Bullshit. Dvorak was spouting off in the mid 1990s about how nobody would ever need anything more than a 200 MHz Pentium. It wasn't true then, it's not true now.
The NES to SNES upgrade was actually relatively minor. You got more colors and bigger sprites, but the CPU was still weak, so the games were the same, but prettier. The SNES -> N64 transition was huge. It was the first console that could do 3D properly. Mario 64 changed platformers completely, and would not have been possible on any previous console. FPS as a genre wasn't really feasible until then either. The PS2 was the first console that had the horsepower to have complex environments, because the N64 and PS could not push enough polygons to do more than very simplistic environments.
This generation is potentially as interesting as the N64 one. The new consoles have an order of magnitude more power than the previous gen, and more importantly, they have a lot of power that's independent of the graphics pipeline. Wheras the main CPU in the PS2 spends much of its time crunching geometry to feed the rasterizer, the geometry processor in the RSX frees the Cell in the PS3 from much of that. Wheras previous consoles had to squeeze in AI and physics into a small slice of time between handling graphics code, the current batch can spend a lot of main CPU time on those things.
Gears of War is really a prime example. Even if you toned down the graphics, such a game could not be done on previous-gen systems. They don't have the horsepower to do either the physics, nor the level complexity (battlefields strewn with junk that serves as cover).
Why should the high volumes of the PS2 scare Sony? It's exactly what happened with the PS1 when the PS2 came out. Sony overlaps their consoles in order to compensate for the fact that they initially take a loss on new ones. The PS2 has a good year in it making Sony money while the PS3 ramps up, while the GC and Xbox are quite dead by now (with TP being the GC's last hurrah).
And you say "processing power" as if it doesn't matter. The SNES was just an NES with more processing power. The PS2 was just a PS1 with more processing power, etc. The new consoles have a *lot* more processing power than the previous gen. Clever developers will use this for more immersive environments, more realistic AI, etc. For example, the 360 and PS3 are really the first consoles that have the horsepower to do realistic physics, allowing for more complex gameplay. They're also the first ones to have the horsepower to do truely interactive environments (destructible geometry, etc). As game developers get more comfortable with these new architectures, you're going to see games that you just couldn't do on previous consoles, graphics aside.
What is so bloody innovative about Singularity? I keep hearing about it, but I can't help but thinking its just a Lisp Machine OS, except based on C#. Security through safe languages is a great idea, it really is. It's just not new --- there has already been more than one production implementation of the concept.
Autoboxing is completely different. Autoboxing isn't an optimization, its a hack*. What I'm talking about is general data structure optimization. The difference is apparent from the following pseudo-code:
function foo()
Vector c = new Vector;
for(i = 0; i N; ++i)
c.Append(i);
return c.Sum();
With auto-boxing, the compiler boxes/unboxes 'i' on every trip through the loop (in C#, this is will involve a heap allocation). The performance is equivalent to a fully-OO representation, its just that the programmer doesn't have to insert the conversion code himself.
Powerful dynamic language runtimes (notably Lisp, Smalltalk, Self, Concurrent Aggregates, Cecil) take a different approach. They either use powerful data-flow analyses (in static compilers), or powerful runtime profiling (in JITs) to actually give you the performance of primitive types without exposing primitive types to the user. For example, type inference will tell the compiler that the concrete class of 'c' is Vector. Escape analysis will tell it that 'c' and 'i' do not escape (their address is not captured by some other function). This let's it perform two optimizations: turn 'c' and 'i' into a primitive vector and primitive integer, respectively, and stack-allocate these primitives.
Now of course these optimizations are not always applicable, so sometimes you will pay the full cost of OO. In practice, they're good enough that Self and Concurrent Aggregates were reaching 50%+ of the performance of C with early 1990s VM technology. In the SBCL Lisp compiler (again, early 1990s technology!), these algorithms are good enough that the code generator is actually the bottleneck for your program (dumb register allocation on x86, no code scheduling for RISCs), not the fact that every single value, right down to your loop counter, is *always* conceptually a full-blown object.
