Why do we have both Mono and DotGNU implementations of Windows.Forms?
Why dedicate four senior Mono developers to solving something that the GNU project has already tackled? Won't the DotGNU version work with Mono? If that's the case why not just take the DotGNU version as a starting point and work from there?
Or is this some kind of licensing issue? It would not surprise me if the DotGNU implementation was lumbered with a full GPL, rather than an LGPL, thus making it unusable for anything appart from "free" software.
I think the big omission in GNUstep though is support for bindings. Last time I checked they didn't have Key Value Coding and Key Value Observation support in their NSObject implementation, which is required for bindings as well as more advanced data models. It would be really nice if they had this, since it would help porting 10.3 apps to Linux.
There is of course the old saying "the enemy of my enemy is my friend".
Bush doesn't understand that.
In Bush-logic if you don't agree with him you must therefore completely disagree, therefore you are the enemy. Unfortunately it seems that much of the country is this simplistic too.
In the immediate aftermath of 9/11 any country that expressed an opinion against the actions of the USA was villified, and some branded as terrorist states. Honestly, renaming "French fries" as "freedom fries" was the most rediculous thing I had ever heard and the general hatred of the French truly demonstrated mass idiocy. If people were more aware that the Statue of Liberty was a gift from France they would have been calling for it to be ripped down too.
Re:What was crazier to think at that time?
on
The Unknown Newton
·
· Score: 1
It's not even crazy these days to think that you can change lead into gold. Gold can be created from lead, with the help of some very expensive equipment and only in very very minute quantities. You just have to knock the right number of protons, neutrons, and electrons off of lead with a supercollider. Simple really.:-)
So robots disguised as humans is old hat. So what? We already knew that a new Battlestar Galactica series wasn't going to be original.
I felt though that it is a refreshing spin on the old series. One possibility that was never present in the original BSG was Cylon infiltration. That is now integral to the plot, and frankly I like it.
Now I did used to like the original series too, but I was 10 years old back then and the unsophisticated story line appealed to me. Now that I'm over 30 I want something a little bit more challenging.
As for the "shakey camera" effect, that was pioneered on Star Trek back in the '60s - Firefly just copied it.:-)
Snooker's a much harder game than pool, since the table is larger, the pockets smaller, and there's more balls on the table to start with. The robot on Horizon could actually play a real game, following the rules or snooker, and making intelligent shots with positioning.
From the article it sounds like this pool playing robot is pretty crummy right now. They hilight its strength as being able to pot the white in any pocket, but make it clear that it has trouble potting any coloured ball. Whether it can actually play a game isn't clear - I get the impression though that it's not programmed yet to put english on the ball.
Whilst this might technically be the first pool playing robot that's only because the British robot that came before it was never reprogrammed to play pool. Maybe the brits should resurrect their old creation and challenge these guys to a game.:-)
Maybe I'm being stupid, but I don't see how this work obseletes the Von Neumann architecture.
Sure, there's no wires connecting the chips together, but the basic ingredients of a Von Neumann machine are still there, i.e. memory and processor.
As somebody has already said it's not going to be just one chip, however even single-chip computers still follow the model. Yes, the processor and memory reside on the same chip, but they're still logically separate units. Indeed most modern processor chips are Von Neumann machines containing CPU and cache in a single unit.
Will this wipe out motherboards? You still need to put these chips somewhere, and I would speculate that they wouldn't be very tolerant of being moved about relative to each other whilst operational. They also need power.
Whilst Mac OS X definitely does make good use of the Velocity Engine to speed things up on machines that have them, and applications also make use of it, much of this is handled through the compiler. From what I've read XCode 1.5 (due any day now) includes some improvements to GCC which do some automatic vectorising. Those improvements are either already part of the main GCC branch or they'll filter into it.
We'll never see Mac OS X for Intel simply because it's not in Apple's interests to support a general Intel platform. They make too much money from their hardware business. I also doubt very much that Apple will return to licensing the Mac OS.
