Panajev2001a
GAF's Pleasant Genius
This was for PSINEXT.com... 
.
The One Console World: a Nightmarishly Happy Dream ()
--------------------------------------------------------------------------------------------------------
The idea of a single standard for console gaming, is an idea that some wish to have, while others wish never to have endure such a "nightmare." Games are the central focus of this industry, yes the games, not the consoles. It is the games which provide a way for people to add some fun and entertainment to their lives.
It is a business that has been steadily growing and moves billions of dollars each year. Why does one have a vested interest in the future of the console business? Perhaps because you think it is a form of art -and games like Ico seems to justify this view for some- or because you want to be part of the industry some day, or because your savings account is banked on the success or failure of the video games business.
This is the reason you should care about how the video game industry is approaching the future generation of consoles. The primary intent of this article is to discuss two fundamental directions I see the industry moving, a turning point perhaps, both routes have good intentions which ground their contention, however in my opinion there is only way direction which the industry has ahead of it- unity.
For example, if you are interested in a CRT (Cathodic Ray Tube) Television Set, maybe even a 26-27'' HDTV and you browse your local Best Buy, your Target or even your Walmart you can see one thing: prices are low.
About two years ago, more or less, you could already find very good deals on name brands: a quality Samsung HDTV with a 4:3 27'' flat-screen could be yours for $600 and now if you shop around for brands like Sanyo you can find tons of other bargains and low prices. What has happened here? Why are the prices low while the technology keeps on improving? Competition.
Because of competition, businesses have to be alert and aggressive on technology, quality of the finished product, and price, as there is a huge market out there trying to find the best deal. If you are not willing to cater to those customers then your lesser known competitor will gain their business and potentially their trust for future purchases.
The current thought of those who wish the gaming industry to continue down the same path as has existed since the gaming industries proverbial inception, relies on the contention that without numerious entities providing competing consoles, the industry would fall victim to its own single threaded approach.
This approach, they contend, would mean yielding higher prices for software and hardware while slowing down to almost a screeching halt the technological evolution the market was propelled by.
Could we blame the people who follow this perspective of the gaming consoles' business? Can we negate how rising development costs are pushing conglomerates like Electronic Arts (EA) to expand and absorb smaller studios and to push for higher games' prices? Can we negate that if Sony / Microsoft / Nintendo had no other competitors in the market that each generation their platform would stagnate and offer only marginal technology improvements? I contend, that yes, we can.
This group, I feel has misunderstood the idea behind the one console world: a single platform does not mean a single Independent Hardware Vendor (IHV) updating the platform's performance and features slowly, releasing buggy devices while raising prices. Of course who can blame them for their fear, perhaps a need only look at one of the console competitors; Microsoft. The Windows operating system during the 9.x period (Windows 95 -> Windows 98 -> Windows ME with all the various steps in between).
The terrible miss-stepings of Microsoft nearly sunk their operating system, for were it not for a consumer based society and good marketing they would most assuredly be extinct. Do their operating systems not suffer the same fate as those who believe in the tradition console market world? Of course, however the philosophy of a single console platform multiple vendor model will succeed and I'll tell you why.
Since we started mentioning Microsoft and their PC Operative Systems (OS's), let's keep the discussion there and let's look at the big winners in the field and how it relates to Computer Graphics and invariably the next generation of consoles.
For starters the two most famous IHV's in the PC space, ATI and nVIDIA. Since the year DirectX and its child Direct3D came into the picture, the PC space has been what people have looked too with the hope to unify the consoles' manufacturers for the good of the consumer, and ultimatly, I contend, for the survival of the gaming business.
Fundamentally the idea behind DirectX was to provide a high level API (Application Programming Interface) that would be the common ground for developers, putting to rest the DOS days of adding the explicit support for all the video cards, sound cards, etc. into to make their games. Thus we can see how we have had a single platform (from a programtic perspective) world in the PC space for almost a decade or more and it is time to check how things are doing.