Your post makes no sense. I say something about the PowerMac, and you come back with something about displays? Look at the mid-range PowerMac. At $2500, it comes with a pair of 2.66 GHz Xeons, 1GB of FB-DIMM, a 7300GT, and a 250GB HDD.
Even online, those are $700 processors. A dual-channel Woodcrest motherboard will run you close to $450 (there are $350 ones, but those are only dual-channel). A pair of 512MB DDR2-667 FB-DIMMs will run you about $180. Throw in another $50 each for the 7300GT, the HDD, and the DVD burner, you're looking at $2200. That leaves you $300 for the PSU, case, cooling, keyboard, mouse, wifi card, and OS. If you cheap-out on all of those things, you might make it under $2500. If you put in parts comparable to what Apple does, you have no hope. The PowerMac's have a 1KW workstation-class power supply. 1KW is probably overkill, and from a total quality perspective a 600W-700W Seasonic is very comparable. Those will run you $150-$200. The PowerMac case is a good 20lbs of aluminum plate with a very sophisticated, microcontroller-controlled cooling system. If we ignore the whole styling issue, I consider a $100 Antec P180 a good alternative (I have both). Throw in another $100 for accessories, and $0 for Linux, you're at $2580 for a machine that's truely comparably-equipped. That's assuming you get free shipping on everything, and never hit sales tax on anything. In practice, you're looking at another $100+ in shipping charges.
That's close to $2700, and you still haven't factored in labor costs or the fact that the home-built machine has no warranty. Then there is the whole resale thing. A Mac keeps its value extremely well. My $2500 October-2005 PowerMac would still fetch $1800+ on EBay. My $2000 August-2005 home-built might fetch $600-$700 (if that). A 3-year old home-built PC is worth next to nothing, while a 3-year-old Mac might still fetch a third of its original price.
I use Visual Studio all the time. It's not terrible, but it was far better in 6.0 before they fucked up the UI beyond all recognition. And.NET might be the best if you think the only VMs in existence are JVM and CLR, but the.NET VM can't hold a candle to some others in many respects. Wake me up when the CLR does a fraction of the high-level optimizations something like Sun's Strongtalk or Self VMs do.
The CLR gets good performance by punting on the whole OO thing and making the programmer use non-objects like integers and structures when performance matters. Those other VMs get a good fraction of the performance of C (often 50-90%), all the while using an object model where everything is fully general. That means your loop counter is a full, boxed, heap-allocated object that automatically upgrades to an infinite-precision integer instead of overflowing, at least as far as the programmer knows. If the integer is used as something like a loop counter, the VM is smart enough to unbox it and stick it in a register. When the CLR can do that, let me know, and well talk about throwing the word "best" around.
Practically identical to PC hardware but much more costly, which is exactly why I stick to PCs, especially now that you can run OSX on just about any PC.The reason most software companies don't port games to mac is because the audience is so much smaller it's just not worth their time and money. Have you looked at Mac prices since the Intel transition? All across the lines, the machines are extremely price-competitive. The iMac and Macbook are within 10% of a comparable PC, and the PowerMac severely undercuts a comparable PC (even if you build it yourself!). Just about the only machine that still carries a premium is the Macbook Pro.
The $37m is really chump change compared to the over $1bn spent developing the PS3. Those charges showed up on the earnings reports over the last year or two, so by now most of the losses have been accounted for.
Their investors understand this, because its a very common business model for technology companies. Sony spent over $1bn developing the PS2 as well, and lost money on the first units, but made back multiples of that over the lifetime of the console. Also, any losses they take on the launch are going to be offset by holiday sales of the PS2 (which are still very strong). At this point, PS2 sales are basically free money, since the cost of development was long-ago paid off, and the actual hardware is so cheap to make.
A console doesn't necessarily sell just for the "great" games either. "Good" games in a wide variety of genres also sells consoles. The PS2 has enormous numbers of "good" games in lots of genres. These games aren't Zelda: TP good, but they're something that any fan of the genre might find it worthwhile to play. One of the biggest problems with Nintendo consoles over the last decade or so has been that while there are lots of great games, there are very few good games. In other words, you can probably find a couple of dozen great games for the system, but its really hard to find a dozen good games in your favorite genre.