What is sad is that Apple killed of Yellow Box for Windows. Yellow Box is what is now called Cocoa on Mac OS X, and the Windows version was basically an updated version of OpenStep. Of course most of the big apps don't use Cocoa right now, but there's many coming that do use it, and it would have been nice to see them portable between platforms. Instead the only portability option is GNUstep, which is a bit behind, to put it mildly.
Now when I was a very young kid the Daleks were frightening.
As I grew up I came to realise that the Daleks were a pretty sad and lame enemy. All the Doctor had to do to run away from them was climb a flight of stairs. Or just run, since they were so slow.
Of course eventually Daleks could climb stairs, along with a cheesy special glowy effect. Pah! It was too late by then, and the childishness of the late Dr. Who was ruining things anyway. No, I don't see them being able to come up with anything decent with Daleks.
As for bringing back the Cybermen, yes, I'd much prefer that. For me that's the real story here, since the Cybermen are also a Terry Nation creation and I believe that this announcement clears their use in the series too. However I'd like to see them toughed up a bit - they always seemed a bit slow and the silly "rub gold on their grill and they die" thing wasn't great.
But why bother resurrecting any of these old villans? There's a whole big universe of time and space out there, and whilst The Doctor may be fascinated with Earth (and might be half human) there's little reason to stick with the old stuff.
Let's face it Bill's a nerd, and he has a particularly bad record at predicting computing trends. What he is good at is business.
Remember too this was in the days of DOS. Microsoft took a while to realise that GUIs would be important. Graphical interfaces suck up memory. If you're using a GUI 640kb isn't even enough memory to display the screen you're looking at these days, let alone run some programs too.
An even greater example of short-sightedness in the area of RAM was Steve Jobs insisting that the Mac should be restricted to 256kb, since it was intended to be a comodity appliance. Fortunately the engineers didn't listen to him and made sure it could be upgraded.
Your post trying to show how America was "first" in space is quite amusing.
You start with the ship rendezvous thing, but point out that the Russians did this first (although a slightly easier version).
The men around and on the moon is a given for America, but that was after the USSR had sent unmanned probes there, beating the USA. This goes with your later point about "human-guided landing"... Surely a machine guided landing is more impressive?
Sure the first untethered space walk may have been the USA but Leonov walked in space many years before then. Not a particularly important achievement really.
"First serious use of Geosynch communications sats" implies that the USSR had put up such birds before the USA, just that they weren't commerically exploited.
The booster docking is fair enough to claim as an achievement for the USA, as is the reusable orbital cargo vehicle (shuttle) claim.
Finally as others have pointed out Salyut came before Skylab, and you're putting an arbitrary and largely unexplained definition on "space station" to make your claim here.
Of course in your efforts to big-up the US space program you also fail to mention that Russia visited a whole load of planets (both orbital and surface), which is a pretty significant achievement.
The USA do have some impressive achievements in space, but the claim that they won the space race is only propaganda - they won the moon race.
Now the moon race win was of course very impressive, but it's not as big a win as you might think. A big reason for the victory is that NASA started their efforts in 1961, whereas the government of the USSR didn't get serious about competing until 1964. They still came fairly close, and if they had those extra few years they may well have won.
I must admit that I haven't looked at your API for your GUI but going in a completely new direction doesn't seem like a great idea to me. A decent high-level API for GUI work is a must.
I'm a Mac programmer thesse days and the Cocoa API (which is very closely related to GNUstep) provides for a very powerful environment in which to develop applications.
That, and a little thingy called 'Velocity Engine'. Oh, and that endian stuff. Trivial, I'm sure...
You forget that Mac OS X runs on machines with G3 processors that have no Velocity Engine. That isn't a big problem.
As for the endian problem, it's not really that tough to handle.
You're also forgetting the heritage of Mac OS X. NeXTStep ran on Intel hardware, and the first versions of Rhapsody (the immediate predecessor to Mac OS X) ran on Intel too.
Requiring patent holders to use their patent immediately would only encourage ideas to stay secret and not be patented until such time as it's practical to exploit the idea embodied within the patent. Many ideas are already "trade secrets" rather than patents, and so long as they are kept secret they'll never be released to the public domain. This is an advantage of patents after all - they do expire.
