Yes. It does.
Also, it has an objective C++ runtime environment going by default as well. I think these are encapsualted in the Cocoa component. I could be wrong though. - Paradox Man of the C!!!
Err, unlike the GNU compiler which no one has been updating in the objc area, the Apple C Compiler uses Objective C++. C++ objects and syntax and items augment the static typing facilities of objective C in this situation. So, therefore, porting all C++ apps will be easy. Isn't the Carbon API C++? I thought only Cocoa was the the Java and Objective-C++ (two excellent languages btw) API set. And sure, those langauges are more flexible than the less dynamic C++. C'est la C++.
However, I _strongly_ encourage anyone open minded to look at Objective C. Just because it is old (and it's not THAT old!) dosen't mean it dosen't rule. Objective C is a really powerful and usefull language, based off of Smalltalk, which is famous for good reason.
One of the reasons Objective-C++ was discarded way back when was that it required garbage collection and the method calling is a tad slower (not much, mind you, and you can eliminate that delay with a line or two more of code). These days, it's not really true anymore. We've come a long way since the days of NeXT and the slow garbage collection and dynamic dispatch.
C++ merely dominates because MS pushes it. I'm not surprised, C++ seems like a Microsofty language to me, quite frankly, with it's not quite object oriented view. - Paradox Man of the C!!!
Well, Since OS X includes Cocoa, which is a fully OpenStep compliant development environment, then I can say with certainty, it rules.
OpenStep is a beatufiul, object oriented and well designed environment to work with! Better yet, you get to use Objective-C++, which makes development of OO programs much faster and cleaner because the language is so much nicer than straight C++ (imnsho, anyways, C++ is a turd compared to ObjC).
All the data structures you need to make a program are already made for you, and the GUI classes are intuitive and easy.
So yes, developing under Cocoa will be easy. Carbon apps are slightly harder from what I hear, but I haven't worked with them. However, it's much easier than the dark ages of the older OS 7-8 approach, which was an amazing pain. - Paradox Man of the C!!!
Hi. Check out IBM's website. You might actually begin to know what you are talking about. To be quite honest, now that OS X is out, even just a beta, to prove it's not a joke, I'm switching over. I am SICK SICK SICK of x86 Hardware that sucks, blows, and otherwise makes me pay sizable amounts of money for crap.
The G4 boxes (not the cubes) have come down in price. Just because they say "400 mhz" dosen't mean they suck. In fact, the G4's, at present speed, are very competetive on the modern market. Further, apple boxes cost more because they have more cool stuff in them. USB, Firewire, DVD RAM stuff is expensive, but apple makes them available for a reasonable price. Of course they overcharge for RAM, but hey no one is perfect.
You're just upset because you want something for your x86 box. Why? To avoid spending money? Fat chance of that working. To avoid losing linux maybe.. ok, I could see that being a concern. A false one, LinuxPPC is in great shape.
Apple can afford to bide a bit of time here, when everyone comes time to buy their next computer, they'll find for less than $300 more, they can get a far superior piece of hardware and monitor, with a terrific GeekMeetsTrendy OS. I don't see how they can go wrong NOT porting it. - Paradox Man of the C!!!
You can make that argument for anything, and it's meaningless. If Hitler hadn't existed, some other tyrant would have taken his place. Riight.
RMS helped to organize the FSF. This has led to more free software under an organizational scheme which is important. I dislike RMS's principle of trying to force people to conform. He is by no means my saint, but he has helped people Accomplish Things. I am compelled therefore to give him some respect. Things I like to use, such as emacs, exist in part because the FSF helped to make them. You can't toss me a hypothetical argument and expect it to hold water.
Of course free software existed before RMS. And of course it will exist without RMS. And yes, a lot of FSF software sucks, I've ranted about this. But we work with what we have. The gcc is an excellent example. It is NOT a cruddy program. Especially since it is free, it is actually a stellar deal.
There may be better compilers, but they are placed out of my reach by prohibitive cost. Lots of developers forget that, being backed by large companies who can afford the liscense fee. What do you want me to do? Go out and buy an SGI and IRIX and their devtools? They're pretty good, better than the gcc anyways. Too bad it costs over $1k just for a year of use.
I'll get right on it. Riiight. - Paradox Man of the C!!!
I apologize, I had it wrong. It's been awhile since I used mac os. The last version I touched was 8.0. - Paradox Man of the C!!!