If you look at the polls, these aren't the radicals. On many issues, the majority of the Christian population is buying into these fundamentalist, anti-scientific and anti-liberal stances.
What I'm incredibly impressed over is the fact that they're mobilizing other space craft in the area to look for the missing probe. The fact that NASA can get spacecraft designed for complete autonomy in extreme environments, and designed years apart by different groups at that, to cooperate with each other all while tens of millions of miles from the closest human, well, that's a pretty impressive bit of engineering.
Not necessarily social retribution, but rather on the fear of making it impossible to live in society productively. Its a concrete (I eat tomorrow if my society doesn't explode with people killing each other), versus an abstract (I may be punished by a divine being if I kill someone) motivation.
The thing that your missing is that the new consoles don't just allow eye candy. They allow you to do graphical effects you couldn't do before, and clever developers (and there are always clever developers), can use that to allow gameplay that wasn't practical before.
Consider polygon counts and shader effects. Yes, you can use those to take the same game and just make it better looking. You can also use them to make organic environments that you just couldn't do before. The current generation of consoles really don't have the horsepower to do, for example, convincing jungle levels. Based on what I've seen with Crysis, the next generation may well have that capability.
Also, consider physics and animation. In the current generation of games, the physics and animation is canned. They don't have the horsepower to do anything more realistic. The new generation of engines have the power to do things like proper bullet ricochets, destructible geometry, etc. To this day I find it ridiculous that you can fire a rocket at a wooden wall in a game, and the only thing that happens is it leaves a scorch mark!
Your little rant misses one thing completely: Sony gets a license fee for *every* game sold on the Playstation platform. That's about $3-$7 of pure profit for each of the 1.89bn games sold on the PS platforms over the last 10 years.
And your analysis of why the PS1 and PS2 were successful is completely off. The reason Sony took the market with the PS1 is because third parties were getting tired of Nintendo's heavy-handedness, and because Nintendo went with the cart format for the N64 (so they could keep their lucrative cart licensing business). Sony has always been very good at relying on their third parties to build the library, instead of trying to do it themselves. It goes all the way back to the PS1, where they broke with the tradition of having a console "mascot" (ie: strong first-party franchises) in favor of letting third parties build the major franchises.
As for Nintendo's "blue ocean" strategy --- I'll believe it when I see it. Letting go of the control they yield over their third parties goes very strongly against Nintendo's corporate culture. They're the company that gave us "quality, not quantity" on the N64, which actually meant "you'll play the 3 good games we release once a year and you'll damn well like it!"
In the end, Nintendo and Sony are different companies. Nintendo is a great software publisher --- its overriding focus on its core franchises plays to that role. They are not good console developers, for the simple reason is that their major concern is about what the console can do for their core franchises, and how it can help push their particular vision of gaming. Sony is almost the exact opposite. They have no particularly strong first-party franchises, and no singular vision of gaming. What they have been doing with the Playstation platform is making it a good substrate for third parties to develop software on.
I'm very sure the Wii will be successful. The question is whether it'll be competitive with the PS3 and 360. Nintendo has its core demographic, as does Sony. Sony's core demographic is a lot more profitable, and a lot larger, and hence a more attractive target for 3rd party game development houses. Over the last 10 years, Sony has sold about 1.9bn games for the PSX and PS2. Nintendo has sold a small fraction of that number for the N64 and GC (less than a fifth).
Without breaking out of its traditional demographic, the Wii isn't going to change that. Sure, it'll make Nintendo a tidy profit, but the third party support is going to be lacking, making it as uninteresting to gamers as the GC was.
Enjoying more mature games has nothing to do with craving violence. More mature games simply tend to have storylines that make it easier for mature gamers to be immersed in the game world. For example, "Mario & Luigi: Partners in Time" is mechanically very competent, and has innovative gameplay, but I found it unplayable because of its suger-coated story. What's the point of an RPG if you can't immerse yourself in it?