Patents used to be for a shorter period and could be extended. I would suggest that is a better solution than the current one, however I would be reluctant to see a proviso added that a patent must be in the process of being developed into a product. Some patented ideas really can't practically be turned into products until quite late into their life.
I just read all the stories that were linked to this article.
None of them blamed Microsoft. In fact the only blame pushed in their direction was your comment...
The articles did say that there was a problem with the operating system. Now we don't know who exactly said this, or what they said precisely, so it is quite possible that this isn't entirely accurate reporting.
I find it very difficult to believe that they would have any single points of failure in a system of that importance.
I agree it's unlikely, but it is possible that there is a single point of failure in their system. There are a great deal of shoddily engineered systems in use today.
It all depends on what you want to use your computer for.
If all you want from your computer is word processing, web browsing and email access then you don't need much processing power at all.
If, in contrast, you want to do video editing, applying plenty of real-time effects, and decoding/encoding to a compressed format then a high-end dual processor machine is handy.
People are getting used to their computers being able to do things in real-time. Consumer-level applications make a lot of use of real-time processing, since a regular person who doesn't understand about computers is less tolerant of the need that a computer has to process things and expect things to happen instantly.
For example, I just wrote an application which is a modern version of an app I helped write over ten years ago. The app makes graphics files. The original app took its input settings and created the output file when you saved it. When using the app it was common to create several files before you got the output you really wanted. The modern version produces the graphics in real-time and displays them, so the user knows exactly what they're getting before they save the file. Users tend to prefer the new app over the old one, even though right now there's a couple of features missing.
As for seeing an overtaking, I think that depends on the application being used for corporate things. Plain web serving of HTML or file and print serving doesn't take that much power, but video-on-demand is much more demanding.
Accepted, Quartz Extreme was not in the first release of Mac OS X, but I didn't claim it was. What I was saying was that most of the technology behind Quartz Extreme was present in the first release of Mac OS X.
There is nothing new about having a vector based rendering model for your graphics system. NeXTStep had that on its first release in 1988. Indeed in 1991 NeXT had accelerated Display PostScript with their NeXTdimension graphics accelerator card.
For the vector stuff (and animations) it seems that Avalon is catching up to where NeXT was in '91.
Quartz is a direct descendant of NeXT's Display PostScript, but you are right, the current version of Quartz Extreme is only using hardware acceleration for compositing windows. It is, in some ways, a step backwards compared to what was possible with NeXT hardware in '91. However in those days we didn't have the kind of GPUs we have today - a NeXTdimension was actually an Intel i860 co-processor along with some video hardware. NeXT ended up focussing on just their OS on commodity PC hardware which meant no real possibility for continuing with accelerated Display PostScript. i860-based graphics cards were significantly more expensive than the 2D and later 3D accelerator cards popular on PCs. It's only been recently with the introduction of programmable GPUs that accelerating vector graphics has been possible again.
Quartz Extreme is developing, but Apple rarely announces features in advance. They're talking about some of the new things coming in Tiger, however there's bound to be plenty more that will also be in there. I wouldn't be surprised if there were some more acceleration features - accelerating the existing display stuff just doesn't make for sexy demos, since it's already damn fast.
One point here though is that it might not actually be sensible to put too much emphasis on using the GPU for rendering, at least at the moment. Balancing rendering tasks between the GPU and CPU can allow for more sophisticated graphics to be created. This is, after all, what the Doom3 engine does to generate its graphics - it's not all GPU work you see on-screen.
Finally you're right - I haven't looked at the Longhorn SDK. I don't plan to either, not unless I get a job programming for Windows for a developer thinking very long-term. Remember that Longhorn is not expected to be available until '06 at the earliest, and judging by the uptake of previous Windows versions won't be common until '08-09.
The sensible way of handling font anti-aliasing on variable DPI displays is to specify the pixel size at which fonts get anti-aliased, rather than the point size. I think that both Microsoft and Apple will work that one out for us.
Core Image in Mac OS X 10.4 (coming early 2005) makes good use of GPU features.
As for when the last time I saw a software rendered engine, well, the best engines are usually a combination of software and hardware. From what I understand Doom3 uses a combined approach making use of SIMD instructions on the CPU to compliment work done by GPUs.
