EU should rule to totally scrap software patents. That would give MS no excuse.
The lobby groups in the EU aren't strong enough to get "full" software patents (i.e. no innovation required) but they are strong enough to prevent an anti-software-patent bill from passing. That's what they did after their last free patents bill got turned into one that strictly locks software patents down.
I'd say they should focus on PvP with each battlefield having deployment limits by point value in each instance. That way even if everyone plays SM they'd still be in fewer numbers in a battle than, say, the orcs. Something like a Dreadnaught would need a special promotion system (but mostly due to the fluff), space marines don't need a real limit.
Though perhaps it would work better if every player controlled a squad of characters so the imperial guard can get enough people together to fight.
An 80% solution today is often more desirable than a 100% solution tomorrow.
Yes but a console will last for longer than just a few weeks so it's important to evaluate both the current offerings and the expected future development. I'm not saying I expect the Wii to become better than the 360 but I don't think a snapshot with such big differences in the system lifetime is a meaningful way to predict how the systems will behave a year or two from now.
Let's say I want to procedurally generate an animal 5 years old. I can see the day where I start the simulation with a single cell along with properties of growth, and run it far enough where the animal is 5 years old.
Let's not forget that that would probably require a supercomputer and a few years worth of CPU time to calculate. Organisms are VERY complex systems, so complex indeed that we don't even understand them fully yet. An organism doesn't grow in a vacuum either so you'd have to simulate a limited environment as well. I don't think this kind of constructive simulation is feasible. While we can do an approximation, growth of a biological system is a very chaotic process and the smallest error could throw the result off completely. By the time we can model that accurately we can probably create AIs that have enough experience with life that they can just produce an entire game by themselves by doing all the steps a normal dev team goes through.
I don't see why not? The interior of a house uses pretty standard components -- wood, metal, plastic, glass, etc.
How do you describe a living room other than describing each object inside by itself? A couch is nothing like a DVD player. Furthermore, wouldn't breaking everything down to its parts be more complex than just describing its shape and how it looks? To properly generate the folds in a couch you'd have to know how the material behaves, how people use it, etc when you could just tell an artist to apply his knowledge of that. Humans know a LOT more about the environment they live in than any computer.
Procedural textures, MAY have infinite detail. But an infinite detail texture, to me at least, seems it would REQUIRE a procedural (fractal) texture.
Yes but there aren't many surfaces that could use a fractal so in a real situation that makes no difference. Games aren't meant to be textbook implementations of technology, they do only what they have to.
What would be really cool, if computers were fast enough to generate procedural textures in real-time, every frame, so the API would be extended to...
Texture MakeTexture( SurfaceProperties, DistanceToCamera )
We use MIP-mapping for that, we precompute a set of textures at different resolutions and apply them as appropriate. There is no reason to recompute the texture every frame.
So, with enough money and enough (brought) representatives, can I give rights to a gun to shoot people, but not the person that wields it? That is inane.
In a sufficiently corrupt government enough money can buy you ANY law, no matter how insane. It's called corruption for a reason.
If this is the true state of intellectual "property" law in this country, I suggest you change specializations before your soul starts dissolving away.
If ignorance were strength that might be appropriate but in the real world you should know your enemy.
Trade secrets don't apply since the document describing AACS is released to the public. Hard to claim something as secret if you've told everyone about it already.
They're lawyers, lawyers are expected to help even the most despicable criminals if they were hired by them*. After all everybody has the right to use the legal system but few have the understanding to do so. If lawyers would refuse to work with those they don't deem worthy they could easily exclude parts of the populace from getting fair treatment under the law. We assume that everyone's innocent until proven guilty so having the lawyers judge who is worthy of standing a chance in the court of law would give them too much power.
*=I know lawyers can decline some things but even a serial killer that's pretty much guaranteed to be guilty can get a lawyer. Hell, even Hussein got a lawyer.
In fact, the specific identification of the content could guide provision of related goods, services and community designed to maximize the consumer's enjoyment of the entertainment experience.