This is why we need QUALITY Free Software
on
A Letter from 2020
·
· Score: 1
The biggest problem with the flood of new developers publishing under the GPL, for GNU, and floods of submissions to Freshmeat.net, is thats a lot of them SUCK. I don't care if fr3d wrote a Perl script to sort his mp3's. Oooh, it's released under the GPL, how trendy.
Has anyone heard a pro developer complain saying, "It's so GNU, it's practically worthless!" ? There is a undercurrent I've discovered of people tired of being told the virtues of software freedom, and trying it only to see that it, quite frankly, sucks sometimes. All these worthless, or neat but poorly written programs, give Free software (by any definition) bad reps. For instance. Look at GNU Emacs. Despite people who dislike emacs in favor of VI, one thing you need to admit is that barring lisp hacking, Emacs dosen't crash. Heck, I've written some pretty wierd stuff in elisp, and I've never once seen it crash, or lost more than a few characters of data, with emacs (with stable versions of course).
Linux is another example. It works, it works well, and fixes it's problems quickly. But unfortunatly in the growing crowd of GPL'ed, BSD'ed,and public domain software, they are exceptions. No one will respect the ideals of GNU, the ideals of the BSD liscense, or the idea that software should be in the hands of and in the control of the USERS, not the developers. Developers only exist to fill the users needs.
I don't care if someone says "Well I'm not getting paid, so don't expect so much." This is a bad attitude, and people like this should STOP MAKING PROGRAMS. Don't release something unless it's usefull. And feel free to release COMPONENT code!!! Libraries are plentiful and easy to modify. Fully working, poorly written programs are less usefull to me than a well written reusable librarly (the GMA library comes to mind).
Also, programmers of Free, Open Sourced, or Public Domain software should APPLY FOR PATENTS and then hold them. Then retire the patent. This way we legally hold work in the public's grasp.
I do not know a service established to help us, lone developers or small project groups, to accomplish this. If there is one, point more people to it!
This is just my personal rant. - Paradox Man of the C!!!
I fully agree. Apple has finally gotten off their ass and "Thought Different" for the first time since their shining pinnacle of glory back in the beginning.
This OS is basically what Be strove to be, UNIX for the rest of us.. but it's not just for the rest of us. It's the real deal here folks. At the same time, we get a nice (themeable, did you know that? Macs went themeing many years ago with Kagi Software's Kaleidoscope) GUI and some good graphics performance. Ask anyone whos serious about *NIX, "What's wrong with *NIX?" and they'll say "Xwindows" right off the bat.
So now we have the best of both worlds here. I for one, am shelling out the cash for dual 500mhz G4 as soon as I get my next paycheck. The hardware is better, and now the OS is better too. What more reason do you want? You don't even need to settle for a 1 button mouse!
"Coming along" meaning what? That one day we might have quantum computers? 'sides. You can make current encryption algorithms work even when there are quantum computers that can do a whole lot of stuff at once in constant time, you just jack up keysizes, prime sizes (in the case of RSA) and the like. At least, this is how I understand it. Imagine if you could generate primes on a real working quantum computer. You could make primes so huge they would foil quantum computers, couldn't you?
But right now quantum computers, with reasonable certainty, don't exist in any useful form. - Paradox Man of the C!!!
You know, everyone loves to hate RMS.
But without him, a lot of software we
take for granted today wouldn't be around.
So, maybe sometimes RMS goes a bit off the deep
end, and goes a bit to far in his quest for
software to be totally free. Maybe he has a bad
singing voice, and maybe he's a little scary.
Without him GNU and the GPL (and it's (better
imho) younger sibling the LGPL) woudln't exist.
These things are Good Things that most people
can agree have caused some very nice software
to be free of charge and with source distributed.
So give him a break folks. No one is perfect. - Paradox Man of the C!!!
Hi.
Go out and buy a G4, people. Do you have any idea how crappy our beige boxes are, even brand new ones, in comparison with these G4s?
"But Paradox, you're wrong, the G4 only runs at 500mhz!" You say. Please try and remember that mhz dosen't mean all that much. Look at the MIPS, FLOPS, and practical performance. Biiiiig difference. The G4 is such a cool machine I can't imagine why people are buying them en masse. The hardware is just too good to shrug it off with the phrase, "It's a mac." One of apple's new G4 cubes makes a damn sexy linux box. And sans monitor they are really quite cost effective for what you are getting. Linux is in a comparable state for the newer motorolla chips, and all your favorite programs will work so long as you have the source, with a few Loki game ports excepted.