How can an OS that's playing catchup to Apple's Panther redefine the industry?
It's also not going to be delivered until a year, or maybe more, after Tiger. Longhorn's Avalon sounds like the Macs Quartz Extreme display engine (which has been out for almost a year now) minus the new Core Image and Core Video extensions that are coming in Tiger. If you want to see what this stuff looks like check out Steve Jobs' keynote from WWDC and fast forward to about 1hr 7mins in to hear something about it, or 1hr 9mins to see a demo.
Given what Apple are doing Microsoft has little choice but to try to catch up, otherwise they're going to look really tired. For those in the know the imaging system in XP really is very tired compared to Mac OS X 10.3, but the differences aren't really dramatic enough for Apple to shout about. Early next year though when 10.4 arrives with Core Image and Core Video I think we can expect them to be shouting very loudly.
As for being sick of hearing about software that won't be in use for another 3-5 years, maybe you should consider switching.
Dude, I hate to break your bubble, but Microsoft are behind the times here.
Avalon is Microsoft's version of Apple's Quartz Extreme. Apple had most of this in the first release of Mac OS X.
Now next year that Mac gets Mac OS X 10.4 "Tiger" which features a great new technology called Core Image. Using Core Image you can apply image effects such as gaussian blur, motion blur, trirl distortions, hue adjustments, perspective transformations all in real-time. Let me say that again - in real-time. Oh, and there's Core Video too which lets you apply all of these effects to video streams, again in real-time. This stuff works right now and is already in the hands of many developers.
Will Avalon offer any of these kind of features? Who knows! It sounds right now as if Avalon is just Quartz Extreme without Core Image, meaning that Longhorn is playing catch-up with Mac OS X 10.3... They've got a way to go to be up with 10.4.
It's not that amazing really - Microsoft are just playing catch-up as usual. As are the Linux boys, although they are much closer to the state of the art.
Mac OS X has an OS-level ability to use 3D acceleration features of graphics cards in more than one application at a time. The technology is called Quartz Extreme. Actually Quartz Extreme means that from 10.3 onwards every desktop application is OpenGL accelerated, since the window manager was revamped. Prior to 10.3 OpenGL was still around and you could use it in multiple applications simultaneously.
I hope that the open source world does manage to unite enough to bring these kind of facilities to mainstream Linux and OpenBSD distributions. It would be even nicer if GnuStep got a little more support, since it's a great application development framework. The nicest result would be improved GnuStep and these advanced graphics systems being API compatible with Quartz. Porting Cocoa stuff from Mac OS X to other Unixes would then be trivial.
For starters, I don't understand why there is a necessity to constantly re-invent the wheel and create gaming engines from scratch just about each time a new game is released. Surely it would be better to throw out the source code to current gaming engines to the Internet community to see what enhancements get added as a result - sure, keep the level design, textures, etc. for a specific commercial game that uses that engine under wraps so that, as a game company, you can make money from it.
Two URLs for you: http://www.renderware.com/ http://www.idsof tware.com/business/technology/
Many games are based on the Renderware engine from Criterion. They were just bought by EA this week.
Many other games are based on the Doom and Quake engines from id.
There are other gaming engines besides those offered by Criterion and id - physics engines, particle engines, rendering engines...
Many game developers don't feel the need to write their own wizzy engine. Grand Theft Auto 3 and its sequels are all based on Renderware, for example. In fact there are several hundred games in development right now that use Renderware.
There's a big difference between open APIs and open source.
OpenGL is an open API, meaning that the API spec is published openly and anyone can implement code that follows that spec, whether it's an application or a driver. The drivers though that follow this API don't have to be open source, and usually aren't. The same applies to applications that use it.
I believe the same applies to the other APIs that are being talked about here.
Sony will, I am sure, have their own proprietary implementations of these open APIs. Sure, some code may be opened to help porting the API to other platforms, but the very low level stuff that talks to hardware is likely to remain proprietary. I also say "opened" here, but in many cases that openness will be restricted to companies within "the club".
Slashdot Mirror
Now this raises an interesting question.