Going on about "pure simple gameplay" is a little ridiculous. People don't play games for "gameplay", they play them for entertainment. Gameplay is like the acting in a movie. The mechanics might be absolutely spot-on, but if the cinematography is off, or the writing isn't very good, audiences still won't be entertained. Even if the movie has great acting, great cinematography, and great writing, if it has a story people can't connect with, people still won't be entertained. To use an example, the mechanics of the show "Gilmore Girls" are very good, with good writing and witty dialogue, but most guys still won't be entertained by it, for obvious reasons.
Its very possible to make a game with a child-friendly story that appeals to mature people. Many of Miyamoto's games fall into this category. However, its very difficult, and very rare. It's the game-equivalent of "Harry Potter". There aren't a lot of Harry Potters out there, and a studio could never make a lot of money just making such movies (what'd they do in the years between HP movies?)
And the PS3 can play Blu-Ray movies? Woop-dee freakin' doo. I don't know about everybody else, but I'm going to buy a gaming console to play games. If I want to watch movies, I'll buy a movie player.
Everybody else is using those phones that are also gaming machines, calendars, portable video players, and cameras.
Nintendo gets ahead with the first unit. If they can't beat their GC numbers, Sony will be winning after the first couple of years, based on the shear volume of their game sales. About 5x as many games were sold for the PS2 as for the GC (well over 700m compared to ~150m as of early 2005), and the license fees on those games are pure profit. Sony made way more on the PS2 than Nintendo did on the GC, even though Sony took a $1bn loss initially, and Nintendo sold each unit at a profit.
I should point out that the average PS2 owner is in his late 20s. The average age of OPM (Official Playstation Magazine) subscribers is 27. The success of the console, at least in the first several years, is going to be determined by the 18-35 demographic. Both Microsoft and Sony priced their consoles pretty out of the range of the mid-teen market.
The PSX didn't have more power in any dimension. It just had more storage, which allowed for better, bigger games.
The N64 was more powerful in almost every way, save for the peak theoretical geometry output of the graphics unit (which wasn't even a fair comparison, because each of the N64's ~100k triangles was anti-aliased, shaded, textured, and filtered, while the PSX could never achieve its ~300k count with anything but flat-shaded, untextured triangles). The N64 was the first system to be a decent 3D system, as opposed to a great 2D system.
And it still lost because Nintendo sucks at working with 3rd parties, and Sony stomped on them with a huge library.
I think that we are approaching the point of diminishing returns in graphics.
Have you seen Gears of War? Like, in person? The gameplay leaves something to be desired, but a team with talent is going to be able to make some really impressive stuff with that engine. There is a lot of potential there for the power being tapped for increased immersion.
Remember the reason why people have been buying Sony consoles in preference to Nintendo ones for more than 10 years now. It asn't power (the N64 _demolished_ the PSX), but because of the games. Specifically, the PS2 had them, and the Nintendo consoles really only had first party titles and ports.
Yet they shipped the PS2 with standard ports (USB, IEEE1394) long before anybody else put standard ports in a console.
The PS3 is about the least proprietary of the three systems being released. It uses a standard media format, supports standard flash formats, uses USB and Bluetooth for the controller, standard 2.5" hard drives, etc. On the software side, it's capable of running a non-proprietary OS (Linux), and uses open, standard APIs (Collada and the Khronos Open* stack). On top of all that, Xenon is a pretty closed chip, while there is a ton of documentation about Cell on IBM's developerworks site, and the Linux/Cell port has been in the mainline kernel since 2.6.16.
Yeah, you can make fun of Sony for Memory Stick and Betamax, but it rings kind of hollow when you're talking about a market in which their primary competitor is using the most proprietary platform (Windows) still in common use. Even MemoryStick is a licensed format (Sandisk and Lexar make them). DirectX, SMB, NTFS, etc, are as proprietary as they come.
Not all hereditary variations are the result of evolution. Evolution tends to be applied to the long term changes in genetic makeup based on natural selection. Skin color is the result of changes in pigmentation levels based on vitamin B and D levels both from diet and sunlight, not the result of reproductive selection.