We'll figure out what he likes and spam him with "related offers".
At least for the Okami - Zelda comparison I'd say the sales numbers are deserved (though maybe they shouldn't be as extreme), Okami just isn't as good as Zelda. I know I found Zelda TP awfully easy but Okami is even easier. I haven't even emptied one stomach yet* in Okami about 30 hours in whereas I've died a few times in TP (mostly during the early game, towards the end you have so many hearts you practically have to fall asleep to die). Okami's puzzles are simpler and even when they aren't Issun will rub the solution into your face pretty quickly. There's lots of things you see that you have to come back to with more abilities to use and most of them only require using that ability in the standard way to retrieve a treasure or some luck. Okami's dungeons are often very short and very linear (though TP's dungeons have that latter fault, too).
*=For people who don't have Okami, a full stomach acts like an extra life.
But textures are easy to generate by hand using PhotoShop. It's meshes and animations that take all the time;
I don't know about animations (depends a lot on how many you need and how much cleanup your motion capture needs, if you're using that) but meshes take a LOT less time than the textures for them (though I think with today's highpoly to lowpoly bakes the time for the modeled and hand-painted parts of the texture is probably about equal). Textures aren't "easily created", that's completely ignoring all the work that goes into them (unless you use touched-up photos but those only work for flat surfaces before getting really complicated to do). Of course a texture for an object like e.g. a car or a character is going to be way too complex to handle with procedurals unless you use the n00b's approach (i.e. all detail in the mesh, just throwing generic procedurals at them). Any serious artist will hand paint a LOT of the surface of the object especially since some surfaces are neither simple nor random with their details (e.g. human skin is a very complex thing to paint if you consider all the color variations caused by veins, bones, etc running under it). Even cinematic characters have most if not all of their surfaces covered with hand-painted textures.
A problem with stuff like kkrieger is that that kind of generation is not meant for reducing development times, it's meant for compressing data. Procedural only has an advantage if it takes less time to mathematically describe a set of assets than to make them by hand. Very few things can be described mathematically in reasonable time and most objects in our lives are too complex to be handled reasonably well by a mathematical description. You can use procedurals to generate lots of similar things but games usually don't have that many of those (usually there's maybe 2-3 different appearances for a set). If you need lots of unique looking people but no specific features on any of them you can use a system that can derive them from a basic shape but few games bother with varying those. Anything that is too different from other things in the game (e.g. characters with unique designs) would need individual attention anyway and it may just be easier to take some basemesh, pull it into the right shape and start modelling the unique bits.
You can use procedurals to generate a city with no points of interest (or manually add those points but the rest will still be an uninteresting mass), to generate a crowd with no identical people or to generate the different trees in a forest but you cannot use them to e.g. model the interior of one house. Also you'd still need to make all components the system can use by hand, it can only combine them.
Also your "infinite details" bit seems more like a misconception to me, unless the surface is a fractal the artist will only describe it in so much detail and anything below that won't be in the game (e.g. if noone told the engine what concrete looks up close it'll look like a solid surface instead of having all the little details on it).
So in summary, procedurals are preferrable if you have a large number of objects that can be derived from a relatively simple rule but they're a waste of time for objects that differ wildly from each other.
We are still using the same 3D modeling tools, the same texturing tools, and the same map-building tools that we've been using since the release of Half-Life 1 8 years ago.
Well, if we ignore all the advances those programs made since 1997... The programs may carry the same names but the functionality has improved so much that your work speed is much, MUCH higher.
American politics has finally reached the point where a hyper-violent rabbit with a gun fetish who promises to nuke the Antarctic is a more attractive candidate than anyone running in the real world.
Slashdot Mirror
EU should rule to totally scrap software patents. That would give MS no excuse.
The lobby groups in the EU aren't strong enough to get "full" software patents (i.e. no innovation required) but they are strong enough to prevent an anti-software-patent bill from passing. That's what they did after their last free patents bill got turned into one that strictly locks software patents down.