Why do we have both Mono and DotGNU implementations of Windows.Forms?
Why dedicate four senior Mono developers to solving something that the GNU project has already tackled? Won't the DotGNU version work with Mono? If that's the case why not just take the DotGNU version as a starting point and work from there?
Or is this some kind of licensing issue? It would not surprise me if the DotGNU implementation was lumbered with a full GPL, rather than an LGPL, thus making it unusable for anything appart from "free" software.
I think the big omission in GNUstep though is support for bindings. Last time I checked they didn't have Key Value Coding and Key Value Observation support in their NSObject implementation, which is required for bindings as well as more advanced data models. It would be really nice if they had this, since it would help porting 10.3 apps to Linux.
There is of course the old saying "the enemy of my enemy is my friend".
Bush doesn't understand that.
In Bush-logic if you don't agree with him you must therefore completely disagree, therefore you are the enemy. Unfortunately it seems that much of the country is this simplistic too.
In the immediate aftermath of 9/11 any country that expressed an opinion against the actions of the USA was villified, and some branded as terrorist states. Honestly, renaming "French fries" as "freedom fries" was the most rediculous thing I had ever heard and the general hatred of the French truly demonstrated mass idiocy. If people were more aware that the Statue of Liberty was a gift from France they would have been calling for it to be ripped down too.
It's not even crazy these days to think that you can change lead into gold. Gold can be created from lead, with the help of some very expensive equipment and only in very very minute quantities. You just have to knock the right number of protons, neutrons, and electrons off of lead with a supercollider. Simple really. :-)
So robots disguised as humans is old hat. So what? We already knew that a new Battlestar Galactica series wasn't going to be original.
:-)
I felt though that it is a refreshing spin on the old series. One possibility that was never present in the original BSG was Cylon infiltration. That is now integral to the plot, and frankly I like it.
Now I did used to like the original series too, but I was 10 years old back then and the unsophisticated story line appealed to me. Now that I'm over 30 I want something a little bit more challenging.
As for the "shakey camera" effect, that was pioneered on Star Trek back in the '60s - Firefly just copied it.
Yeah, I remember that.
:-)
Snooker's a much harder game than pool, since the table is larger, the pockets smaller, and there's more balls on the table to start with. The robot on Horizon could actually play a real game, following the rules or snooker, and making intelligent shots with positioning.
From the article it sounds like this pool playing robot is pretty crummy right now. They hilight its strength as being able to pot the white in any pocket, but make it clear that it has trouble potting any coloured ball. Whether it can actually play a game isn't clear - I get the impression though that it's not programmed yet to put english on the ball.
Whilst this might technically be the first pool playing robot that's only because the British robot that came before it was never reprogrammed to play pool. Maybe the brits should resurrect their old creation and challenge these guys to a game.
Maybe I'm being stupid, but I don't see how this work obseletes the Von Neumann architecture.
Sure, there's no wires connecting the chips together, but the basic ingredients of a Von Neumann machine are still there, i.e. memory and processor.
As somebody has already said it's not going to be just one chip, however even single-chip computers still follow the model. Yes, the processor and memory reside on the same chip, but they're still logically separate units. Indeed most modern processor chips are Von Neumann machines containing CPU and cache in a single unit.
Will this wipe out motherboards? You still need to put these chips somewhere, and I would speculate that they wouldn't be very tolerant of being moved about relative to each other whilst operational. They also need power.
Whilst Mac OS X definitely does make good use of the Velocity Engine to speed things up on machines that have them, and applications also make use of it, much of this is handled through the compiler. From what I've read XCode 1.5 (due any day now) includes some improvements to GCC which do some automatic vectorising. Those improvements are either already part of the main GCC branch or they'll filter into it.
We'll never see Mac OS X for Intel simply because it's not in Apple's interests to support a general Intel platform. They make too much money from their hardware business. I also doubt very much that Apple will return to licensing the Mac OS.
What is sad is that Apple killed of Yellow Box for Windows. Yellow Box is what is now called Cocoa on Mac OS X, and the Windows version was basically an updated version of OpenStep. Of course most of the big apps don't use Cocoa right now, but there's many coming that do use it, and it would have been nice to see them portable between platforms. Instead the only portability option is GNUstep, which is a bit behind, to put it mildly.