I suppose if you define evolution broadly, then you can define changes in skin tone, as well as characteristics such as height, etc, as evolutionary attributes. That's a broader definition of evolution than is usually seen in practice, though.
Slashdot Mirror
Bullshit. Dvorak was spouting off in the mid 1990s about how nobody would ever need anything more than a 200 MHz Pentium. It wasn't true then, it's not true now.
The NES to SNES upgrade was actually relatively minor. You got more colors and bigger sprites, but the CPU was still weak, so the games were the same, but prettier. The SNES -> N64 transition was huge. It was the first console that could do 3D properly. Mario 64 changed platformers completely, and would not have been possible on any previous console. FPS as a genre wasn't really feasible until then either. The PS2 was the first console that had the horsepower to have complex environments, because the N64 and PS could not push enough polygons to do more than very simplistic environments.
This generation is potentially as interesting as the N64 one. The new consoles have an order of magnitude more power than the previous gen, and more importantly, they have a lot of power that's independent of the graphics pipeline. Wheras the main CPU in the PS2 spends much of its time crunching geometry to feed the rasterizer, the geometry processor in the RSX frees the Cell in the PS3 from much of that. Wheras previous consoles had to squeeze in AI and physics into a small slice of time between handling graphics code, the current batch can spend a lot of main CPU time on those things.
Gears of War is really a prime example. Even if you toned down the graphics, such a game could not be done on previous-gen systems. They don't have the horsepower to do either the physics, nor the level complexity (battlefields strewn with junk that serves as cover).
Why should the high volumes of the PS2 scare Sony? It's exactly what happened with the PS1 when the PS2 came out. Sony overlaps their consoles in order to compensate for the fact that they initially take a loss on new ones. The PS2 has a good year in it making Sony money while the PS3 ramps up, while the GC and Xbox are quite dead by now (with TP being the GC's last hurrah).
And you say "processing power" as if it doesn't matter. The SNES was just an NES with more processing power. The PS2 was just a PS1 with more processing power, etc. The new consoles have a *lot* more processing power than the previous gen. Clever developers will use this for more immersive environments, more realistic AI, etc. For example, the 360 and PS3 are really the first consoles that have the horsepower to do realistic physics, allowing for more complex gameplay. They're also the first ones to have the horsepower to do truely interactive environments (destructible geometry, etc). As game developers get more comfortable with these new architectures, you're going to see games that you just couldn't do on previous consoles, graphics aside.
What is so bloody innovative about Singularity? I keep hearing about it, but I can't help but thinking its just a Lisp Machine OS, except based on C#. Security through safe languages is a great idea, it really is. It's just not new --- there has already been more than one production implementation of the concept.
Autoboxing is completely different. Autoboxing isn't an optimization, its a hack*. What I'm talking about is general data structure optimization. The difference is apparent from the following pseudo-code:
function foo()
Vector c = new Vector;
for(i = 0; i N; ++i)
c.Append(i);
return c.Sum();
With auto-boxing, the compiler boxes/unboxes 'i' on every trip through the loop (in C#, this is will involve a heap allocation). The performance is equivalent to a fully-OO representation, its just that the programmer doesn't have to insert the conversion code himself.
Powerful dynamic language runtimes (notably Lisp, Smalltalk, Self, Concurrent Aggregates, Cecil) take a different approach. They either use powerful data-flow analyses (in static compilers), or powerful runtime profiling (in JITs) to actually give you the performance of primitive types without exposing primitive types to the user. For example, type inference will tell the compiler that the concrete class of 'c' is Vector. Escape analysis will tell it that 'c' and 'i' do not escape (their address is not captured by some other function). This let's it perform two optimizations: turn 'c' and 'i' into a primitive vector and primitive integer, respectively, and stack-allocate these primitives.
Now of course these optimizations are not always applicable, so sometimes you will pay the full cost of OO. In practice, they're good enough that Self and Concurrent Aggregates were reaching 50%+ of the performance of C with early 1990s VM technology. In the SBCL Lisp compiler (again, early 1990s technology!), these algorithms are good enough that the code generator is actually the bottleneck for your program (dumb register allocation on x86, no code scheduling for RISCs), not the fact that every single value, right down to your loop counter, is *always* conceptually a full-blown object.