I'd say they should focus on PvP with each battlefield having deployment limits by point value in each instance. That way even if everyone plays SM they'd still be in fewer numbers in a battle than, say, the orcs. Something like a Dreadnaught would need a special promotion system (but mostly due to the fluff), space marines don't need a real limit.
Though perhaps it would work better if every player controlled a squad of characters so the imperial guard can get enough people together to fight.
Well, THQ is the publisher for Relic who did Dawn of War.
An 80% solution today is often more desirable than a 100% solution tomorrow.
Yes but a console will last for longer than just a few weeks so it's important to evaluate both the current offerings and the expected future development. I'm not saying I expect the Wii to become better than the 360 but I don't think a snapshot with such big differences in the system lifetime is a meaningful way to predict how the systems will behave a year or two from now.
Let's say I want to procedurally generate an animal 5 years old. I can see the day where I start the simulation with a single cell along with properties of growth, and run it far enough where the animal is 5 years old.
Let's not forget that that would probably require a supercomputer and a few years worth of CPU time to calculate. Organisms are VERY complex systems, so complex indeed that we don't even understand them fully yet. An organism doesn't grow in a vacuum either so you'd have to simulate a limited environment as well. I don't think this kind of constructive simulation is feasible. While we can do an approximation, growth of a biological system is a very chaotic process and the smallest error could throw the result off completely. By the time we can model that accurately we can probably create AIs that have enough experience with life that they can just produce an entire game by themselves by doing all the steps a normal dev team goes through.
I don't see why not? The interior of a house uses pretty standard components -- wood, metal, plastic, glass, etc.
How do you describe a living room other than describing each object inside by itself? A couch is nothing like a DVD player. Furthermore, wouldn't breaking everything down to its parts be more complex than just describing its shape and how it looks? To properly generate the folds in a couch you'd have to know how the material behaves, how people use it, etc when you could just tell an artist to apply his knowledge of that. Humans know a LOT more about the environment they live in than any computer.
Procedural textures, MAY have infinite detail.
But an infinite detail texture, to me at least, seems it would REQUIRE a procedural (fractal) texture.
Yes but there aren't many surfaces that could use a fractal so in a real situation that makes no difference. Games aren't meant to be textbook implementations of technology, they do only what they have to.
What would be really cool, if computers were fast enough to generate procedural textures in real-time, every frame, so the API would be extended to...
Texture MakeTexture( SurfaceProperties, DistanceToCamera )
We use MIP-mapping for that, we precompute a set of textures at different resolutions and apply them as appropriate. There is no reason to recompute the texture every frame.
So, with enough money and enough (brought) representatives, can I give rights to a gun to shoot people, but not the person that wields it? That is inane.
In a sufficiently corrupt government enough money can buy you ANY law, no matter how insane. It's called corruption for a reason.
Why? The DMCA doesn't cover nor needs to cover theft.
In math, the rational is a subset of the real. In law, both sets are disjoint.
If this is the true state of intellectual "property" law in this country, I suggest you change specializations before your soul starts dissolving away.
If ignorance were strength that might be appropriate but in the real world you should know your enemy.
Yes but the DMCA supercedes Fair Use.
6) Chuck Norris
Trade secrets don't apply since the document describing AACS is released to the public. Hard to claim something as secret if you've told everyone about it already.
They're lawyers, lawyers are expected to help even the most despicable criminals if they were hired by them*. After all everybody has the right to use the legal system but few have the understanding to do so. If lawyers would refuse to work with those they don't deem worthy they could easily exclude parts of the populace from getting fair treatment under the law. We assume that everyone's innocent until proven guilty so having the lawyers judge who is worthy of standing a chance in the court of law would give them too much power.
*=I know lawyers can decline some things but even a serial killer that's pretty much guaranteed to be guilty can get a lawyer. Hell, even Hussein got a lawyer.
Work overtime (or look for a job that pays more) until you have enough money?
In fact, the specific identification of the content could guide provision of related goods, services and community designed to maximize the consumer's enjoyment of the entertainment experience.
We'll figure out what he likes and spam him with "related offers".