Apple should be hardware company. Their machines kick ass once you unshackle them from MacOS. So stop poking at the big guy (Intel) and rooting for the underdog (AMD) and actally buy a really nice computer based on a very nice RISC design. - Paradox Man of the C!!!
Hmm.
Debian needs to fix this problem. Debian has given me nothing but problems, which is why I switched to slackware (and am happy now). I actually left debian because it took me 2, count them 2 weeks to figure out how to shore up debian so X would compile. Every other system I know of, it just works. It's ridiculous.
Compiling X4.0 on any distro but debian is a matter of chaning ProjectRoot (as specified clearly in the INSTALL file, one of the first things it says!) and typing make World &> world.log ; make Install &> install.log ; echo ALL DONE!
You _can_ use the binary distro, but it break's debians overly fragile package system, which means you can't use the cool features apt-get offers you. Anyone using debian is screwed here until SOMEONE working with potato fixes this problem.
Oh, and people who use Linux should expect compilaitions sometimes. It's just The Way Things Are. Probably the way they always will be, and that's not a bad thing, even for end-users who don't code. automake, autoconf.. these make life too easy.
Good luck getting X to work. And remember that Nvidia's drivers fail if you load the DRI modules in your XF86Config file, Nvidia uses DRI through their own modules, because they didn't want to wait for linux to catch up (everyone was complaining). - Paradox Man of the C!!!
Well, you could always run opengl apps and see how they handle. Duh!
Maybe it's not the best benchmark, but running simple linux games like quake3, heavy gear 2 and the like are reasonable judges of how well a 3d card will perform under linux.
I have now tried almost every major brand of card on my machine, and without a question, for games and standard opengl, Nvidia's drivers, in the case of the TNT2 and the Geforce2 I've tried, kick butt.
Try running the program "mtri" under your mesa's demo dir. Make sure to link it to the right opengl libs. Like magic, you can see how many triangles your card can push in-window (which is slower than full screen btw).
In window, the Matrox G400 I tried scored about 1m triangles. The Nvidia geforce score more than that, by almost 3.5 times! (taken at 32-bit color 1280x1024 on an athalon t-bird 800 w/ 256 megs pc100 ram, for those who care). The TNT2 scores at just below 2mtri. For reference, the Voodoo3 3000 scored just below the TNT2, at about 1.8 mtri (this test was done in 16-bit color tho, so ymmv).
That's plenty enough benchmarking for me. Code I write runs well with one card, and at 1/4 speed on a Matrox card. Now, don't get me wrong. I love Matrox. I think it's admirable what they are doing, and I plan to buy a g450 w/ dual heading for my workstatin computer because dual monitors make coding easier. But for raw 3d performace, both in games and simulations, nvidia cards give you the best performance you are going to get on an x86 machine, open source or closed. - Paradox Man of the C!!!
Only if you have shitty level designers and you don't use curves. The Q3 engine, when used right, makes for some very nice detail, because of the
bezier curve stuff. (It is bezier curves, right?).
'sides.. for man-made structures, usually the wall IS a small little sharp corner. I look at the corner in my office and I don't see a smoothly blending curve, I see a very tight corner. If I look very close I see the curve, but it has to be pretty close! - Paradox Man of the C!!!
Fine. Ray Tracing.:P
You still cast rays in ray tracing. Ray casting is just a special case of ray tracing. I hate that stupid semantic separation, logically it makes little sense. - Paradox Man of the C!!!
First let me address the old open-driver issue.
Nvidia doesn't make all the tech they use. They purchase rights to use it. These rights do not include the right to tell everyone how all these things work. So Nvidia, whether it would like to open drivers or not, CANNOT. It's not an issue of Nvidia being dicks, it's an issue of Nvidia being legally bound.
Nvidia is going to try and get drivers out for every market they feel is viable. They're a very busy company, so we'll see about the Be thing. Porting it to BSD shouldn't be too hard, so if Nvidia gets enough email saying, "I use FreeBSD, I'll buy your card if you give me a driver" then thats great. But I doubt many freebsd users (I'm one) use freebsd or openbsd or netbsd for games. It just isn't the focus of the OS. FreeBSD is for high performance stuff on PC hardware. OpenBSD is security, NetBSD is just about propagating. Of all of them, FreeBSD x86 is the most promising. But they need a reason.