Did technology succumb to the lack of competition that surely, the traditionalists would contend, a single platform future would bring? When PlayStation 2 launched in early 2000, nVIDIA was launching their GeForce 2 line -the code name: NV15. When PlayStation 3 is released (either late 2005 or in the year 2006) we will be able to enjoy a customized version of their GeForce 7 series GPU code name: NV50, optimized to work with the CELL processor acting as member CPU of the system.
The GeForce 2 did not have Vertex Shaders, but a Hardwired Transformation & Lighting unit and what could be called Pixel Shaders 0.x, powered by the nVIDIA's Register Combiners which were basically able to perform a relatively small set of mathematical operations over a given pixel the GPU was rendering. The Register Combiners are said by many, to have been one of the basis for the first generation of Pixel Shaders, introduced along with the Vertex Shaders in the DirectX world with DirectX 8.0.
The jump made from DirectX 7 to DirectX 9 and along with it, Shader Model 3.0 was a huge jump both in features and in processing speed: GPU's have gained in flexibility and usefulness even in General Purpose tasks while massively increasing in complexity, transistor count, clock speed and parallel processing performance and efficiency and yet the programatic interface has remaind roughly the same.
Why do our latest Radeon x800 and GeForce 6800 GPUs then lack the massive amount of games showing in full what they can achieve if properly utilized? This question will be presented to us in the attempt to show that, effectively, the common platform provided by the Win32 API + DirectX on the PC side of things effectively killed or seriously hurt competitiveness and hardware innovation.
The answer is obvious and we clarify "on the PC side of things": games are developed with the lowest common denominator in terms of hardware configuration and although the developer may try to scale up the rendering quality for the owners of the fast and new GPUs the result is not comparable to games that start their development with those high-end GPUs in mind.
It is true that in the PC development world, to search for a wide user-base, means taking into account the vast number of users who have old GPUs or low-end GPUs. This does not make the jump from NV1X to NV4X less astonishing as the issue we just mentioned. If we look at Half-Life 2 running on a GeForce 6800 GPU compared to Half-Life 1 running on a GeForce 2 we still see an amazing leap even though on each GPU you can probably do better if you designed the game around its strengths while minimizing its weaknesses, none-the-less, there is still a dramatic difference.
If we admit the "lowest common denominator" problem as a real problem for which traditionalist will contend will hurt the console business, if approached, then what is the answer? The answer is simpler than it appears: hardcore users. Hardcore gamers / users care about who is the current champion of GPUs, users are influenced in their purchase of a low-end GPU based on what GPU maker was able to achieve the performance crown on the most popular games and applications of the period.
The good image of the performance leader reflects on the low-end GPUs that are sold in such high quantities. Sadly, we all remember how this was used to scam people at times... do you still remember the GeForce 4 MX and the Radeon 9000? The newest GPUs still do not get to really shine on PC's because of the aforementioned issues. Furthermore it is important to remember just how fragmented, and open the PC market is, this fragmentaiton, would certainly not be nearly as realized in the single threaded console world scenario.
The reason being is that PC users like how versitile PC's are, they like ability to upgrade them frequently and customize them; installing new peripherals coming from a huge multitude of vendors. Conversley, console users and the average user of a Consumer Electronics (CE) device have accepted the closed box nature of those devices and the longer life-cycles enjoying the higher stability, better utilization of the components by the applications that will run on the device, and greater ease of use, coupled with a relatively low price-tag. Consider the console; insert the disk in the console and in a few seconds you are playing your entertainment content: no long installation procedure needed, no huge amount of patches to download, no worries about updating your graphics card drivers or having to troubleshoot the game.
This closed environment of CE devices, and the console market in general provide develoeprs with a stable and fixed set of specifications before a machine is launched which is not changed until the console goes end of life. This facilitates the ability to have a good software line-up ready for when the machine hits the stores and to better utilize the hardware over the span of thier life.