Now when I was a very young kid the Daleks were frightening.
As I grew up I came to realise that the Daleks were a pretty sad and lame enemy. All the Doctor had to do to run away from them was climb a flight of stairs. Or just run, since they were so slow.
Of course eventually Daleks could climb stairs, along with a cheesy special glowy effect. Pah! It was too late by then, and the childishness of the late Dr. Who was ruining things anyway. No, I don't see them being able to come up with anything decent with Daleks.
As for bringing back the Cybermen, yes, I'd much prefer that. For me that's the real story here, since the Cybermen are also a Terry Nation creation and I believe that this announcement clears their use in the series too. However I'd like to see them toughed up a bit - they always seemed a bit slow and the silly "rub gold on their grill and they die" thing wasn't great.
But why bother resurrecting any of these old villans? There's a whole big universe of time and space out there, and whilst The Doctor may be fascinated with Earth (and might be half human) there's little reason to stick with the old stuff.
640k my ass billy G!
Let's face it Bill's a nerd, and he has a particularly bad record at predicting computing trends. What he is good at is business.
Remember too this was in the days of DOS. Microsoft took a while to realise that GUIs would be important. Graphical interfaces suck up memory. If you're using a GUI 640kb isn't even enough memory to display the screen you're looking at these days, let alone run some programs too.
An even greater example of short-sightedness in the area of RAM was Steve Jobs insisting that the Mac should be restricted to 256kb, since it was intended to be a comodity appliance. Fortunately the engineers didn't listen to him and made sure it could be upgraded.
Your post trying to show how America was "first" in space is quite amusing.
You start with the ship rendezvous thing, but point out that the Russians did this first (although a slightly easier version).
The men around and on the moon is a given for America, but that was after the USSR had sent unmanned probes there, beating the USA. This goes with your later point about "human-guided landing"... Surely a machine guided landing is more impressive?
Sure the first untethered space walk may have been the USA but Leonov walked in space many years before then. Not a particularly important achievement really.
"First serious use of Geosynch communications sats" implies that the USSR had put up such birds before the USA, just that they weren't commerically exploited.
The booster docking is fair enough to claim as an achievement for the USA, as is the reusable orbital cargo vehicle (shuttle) claim.
Finally as others have pointed out Salyut came before Skylab, and you're putting an arbitrary and largely unexplained definition on "space station" to make your claim here.
Of course in your efforts to big-up the US space program you also fail to mention that Russia visited a whole load of planets (both orbital and surface), which is a pretty significant achievement.
The USA do have some impressive achievements in space, but the claim that they won the space race is only propaganda - they won the moon race.
Now the moon race win was of course very impressive, but it's not as big a win as you might think. A big reason for the victory is that NASA started their efforts in 1961, whereas the government of the USSR didn't get serious about competing until 1964. They still came fairly close, and if they had those extra few years they may well have won.
Have you considered porting GNUstep to Syllable?
I must admit that I haven't looked at your API for your GUI but going in a completely new direction doesn't seem like a great idea to me. A decent high-level API for GUI work is a must.
I'm a Mac programmer thesse days and the Cocoa API (which is very closely related to GNUstep) provides for a very powerful environment in which to develop applications.
That, and a little thingy called 'Velocity Engine'. Oh, and that endian stuff. Trivial, I'm sure...
You forget that Mac OS X runs on machines with G3 processors that have no Velocity Engine. That isn't a big problem.
As for the endian problem, it's not really that tough to handle.
You're also forgetting the heritage of Mac OS X. NeXTStep ran on Intel hardware, and the first versions of Rhapsody (the immediate predecessor to Mac OS X) ran on Intel too.
Requiring patent holders to use their patent immediately would only encourage ideas to stay secret and not be patented until such time as it's practical to exploit the idea embodied within the patent. Many ideas are already "trade secrets" rather than patents, and so long as they are kept secret they'll never be released to the public domain. This is an advantage of patents after all - they do expire.