Your post makes no sense. I say something about the PowerMac, and you come back with something about displays? Look at the mid-range PowerMac. At $2500, it comes with a pair of 2.66 GHz Xeons, 1GB of FB-DIMM, a 7300GT, and a 250GB HDD.
Even online, those are $700 processors. A dual-channel Woodcrest motherboard will run you close to $450 (there are $350 ones, but those are only dual-channel). A pair of 512MB DDR2-667 FB-DIMMs will run you about $180. Throw in another $50 each for the 7300GT, the HDD, and the DVD burner, you're looking at $2200. That leaves you $300 for the PSU, case, cooling, keyboard, mouse, wifi card, and OS. If you cheap-out on all of those things, you might make it under $2500. If you put in parts comparable to what Apple does, you have no hope. The PowerMac's have a 1KW workstation-class power supply. 1KW is probably overkill, and from a total quality perspective a 600W-700W Seasonic is very comparable. Those will run you $150-$200. The PowerMac case is a good 20lbs of aluminum plate with a very sophisticated, microcontroller-controlled cooling system. If we ignore the whole styling issue, I consider a $100 Antec P180 a good alternative (I have both). Throw in another $100 for accessories, and $0 for Linux, you're at $2580 for a machine that's truely comparably-equipped. That's assuming you get free shipping on everything, and never hit sales tax on anything. In practice, you're looking at another $100+ in shipping charges.
That's close to $2700, and you still haven't factored in labor costs or the fact that the home-built machine has no warranty. Then there is the whole resale thing. A Mac keeps its value extremely well. My $2500 October-2005 PowerMac would still fetch $1800+ on EBay. My $2000 August-2005 home-built might fetch $600-$700 (if that). A 3-year old home-built PC is worth next to nothing, while a 3-year-old Mac might still fetch a third of its original price.
I use Visual Studio all the time. It's not terrible, but it was far better in 6.0 before they fucked up the UI beyond all recognition. And .NET might be the best if you think the only VMs in existence are JVM and CLR, but the .NET VM can't hold a candle to some others in many respects. Wake me up when the CLR does a fraction of the high-level optimizations something like Sun's Strongtalk or Self VMs do.
The CLR gets good performance by punting on the whole OO thing and making the programmer use non-objects like integers and structures when performance matters. Those other VMs get a good fraction of the performance of C (often 50-90%), all the while using an object model where everything is fully general. That means your loop counter is a full, boxed, heap-allocated object that automatically upgrades to an infinite-precision integer instead of overflowing, at least as far as the programmer knows. If the integer is used as something like a loop counter, the VM is smart enough to unbox it and stick it in a register. When the CLR can do that, let me know, and well talk about throwing the word "best" around.
The $37m is really chump change compared to the over $1bn spent developing the PS3. Those charges showed up on the earnings reports over the last year or two, so by now most of the losses have been accounted for.
Their investors understand this, because its a very common business model for technology companies. Sony spent over $1bn developing the PS2 as well, and lost money on the first units, but made back multiples of that over the lifetime of the console. Also, any losses they take on the launch are going to be offset by holiday sales of the PS2 (which are still very strong). At this point, PS2 sales are basically free money, since the cost of development was long-ago paid off, and the actual hardware is so cheap to make.
A console doesn't necessarily sell just for the "great" games either. "Good" games in a wide variety of genres also sells consoles. The PS2 has enormous numbers of "good" games in lots of genres. These games aren't Zelda: TP good, but they're something that any fan of the genre might find it worthwhile to play. One of the biggest problems with Nintendo consoles over the last decade or so has been that while there are lots of great games, there are very few good games. In other words, you can probably find a couple of dozen great games for the system, but its really hard to find a dozen good games in your favorite genre.
If you look at the polls, these aren't the radicals. On many issues, the majority of the Christian population is buying into these fundamentalist, anti-scientific and anti-liberal stances.