Hardware manufacturers have traditionally pushed the consoles' performance and technology ahead of the contemporary PC because they plan to loose money on hardware production in the first year of the consoles' life, waiting for the technology to advance and allow the chip-set to be re-worked (typically tigher intergration of curcuits) to cut costs: Sony pursued this strategy very aggressively with orginal PlayStation as well as the PlayStation 2. The losses they were taking on each unit sold at launch in Japan and in the U.S. were relatively high in both instances, that allowed the console to maintain a very attractive combination of power+features+price.
Sony pursued two strategies to bring the platform to profitability. The first one was linked with the R&D for their next generation PlayStation and involved development of new manufacturing processes with partners such as Toshiba and later IBM combined with new factories: they proceeded to shrink the Emotion Engine CPU and the Graphics Synthesizer GPU into smaller and smaller chips -which means more chips produced per silicon wafer, reducing the cost of each chip- to the point that they were able to make a single IC (Integrated Circuit) measuring only about 86 mm2 containing both CPU and GPU called "EE+GS@90 nm" while also combining other chips used in PlayStation 2 together into other IC's.
This allowed both their Nagasaki #1 and their Oita #1 (latter co-owned with Toshiba) manufacturing plants to both produce the combined CPU+GPU chip, effectively doubling their monthly output and further reducing manufacturing costs. When you build a manufacturing plant, invest money, purchasing or developing silicon wafers and all the technologies you need to use to manufacture the final product, the "manufacturing cost" of each product varies.
If we want to simplify the equation and forget for now the percentage of profit the company wants to make on the product, on those R&D costs and on the volume of production, or on how many chips you divide those amortized R&D costs over, the more chips your plant can produce and you can sell to customers, and consequently the less each chip (unit) is going to cost you to produce.
Closed box concepts like game consoles and open platforms like PC's are not exclusive concepts. There will be a more typical closed console and a more open PC -whether a CELL WorkStation or a regular PC, the two will co-exist, each one having functions/purposes that the other does not have, we have seen this already in the PlayStation 2 "optimized" Linux release -the Linux kit- being offered for PlayStation 2.
It seems we have covered and refuted pretty much a lot of the criticism against the vision of a one game console world, but we forgot one thing: we have forgotten to look at the arguments people generally presents in favor of a one game console world. These arguments can be classified under two categories for the most part: game making issues and users' convenience.
It is no joke that each generation of consoles game development costs have risen, that you need more and more people to program a game, to create content for it and to satisfy the growing demand of modern gamers for quality and polish: Xbox 2 and PlayStation 3 will be much more powerful than their predecessors, but it will be challenge for some developers to produce the visual proof that is expected of this jump.
Say for example that you have computers that can render at 60 frames per second, with graphics of the same level as Toy Story 2 or Finding Nemo, if well programmed: how good would it be if you only have the staff necessary to make a nice game that uses the appropriate quality game content (background detail, characters' models detail, special effects, etc.) in about 10 years or have to spend $40-80 Million developing each game? Not very, however this is the future developers are facing.
With the existing reality of two or three hardware manufacturers and the market split in chunks and you have to face the choice, as a publisher of either sticking to one platform and renounce the users who purchase the other platforms or you spend money supporting all three platforms which means having dedicated teams for each console, and / or rely on multi-platform Middleware that is targeted at the currently leading platform and is not really very well optimized for any system.
We have seen during this current console generation that the effective performance of each of the different consoles was on a very similar level, yet each platform was incompatible with software from the other platforms and was developed with very different strengths, weaknesses, and features. Games need to under go quite a laborious process to maintain the same quality when released for each platform and this is not an in-expensive proposition for games publishers.
The more systems a developer supports with unique titles, the more you have to spend, we have seen that less and less, developers are willing to stay exclusive to any platform- multi-platform releases are becoming more and more prevalent- unless the market leader has such an advantage over the competition as PlayStation 2 does, but even then the exclusivity period is not limit-less and it is often only a 4-6 months exclusivity window.
