probably pretty close to the mark, if not conservative; but it is irrelevant now, thanks to the ever advancing semiconductor progress train.
> How come Sony and Nintendo didn't put a 10/100 ethernet controller onboard?
Because there's no proven business model for making money off network games? And what percentage of the market has broadband?
> How come Sony and Nintendo don't support real Hi-Def?
How many Xbox games support real Hi-Def? What percentage of the market has a real HDTV?
> How come Sony and Nintendo don't have a built-in hard drive?
What concrete benefit has the hard drive provided to Xbox? What percentage of the market is interested in this benefit?
Here's a question to consider:
Why are Sony and Nintendo (currently) profiting from their hardware sales, while Microsoft is losing lots of money?
Market penetration of the console is only part of the picture that potential developers look at, I would hope. Actual game sales figures are another important factor. Thus, even if Microsoft ships an Xbox to every home in America, it won't help them much if everyone puts the box in the closet and buys no games. Sure, it may give them some leverage, but not nearly so much as strong game sales figures would.
The original line probably had something about an axe. Anyone know it?
Anyone know what the actual line is?
So that's one small step for MAN... ?
Mod parent up! This is excellent material:
http://www.natcap.org
Ever heard of the one-child policy?
> take a course in statistics... minimum sample size needed is 30 [...]
There's also a difference between statistics and data points. The previous poster didn't make any statistical comments; he was just providing his data points.
I've done a little research, and I think I've figured out the aiming:
The phase delay from each micro-antenna must be electronically adjustable. By carefully tuning the delay from each element, you can effectively aim the antenna.
This makes sense if you imagine the wave fronts as they intersect with the antenna. By carefully choosing what "points in time" are added up from each element, you can put together the wave fronts coming from the direction of your choice.
I still haven't figured out the polarization issue, but I imagine that they effectively have two antennas at right angles to each other, and that by doing some trignometric calculations on the two signals, they can figure out the original cross-polarized signals.
I know the very basics of a phased-array antenna: you've got a bunch of little antennas, and you combine their signals by matching the phase delay from the feed line for each one.
But how do you aim them in the case of a mobile unit like this? The only other phased-array DBS antenna I've seen requires you to point the whole thing.
Also, DirecWay uses linear polarization (vs. circular polarization for DBS). How is that handled? (ie, you've got to deal with not only azimuth and elevation, but also skew angle.)