What I'm incredibly impressed over is the fact that they're mobilizing other space craft in the area to look for the missing probe. The fact that NASA can get spacecraft designed for complete autonomy in extreme environments, and designed years apart by different groups at that, to cooperate with each other all while tens of millions of miles from the closest human, well, that's a pretty impressive bit of engineering.
Trolltech does almost all the Qt development.
Not necessarily social retribution, but rather on the fear of making it impossible to live in society productively. Its a concrete (I eat tomorrow if my society doesn't explode with people killing each other), versus an abstract (I may be punished by a divine being if I kill someone) motivation.
The thing that your missing is that the new consoles don't just allow eye candy. They allow you to do graphical effects you couldn't do before, and clever developers (and there are always clever developers), can use that to allow gameplay that wasn't practical before.
Consider polygon counts and shader effects. Yes, you can use those to take the same game and just make it better looking. You can also use them to make organic environments that you just couldn't do before. The current generation of consoles really don't have the horsepower to do, for example, convincing jungle levels. Based on what I've seen with Crysis, the next generation may well have that capability.
Also, consider physics and animation. In the current generation of games, the physics and animation is canned. They don't have the horsepower to do anything more realistic. The new generation of engines have the power to do things like proper bullet ricochets, destructible geometry, etc. To this day I find it ridiculous that you can fire a rocket at a wooden wall in a game, and the only thing that happens is it leaves a scorch mark!
Your little rant misses one thing completely: Sony gets a license fee for *every* game sold on the Playstation platform. That's about $3-$7 of pure profit for each of the 1.89bn games sold on the PS platforms over the last 10 years.
And your analysis of why the PS1 and PS2 were successful is completely off. The reason Sony took the market with the PS1 is because third parties were getting tired of Nintendo's heavy-handedness, and because Nintendo went with the cart format for the N64 (so they could keep their lucrative cart licensing business). Sony has always been very good at relying on their third parties to build the library, instead of trying to do it themselves. It goes all the way back to the PS1, where they broke with the tradition of having a console "mascot" (ie: strong first-party franchises) in favor of letting third parties build the major franchises.
As for Nintendo's "blue ocean" strategy --- I'll believe it when I see it. Letting go of the control they yield over their third parties goes very strongly against Nintendo's corporate culture. They're the company that gave us "quality, not quantity" on the N64, which actually meant "you'll play the 3 good games we release once a year and you'll damn well like it!"
In the end, Nintendo and Sony are different companies. Nintendo is a great software publisher --- its overriding focus on its core franchises plays to that role. They are not good console developers, for the simple reason is that their major concern is about what the console can do for their core franchises, and how it can help push their particular vision of gaming. Sony is almost the exact opposite. They have no particularly strong first-party franchises, and no singular vision of gaming. What they have been doing with the Playstation platform is making it a good substrate for third parties to develop software on.
I'm very sure the Wii will be successful. The question is whether it'll be competitive with the PS3 and 360. Nintendo has its core demographic, as does Sony. Sony's core demographic is a lot more profitable, and a lot larger, and hence a more attractive target for 3rd party game development houses. Over the last 10 years, Sony has sold about 1.9bn games for the PSX and PS2. Nintendo has sold a small fraction of that number for the N64 and GC (less than a fifth).
Without breaking out of its traditional demographic, the Wii isn't going to change that. Sure, it'll make Nintendo a tidy profit, but the third party support is going to be lacking, making it as uninteresting to gamers as the GC was.
Enjoying more mature games has nothing to do with craving violence. More mature games simply tend to have storylines that make it easier for mature gamers to be immersed in the game world. For example, "Mario & Luigi: Partners in Time" is mechanically very competent, and has innovative gameplay, but I found it unplayable because of its suger-coated story. What's the point of an RPG if you can't immerse yourself in it?