Users on their part have to purchase up to three platforms if they want to play all the exclusive games that first party developers and some third party developers (often with some form of incentive) have published. GT4 and Metal Gear Solid 3: Snake Eater on PlayStation 2, Halo 2 and Ninja Gaiden on Xbox, Super Mario Sunshine and Super Smash Brothers: Melee and Zelda: The Wind Waker on GameCube are all good examples of such. Overall this hurts consumers and third party developers alike.
The consumers who want all consoles, having to spend hundreds of dollars on each console, have less money to spend on games. When the market has three game consoles with relatively similar user bases there is a risk for multi-platform developers if many of those users have more than one console: if a person owns three different consoles by three different makers it is difficult / unlikely for said uers to buy the same game for all three consoles. Publishers would like to target a common platform with a user base of 150+ Million consoles owners more than supporting three platforms with about 50 Million "owners" each.
Even if everyone of those owners had only 1 of the three consoles, writing multi-platform software is still more expensive than developing for a single platform with a giant user-base, hence my contention of a single platform multiple vendor market. Programmaticly a single platform with multiple IHV's is still simpler than having multiple platforms, each with its own set of API's and very different design philosophies.
Consider a game designed for a DirectX 9.0c compatible GPU and it will run on all the DirectX 9.0c cards which expose all the required DirectX 9.0c features.
The speed the game runs at might change, but the game will run and most of the time. The optimization of the Shader code-path for each specific GPU can be optimized to handle a profile in the High Level Shading Language (HLSL)-to-DirectX ASM-like level which can surely be provided by the IHV as it is the case for nVIDIA's Cg shading language and compiler.
As you can see the future direction of the gaming industry into a single platform model is not all the doom and gloom. I hope you enjoyed the ride and you will be so patient and curious to read my future editorials: I hope to be welcoming you all in a One Console World one day !
Panajev
.The One Console World: a Nightmarishly Happy Dream (
)--------------------------------------------------------------------------------------------------------
The idea of a single standard for console gaming, is an idea that some wish to have, while others wish never to have endure such a "nightmare." Games are the central focus of this industry, yes the games, not the consoles. It is the games which provide a way for people to add some fun and entertainment to their lives.
It is a business that has been steadily growing and moves billions of dollars each year. Why does one have a vested interest in the future of the console business? Perhaps because you think it is a form of art -and games like Ico seems to justify this view for some- or because you want to be part of the industry some day, or because your savings account is banked on the success or failure of the video games business.
This is the reason you should care about how the video game industry is approaching the future generation of consoles. The primary intent of this article is to discuss two fundamental directions I see the industry moving, a turning point perhaps, both routes have good intentions which ground their contention, however in my opinion there is only way direction which the industry has ahead of it- unity.
For example, if you are interested in a CRT (Cathodic Ray Tube) Television Set, maybe even a 26-27'' HDTV and you browse your local Best Buy, your Target or even your Walmart you can see one thing: prices are low.
About two years ago, more or less, you could already find very good deals on name brands: a quality Samsung HDTV with a 4:3 27'' flat-screen could be yours for $600 and now if you shop around for brands like Sanyo you can find tons of other bargains and low prices. What has happened here? Why are the prices low while the technology keeps on improving? Competition.
Because of competition, businesses have to be alert and aggressive on technology, quality of the finished product, and price, as there is a huge market out there trying to find the best deal. If you are not willing to cater to those customers then your lesser known competitor will gain their business and potentially their trust for future purchases.
The current thought of those who wish the gaming industry to continue down the same path as has existed since the gaming industries proverbial inception, relies on the contention that without numerious entities providing competing consoles, the industry would fall victim to its own single threaded approach.
This approach, they contend, would mean yielding higher prices for software and hardware while slowing down to almost a screeching halt the technological evolution the market was propelled by.