Patents used to be for a shorter period and could be extended. I would suggest that is a better solution than the current one, however I would be reluctant to see a proviso added that a patent must be in the process of being developed into a product. Some patented ideas really can't practically be turned into products until quite late into their life.
Blaming MS is the easy way out.
I just read all the stories that were linked to this article.
None of them blamed Microsoft. In fact the only blame pushed in their direction was your comment...
The articles did say that there was a problem with the operating system. Now we don't know who exactly said this, or what they said precisely, so it is quite possible that this isn't entirely accurate reporting.
I find it very difficult to believe that they would have any single points of failure in a system of that importance.
I agree it's unlikely, but it is possible that there is a single point of failure in their system. There are a great deal of shoddily engineered systems in use today.
It all depends on what you want to use your computer for.
If all you want from your computer is word processing, web browsing and email access then you don't need much processing power at all.
If, in contrast, you want to do video editing, applying plenty of real-time effects, and decoding/encoding to a compressed format then a high-end dual processor machine is handy.
People are getting used to their computers being able to do things in real-time. Consumer-level applications make a lot of use of real-time processing, since a regular person who doesn't understand about computers is less tolerant of the need that a computer has to process things and expect things to happen instantly.
For example, I just wrote an application which is a modern version of an app I helped write over ten years ago. The app makes graphics files. The original app took its input settings and created the output file when you saved it. When using the app it was common to create several files before you got the output you really wanted. The modern version produces the graphics in real-time and displays them, so the user knows exactly what they're getting before they save the file. Users tend to prefer the new app over the old one, even though right now there's a couple of features missing.
As for seeing an overtaking, I think that depends on the application being used for corporate things. Plain web serving of HTML or file and print serving doesn't take that much power, but video-on-demand is much more demanding.
Accepted, Quartz Extreme was not in the first release of Mac OS X, but I didn't claim it was. What I was saying was that most of the technology behind Quartz Extreme was present in the first release of Mac OS X.
There is nothing new about having a vector based rendering model for your graphics system. NeXTStep had that on its first release in 1988. Indeed in 1991 NeXT had accelerated Display PostScript with their NeXTdimension graphics accelerator card.
For the vector stuff (and animations) it seems that Avalon is catching up to where NeXT was in '91.
Quartz is a direct descendant of NeXT's Display PostScript, but you are right, the current version of Quartz Extreme is only using hardware acceleration for compositing windows. It is, in some ways, a step backwards compared to what was possible with NeXT hardware in '91. However in those days we didn't have the kind of GPUs we have today - a NeXTdimension was actually an Intel i860 co-processor along with some video hardware. NeXT ended up focussing on just their OS on commodity PC hardware which meant no real possibility for continuing with accelerated Display PostScript. i860-based graphics cards were significantly more expensive than the 2D and later 3D accelerator cards popular on PCs. It's only been recently with the introduction of programmable GPUs that accelerating vector graphics has been possible again.
Quartz Extreme is developing, but Apple rarely announces features in advance. They're talking about some of the new things coming in Tiger, however there's bound to be plenty more that will also be in there. I wouldn't be surprised if there were some more acceleration features - accelerating the existing display stuff just doesn't make for sexy demos, since it's already damn fast.
One point here though is that it might not actually be sensible to put too much emphasis on using the GPU for rendering, at least at the moment. Balancing rendering tasks between the GPU and CPU can allow for more sophisticated graphics to be created. This is, after all, what the Doom3 engine does to generate its graphics - it's not all GPU work you see on-screen.
Finally you're right - I haven't looked at the Longhorn SDK. I don't plan to either, not unless I get a job programming for Windows for a developer thinking very long-term. Remember that Longhorn is not expected to be available until '06 at the earliest, and judging by the uptake of previous Windows versions won't be common until '08-09.
No, I don't work for Criterion. :-)
The sensible way of handling font anti-aliasing on variable DPI displays is to specify the pixel size at which fonts get anti-aliased, rather than the point size. I think that both Microsoft and Apple will work that one out for us.
Core Image in Mac OS X 10.4 (coming early 2005) makes good use of GPU features.