Going on about "pure simple gameplay" is a little ridiculous. People don't play games for "gameplay", they play them for entertainment. Gameplay is like the acting in a movie. The mechanics might be absolutely spot-on, but if the cinematography is off, or the writing isn't very good, audiences still won't be entertained. Even if the movie has great acting, great cinematography, and great writing, if it has a story people can't connect with, people still won't be entertained. To use an example, the mechanics of the show "Gilmore Girls" are very good, with good writing and witty dialogue, but most guys still won't be entertained by it, for obvious reasons.
Its very possible to make a game with a child-friendly story that appeals to mature people. Many of Miyamoto's games fall into this category. However, its very difficult, and very rare. It's the game-equivalent of "Harry Potter". There aren't a lot of Harry Potters out there, and a studio could never make a lot of money just making such movies (what'd they do in the years between HP movies?)
And the PS3 can play Blu-Ray movies? Woop-dee freakin' doo. I don't know about everybody else, but I'm going to buy a gaming console to play games. If I want to watch movies, I'll buy a movie player.
Everybody else is using those phones that are also gaming machines, calendars, portable video players, and cameras.
Nintendo gets ahead with the first unit. If they can't beat their GC numbers, Sony will be winning after the first couple of years, based on the shear volume of their game sales. About 5x as many games were sold for the PS2 as for the GC (well over 700m compared to ~150m as of early 2005), and the license fees on those games are pure profit. Sony made way more on the PS2 than Nintendo did on the GC, even though Sony took a $1bn loss initially, and Nintendo sold each unit at a profit.
I should point out that the average PS2 owner is in his late 20s. The average age of OPM (Official Playstation Magazine) subscribers is 27. The success of the console, at least in the first several years, is going to be determined by the 18-35 demographic. Both Microsoft and Sony priced their consoles pretty out of the range of the mid-teen market.
The PSX didn't have more power in any dimension. It just had more storage, which allowed for better, bigger games.
The N64 was more powerful in almost every way, save for the peak theoretical geometry output of the graphics unit (which wasn't even a fair comparison, because each of the N64's ~100k triangles was anti-aliased, shaded, textured, and filtered, while the PSX could never achieve its ~300k count with anything but flat-shaded, untextured triangles). The N64 was the first system to be a decent 3D system, as opposed to a great 2D system.
And it still lost because Nintendo sucks at working with 3rd parties, and Sony stomped on them with a huge library.
I think that we are approaching the point of diminishing returns in graphics.
Have you seen Gears of War? Like, in person? The gameplay leaves something to be desired, but a team with talent is going to be able to make some really impressive stuff with that engine. There is a lot of potential there for the power being tapped for increased immersion.
It's the *potential* of games.
Remember the reason why people have been buying Sony consoles in preference to Nintendo ones for more than 10 years now. It asn't power (the N64 _demolished_ the PSX), but because of the games. Specifically, the PS2 had them, and the Nintendo consoles really only had first party titles and ports.
Yet they shipped the PS2 with standard ports (USB, IEEE1394) long before anybody else put standard ports in a console.
The PS3 is about the least proprietary of the three systems being released. It uses a standard media format, supports standard flash formats, uses USB and Bluetooth for the controller, standard 2.5" hard drives, etc. On the software side, it's capable of running a non-proprietary OS (Linux), and uses open, standard APIs (Collada and the Khronos Open* stack). On top of all that, Xenon is a pretty closed chip, while there is a ton of documentation about Cell on IBM's developerworks site, and the Linux/Cell port has been in the mainline kernel since 2.6.16.
Yeah, you can make fun of Sony for Memory Stick and Betamax, but it rings kind of hollow when you're talking about a market in which their primary competitor is using the most proprietary platform (Windows) still in common use. Even MemoryStick is a licensed format (Sandisk and Lexar make them). DirectX, SMB, NTFS, etc, are as proprietary as they come.
Not all hereditary variations are the result of evolution. Evolution tends to be applied to the long term changes in genetic makeup based on natural selection. Skin color is the result of changes in pigmentation levels based on vitamin B and D levels both from diet and sunlight, not the result of reproductive selection.
I suppose if you define evolution broadly, then you can define changes in skin tone, as well as characteristics such as height, etc, as evolutionary attributes. That's a broader definition of evolution than is usually seen in practice, though.