Could we blame the people who follow this perspective of the gaming consoles' business? Can we negate how rising development costs are pushing conglomerates like Electronic Arts (EA) to expand and absorb smaller studios and to push for higher games' prices? Can we negate that if Sony / Microsoft / Nintendo had no other competitors in the market that each generation their platform would stagnate and offer only marginal technology improvements? I contend, that yes, we can.
This group, I feel has misunderstood the idea behind the one console world: a single platform does not mean a single Independent Hardware Vendor (IHV) updating the platform's performance and features slowly, releasing buggy devices while raising prices. Of course who can blame them for their fear, perhaps a need only look at one of the console competitors; Microsoft. The Windows operating system during the 9.x period (Windows 95 -> Windows 98 -> Windows ME with all the various steps in between).
The terrible miss-stepings of Microsoft nearly sunk their operating system, for were it not for a consumer based society and good marketing they would most assuredly be extinct. Do their operating systems not suffer the same fate as those who believe in the tradition console market world? Of course, however the philosophy of a single console platform multiple vendor model will succeed and I'll tell you why.
Since we started mentioning Microsoft and their PC Operative Systems (OS's), let's keep the discussion there and let's look at the big winners in the field and how it relates to Computer Graphics and invariably the next generation of consoles.
For starters the two most famous IHV's in the PC space, ATI and nVIDIA. Since the year DirectX and its child Direct3D came into the picture, the PC space has been what people have looked too with the hope to unify the consoles' manufacturers for the good of the consumer, and ultimatly, I contend, for the survival of the gaming business.
Fundamentally the idea behind DirectX was to provide a high level API (Application Programming Interface) that would be the common ground for developers, putting to rest the DOS days of adding the explicit support for all the video cards, sound cards, etc. into to make their games. Thus we can see how we have had a single platform (from a programtic perspective) world in the PC space for almost a decade or more and it is time to check how things are doing.
Did technology succumb to the lack of competition that surely, the traditionalists would contend, a single platform future would bring? When PlayStation 2 launched in early 2000, nVIDIA was launching their GeForce 2 line -the code name: NV15. When PlayStation 3 is released (either late 2005 or in the year 2006) we will be able to enjoy a customized version of their GeForce 7 series GPU code name: NV50, optimized to work with the CELL processor acting as member CPU of the system.
The GeForce 2 did not have Vertex Shaders, but a Hardwired Transformation & Lighting unit and what could be called Pixel Shaders 0.x, powered by the nVIDIA's Register Combiners which were basically able to perform a relatively small set of mathematical operations over a given pixel the GPU was rendering. The Register Combiners are said by many, to have been one of the basis for the first generation of Pixel Shaders, introduced along with the Vertex Shaders in the DirectX world with DirectX 8.0.
The jump made from DirectX 7 to DirectX 9 and along with it, Shader Model 3.0 was a huge jump both in features and in processing speed: GPU's have gained in flexibility and usefulness even in General Purpose tasks while massively increasing in complexity, transistor count, clock speed and parallel processing performance and efficiency and yet the programatic interface has remaind roughly the same.
Why do our latest Radeon x800 and GeForce 6800 GPUs then lack the massive amount of games showing in full what they can achieve if properly utilized? This question will be presented to us in the attempt to show that, effectively, the common platform provided by the Win32 API + DirectX on the PC side of things effectively killed or seriously hurt competitiveness and hardware innovation.
The answer is obvious and we clarify "on the PC side of things": games are developed with the lowest common denominator in terms of hardware configuration and although the developer may try to scale up the rendering quality for the owners of the fast and new GPUs the result is not comparable to games that start their development with those high-end GPUs in mind.