At least for the Okami - Zelda comparison I'd say the sales numbers are deserved (though maybe they shouldn't be as extreme), Okami just isn't as good as Zelda. I know I found Zelda TP awfully easy but Okami is even easier. I haven't even emptied one stomach yet* in Okami about 30 hours in whereas I've died a few times in TP (mostly during the early game, towards the end you have so many hearts you practically have to fall asleep to die). Okami's puzzles are simpler and even when they aren't Issun will rub the solution into your face pretty quickly. There's lots of things you see that you have to come back to with more abilities to use and most of them only require using that ability in the standard way to retrieve a treasure or some luck. Okami's dungeons are often very short and very linear (though TP's dungeons have that latter fault, too).
*=For people who don't have Okami, a full stomach acts like an extra life.
But textures are easy to generate by hand using PhotoShop. It's meshes and animations that take all the time;
I don't know about animations (depends a lot on how many you need and how much cleanup your motion capture needs, if you're using that) but meshes take a LOT less time than the textures for them (though I think with today's highpoly to lowpoly bakes the time for the modeled and hand-painted parts of the texture is probably about equal). Textures aren't "easily created", that's completely ignoring all the work that goes into them (unless you use touched-up photos but those only work for flat surfaces before getting really complicated to do). Of course a texture for an object like e.g. a car or a character is going to be way too complex to handle with procedurals unless you use the n00b's approach (i.e. all detail in the mesh, just throwing generic procedurals at them). Any serious artist will hand paint a LOT of the surface of the object especially since some surfaces are neither simple nor random with their details (e.g. human skin is a very complex thing to paint if you consider all the color variations caused by veins, bones, etc running under it). Even cinematic characters have most if not all of their surfaces covered with hand-painted textures.
A problem with stuff like kkrieger is that that kind of generation is not meant for reducing development times, it's meant for compressing data. Procedural only has an advantage if it takes less time to mathematically describe a set of assets than to make them by hand. Very few things can be described mathematically in reasonable time and most objects in our lives are too complex to be handled reasonably well by a mathematical description. You can use procedurals to generate lots of similar things but games usually don't have that many of those (usually there's maybe 2-3 different appearances for a set). If you need lots of unique looking people but no specific features on any of them you can use a system that can derive them from a basic shape but few games bother with varying those. Anything that is too different from other things in the game (e.g. characters with unique designs) would need individual attention anyway and it may just be easier to take some basemesh, pull it into the right shape and start modelling the unique bits.
You can use procedurals to generate a city with no points of interest (or manually add those points but the rest will still be an uninteresting mass), to generate a crowd with no identical people or to generate the different trees in a forest but you cannot use them to e.g. model the interior of one house. Also you'd still need to make all components the system can use by hand, it can only combine them.
Also your "infinite details" bit seems more like a misconception to me, unless the surface is a fractal the artist will only describe it in so much detail and anything below that won't be in the game (e.g. if noone told the engine what concrete looks up close it'll look like a solid surface instead of having all the little details on it).
So in summary, procedurals are preferrable if you have a large number of objects that can be derived from a relatively simple rule but they're a waste of time for objects that differ wildly from each other.
We are still using the same 3D modeling tools, the same texturing tools, and the same map-building tools that we've been using since the release of Half-Life 1 8 years ago.
Well, if we ignore all the advances those programs made since 1997... The programs may carry the same names but the functionality has improved so much that your work speed is much, MUCH higher.
Same place they get the money for starting their company.
That's cool but it would be cooler if it was called "Multiple Use Labor Element" instead.
To quote Gamespy's Sam & Max episode 4 review:
American politics has finally reached the point where a hyper-violent rabbit with a gun fetish who promises to nuke the Antarctic is a more attractive candidate than anyone running in the real world.
Yep, noone liked No One Lives Forever.
Even in Quake 4, people mostly look like they're plastic -- but Alyx looks as good as ever.
However the Combine in HL2 do look like action figures.
Discworld trolls believe they move backwards through time.