As for when the last time I saw a software rendered engine, well, the best engines are usually a combination of software and hardware. From what I understand Doom3 uses a combined approach making use of SIMD instructions on the CPU to compliment work done by GPUs.
Redefined?
How can an OS that's playing catchup to Apple's Panther redefine the industry?
It's also not going to be delivered until a year, or maybe more, after Tiger. Longhorn's Avalon sounds like the Macs Quartz Extreme display engine (which has been out for almost a year now) minus the new Core Image and Core Video extensions that are coming in Tiger. If you want to see what this stuff looks like check out Steve Jobs' keynote from WWDC and fast forward to about 1hr 7mins in to hear something about it, or 1hr 9mins to see a demo.
Given what Apple are doing Microsoft has little choice but to try to catch up, otherwise they're going to look really tired. For those in the know the imaging system in XP really is very tired compared to Mac OS X 10.3, but the differences aren't really dramatic enough for Apple to shout about. Early next year though when 10.4 arrives with Core Image and Core Video I think we can expect them to be shouting very loudly.
As for being sick of hearing about software that won't be in use for another 3-5 years, maybe you should consider switching.
Dude, I hate to break your bubble, but Microsoft are behind the times here.
Avalon is Microsoft's version of Apple's Quartz Extreme. Apple had most of this in the first release of Mac OS X.
Now next year that Mac gets Mac OS X 10.4 "Tiger" which features a great new technology called Core Image. Using Core Image you can apply image effects such as gaussian blur, motion blur, trirl distortions, hue adjustments, perspective transformations all in real-time. Let me say that again - in real-time. Oh, and there's Core Video too which lets you apply all of these effects to video streams, again in real-time. This stuff works right now and is already in the hands of many developers.
Will Avalon offer any of these kind of features? Who knows! It sounds right now as if Avalon is just Quartz Extreme without Core Image, meaning that Longhorn is playing catch-up with Mac OS X 10.3... They've got a way to go to be up with 10.4.
It's not that amazing really - Microsoft are just playing catch-up as usual. As are the Linux boys, although they are much closer to the state of the art.
Mac OS X has an OS-level ability to use 3D acceleration features of graphics cards in more than one application at a time. The technology is called Quartz Extreme. Actually Quartz Extreme means that from 10.3 onwards every desktop application is OpenGL accelerated, since the window manager was revamped. Prior to 10.3 OpenGL was still around and you could use it in multiple applications simultaneously.
I hope that the open source world does manage to unite enough to bring these kind of facilities to mainstream Linux and OpenBSD distributions. It would be even nicer if GnuStep got a little more support, since it's a great application development framework. The nicest result would be improved GnuStep and these advanced graphics systems being API compatible with Quartz. Porting Cocoa stuff from Mac OS X to other Unixes would then be trivial.
For starters, I don't understand why there is a necessity to constantly re-invent the wheel and create gaming engines from scratch just about each time a new game is released. Surely it would be better to throw out the source code to current gaming engines to the Internet community to see what enhancements get added as a result - sure, keep the level design, textures, etc. for a specific commercial game that uses that engine under wraps so that, as a game company, you can make money from it.
f tware.com/business/technology/
Two URLs for you:
http://www.renderware.com/
http://www.idso
Many games are based on the Renderware engine from Criterion. They were just bought by EA this week.
Many other games are based on the Doom and Quake engines from id.
There are other gaming engines besides those offered by Criterion and id - physics engines, particle engines, rendering engines...
Many game developers don't feel the need to write their own wizzy engine. Grand Theft Auto 3 and its sequels are all based on Renderware, for example. In fact there are several hundred games in development right now that use Renderware.
There's a big difference between open APIs and open source.
OpenGL is an open API, meaning that the API spec is published openly and anyone can implement code that follows that spec, whether it's an application or a driver. The drivers though that follow this API don't have to be open source, and usually aren't. The same applies to applications that use it.
I believe the same applies to the other APIs that are being talked about here.
Sony will, I am sure, have their own proprietary implementations of these open APIs. Sure, some code may be opened to help porting the API to other platforms, but the very low level stuff that talks to hardware is likely to remain proprietary. I also say "opened" here, but in many cases that openness will be restricted to companies within "the club".