It is true that in the PC development world, to search for a wide user-base, means taking into account the vast number of users who have old GPUs or low-end GPUs. This does not make the jump from NV1X to NV4X less astonishing as the issue we just mentioned. If we look at Half-Life 2 running on a GeForce 6800 GPU compared to Half-Life 1 running on a GeForce 2 we still see an amazing leap even though on each GPU you can probably do better if you designed the game around its strengths while minimizing its weaknesses, none-the-less, there is still a dramatic difference.
If we admit the "lowest common denominator" problem as a real problem for which traditionalist will contend will hurt the console business, if approached, then what is the answer? The answer is simpler than it appears: hardcore users. Hardcore gamers / users care about who is the current champion of GPUs, users are influenced in their purchase of a low-end GPU based on what GPU maker was able to achieve the performance crown on the most popular games and applications of the period.
The good image of the performance leader reflects on the low-end GPUs that are sold in such high quantities. Sadly, we all remember how this was used to scam people at times... do you still remember the GeForce 4 MX and the Radeon 9000? The newest GPUs still do not get to really shine on PC's because of the aforementioned issues. Furthermore it is important to remember just how fragmented, and open the PC market is, this fragmentaiton, would certainly not be nearly as realized in the single threaded console world scenario.
The reason being is that PC users like how versitile PC's are, they like ability to upgrade them frequently and customize them; installing new peripherals coming from a huge multitude of vendors. Conversley, console users and the average user of a Consumer Electronics (CE) device have accepted the closed box nature of those devices and the longer life-cycles enjoying the higher stability, better utilization of the components by the applications that will run on the device, and greater ease of use, coupled with a relatively low price-tag. Consider the console; insert the disk in the console and in a few seconds you are playing your entertainment content: no long installation procedure needed, no huge amount of patches to download, no worries about updating your graphics card drivers or having to troubleshoot the game.
This closed environment of CE devices, and the console market in general provide develoeprs with a stable and fixed set of specifications before a machine is launched which is not changed until the console goes end of life. This facilitates the ability to have a good software line-up ready for when the machine hits the stores and to better utilize the hardware over the span of thier life.
Hardware manufacturers have traditionally pushed the consoles' performance and technology ahead of the contemporary PC because they plan to loose money on hardware production in the first year of the consoles' life, waiting for the technology to advance and allow the chip-set to be re-worked (typically tigher intergration of curcuits) to cut costs: Sony pursued this strategy very aggressively with orginal PlayStation as well as the PlayStation 2. The losses they were taking on each unit sold at launch in Japan and in the U.S. were relatively high in both instances, that allowed the console to maintain a very attractive combination of power+features+price.
Sony pursued two strategies to bring the platform to profitability. The first one was linked with the R&D for their next generation PlayStation and involved development of new manufacturing processes with partners such as Toshiba and later IBM combined with new factories: they proceeded to shrink the Emotion Engine CPU and the Graphics Synthesizer GPU into smaller and smaller chips -which means more chips produced per silicon wafer, reducing the cost of each chip- to the point that they were able to make a single IC (Integrated Circuit) measuring only about 86 mm2 containing both CPU and GPU called "EE+GS@90 nm" while also combining other chips used in PlayStation 2 together into other IC's.
This allowed both their Nagasaki #1 and their Oita #1 (latter co-owned with Toshiba) manufacturing plants to both produce the combined CPU+GPU chip, effectively doubling their monthly output and further reducing manufacturing costs. When you build a manufacturing plant, invest money, purchasing or developing silicon wafers and all the technologies you need to use to manufacture the final product, the "manufacturing cost" of each product varies.
If we want to simplify the equation and forget for now the percentage of profit the company wants to make on the product, on those R&D costs and on the volume of production, or on how many chips you divide those amortized R&D costs over, the more chips your plant can produce and you can sell to customers, and consequently the less each chip (unit) is going to cost you to produce.
Closed box concepts like game consoles and open platforms like PC's are not exclusive concepts. There will be a more typical closed console and a more open PC -whether a CELL WorkStation or a regular PC, the two will co-exist, each one having functions/purposes that the other does not have, we have seen this already in the PlayStation 2 "optimized" Linux release -the Linux kit- being offered for PlayStation 2.
It seems we have covered and refuted pretty much a lot of the criticism against the vision of a one game console world, but we forgot one thing: we have forgotten to look at the arguments people generally presents in favor of a one game console world. These arguments can be classified under two categories for the most part: game making issues and users' convenience.
It is no joke that each generation of consoles game development costs have risen, that you need more and more people to program a game, to create content for it and to satisfy the growing demand of modern gamers for quality and polish: Xbox 2 and PlayStation 3 will be much more powerful than their predecessors, but it will be challenge for some developers to produce the visual proof that is expected of this jump.
Say for example that you have computers that can render at 60 frames per second, with graphics of the same level as Toy Story 2 or Finding Nemo, if well programmed: how good would it be if you only have the staff necessary to make a nice game that uses the appropriate quality game content (background detail, characters' models detail, special effects, etc.) in about 10 years or have to spend $40-80 Million developing each game? Not very, however this is the future developers are facing.
With the existing reality of two or three hardware manufacturers and the market split in chunks and you have to face the choice, as a publisher of either sticking to one platform and renounce the users who purchase the other platforms or you spend money supporting all three platforms which means having dedicated teams for each console, and / or rely on multi-platform Middleware that is targeted at the currently leading platform and is not really very well optimized for any system.
We have seen during this current console generation that the effective performance of each of the different consoles was on a very similar level, yet each platform was incompatible with software from the other platforms and was developed with very different strengths, weaknesses, and features. Games need to under go quite a laborious process to maintain the same quality when released for each platform and this is not an in-expensive proposition for games publishers.
The more systems a developer supports with unique titles, the more you have to spend, we have seen that less and less, developers are willing to stay exclusive to any platform- multi-platform releases are becoming more and more prevalent- unless the market leader has such an advantage over the competition as PlayStation 2 does, but even then the exclusivity period is not limit-less and it is often only a 4-6 months exclusivity window.
Users on their part have to purchase up to three platforms if they want to play all the exclusive games that first party developers and some third party developers (often with some form of incentive) have published. GT4 and Metal Gear Solid 3: Snake Eater on PlayStation 2, Halo 2 and Ninja Gaiden on Xbox, Super Mario Sunshine and Super Smash Brothers: Melee and Zelda: The Wind Waker on GameCube are all good examples of such. Overall this hurts consumers and third party developers alike.
The consumers who want all consoles, having to spend hundreds of dollars on each console, have less money to spend on games. When the market has three game consoles with relatively similar user bases there is a risk for multi-platform developers if many of those users have more than one console: if a person owns three different consoles by three different makers it is difficult / unlikely for said uers to buy the same game for all three consoles. Publishers would like to target a common platform with a user base of 150+ Million consoles owners more than supporting three platforms with about 50 Million "owners" each.
Even if everyone of those owners had only 1 of the three consoles, writing multi-platform software is still more expensive than developing for a single platform with a giant user-base, hence my contention of a single platform multiple vendor market. Programmaticly a single platform with multiple IHV's is still simpler than having multiple platforms, each with its own set of API's and very different design philosophies.
Consider a game designed for a DirectX 9.0c compatible GPU and it will run on all the DirectX 9.0c cards which expose all the required DirectX 9.0c features.
The speed the game runs at might change, but the game will run and most of the time. The optimization of the Shader code-path for each specific GPU can be optimized to handle a profile in the High Level Shading Language (HLSL)-to-DirectX ASM-like level which can surely be provided by the IHV as it is the case for nVIDIA's Cg shading language and compiler.
As you can see the future direction of the gaming industry into a single platform model is not all the doom and gloom. I hope you enjoyed the ride and you will be so patient and curious to read my future editorials: I hope to be welcoming you all in a One Console World one day !
Panajev