I, being a graphics programmer who is devout in his belief OpenGL rules, love Nvidia. They make great drivers for the platforms they DO support, and their linux support is, without question, on par with windows support. Futher, as I've said in other posts, their drivers are very "complete." If you're not an opengl programmer, you may know know about opengl extensions and how much better they can make a given image look. OpenGL drivers with a bunch of offical (SGI recognizes them) extensions are easier to use to make fast, realistic effects.
Since these drivers are complete, Nvidia cards can be used for more than just games. They can be used for simulation, medical imaging, satellite image display, a bunch of cool stuff!
Does 3dfx do this? No. And no one makes a technically better card, with better drivers. I wish they could open source and port to every platform, but I'll take performance over an ethical pat on the shoulder any day. - Paradox Man of the C!!!
Get off your ass and compile Xfree 4.0.
Set the target dir of the install to/use/local/X11R4 or something like that. Then adjust your libpaths and PATH variable to look at the new libraries and binaries.
Taadaa! Finished.
It works, I've got it working right now.
It's really worth it.
Oh, the Nvidia drivers for Xfree 4.0 are:
A) Very fast. As fast as the detanator2 series when they came out for fullscreen stuff, and almost as fast in windowing. For those who love quake3 framerates at full detail 800x600 I never drop below 80, and peak at 150. At optimized detail with low geometry this lower bound goes to 100, similar peak, and I can do it at 1280x1024!
(w/ an 800mhz athalon thunderbird to handle all the culling, portals, and BSP stuff)
I've coded some old opengl demos that ran like crap on a TNT2, because I didn't know what I was doing and coded poorly. The Geforce2 gts pumps them out smoothly, despite my poor old code. Optimized code runs even faster.
B) Very stable. I've not had one crash yet. The older nvidia drivers, the ones for Xfree3 were buggy and frequently put your videocard in a state that required a full hard reboot. Virtual consoles still mess it up, but only when you have a 3d accelerated window running. This is more than I could say for the Voodoo3 drivers I ran for awhile.
C) Being updated! There are new versions coming out with speed improvements and support for new versions of Xfree4.0. As soon as the 2.4 series becomes the real real reality and settles down a bit, Nvidia promises to have those drivers out.
D) Very Complete! Take a look at the opengl extensins support when the geforce drivers run. Try it! Use quake3 if you don't know the gl call. Under driver info. It has a ton of stuff, including Nvidia's cool register combiners and a bunch of SGIS stuff. Stuff opengl programmers love to see, I assure you.
Get a good tech job and buy the damn card. You know you want to:) - Paradox Man of the C!!!
Make some money and buy a nice video card.
Nice video cards with more features mean my opengl programming produces better results, faster. Screw gaming. With Nvidia's great opengl
drivers EVERYTHING works better. Not just my q3 framerate.
I fail to see any penis issues here. Other than the fact you're being dick. - Paradox Man of the C!!!
Just because the japanese use that grammar structure dosen't mean it would affect things. One of the reasons a function header is in front is for indentification purposes. Makes it conveniant to have a nice tag at the start of a block.
Besides, wouldn't compilers have a problem with languages that required that much lookahead? Fast compilers and interpreters usually use predictive parsing and similar stuff for translation. I don't think it would work very well.
Watashi wa Perl ga sukurimas (I use Perl) (I think it's sukao for use, isn't it?) probably isn't the best example anyways. If you look it actually follows a regular pattern that reminds me, at least, of programming. You specify what something is, then what is being done with it.
It dosen't take that much CPU time, you know.
'sides, just breaking a voxel model appart would look ugly! Things don't shatter in nice little pieces that way, they explode in an irregular fashion. You don't need to segment an existing polygonal model, you just add a few more triangles to begin with, or a different model entirely, and billboard up an explosion while you switch the models and send some tank-colored particles everywhere.
It may sound hackier, but it looks pretty good, and it's not very expensive to do. That's what's important in real time rendering.
Ramp-up-down terrain is a thing of the past or an artifact of bad terrain programming. These days even a mediocre OpenGL or D3D programmer can make a very smooth heightmap. It's easy when 2000 triangles is only a small fraction of what your card can push.
Some algorithms you could look at if you are interested: Quadtree LOD, ROAM, Triangle Bintrees w/out ROAM. - Paradox Man of the C!!!
It tessalates them for you, doof. Those aren't openGL calls you are making, they are calls to the GLU which reduces them to polygons for the machine to render.
This isn't acceleration. When cards can be FED a set of NURBS knots and control points, and do it all itself, then it's accelerated.
Look! My card draws quake models in an accelerated fashion. I call myGGLURenderModel() on a string "Ranger" and it draws the ranger model to screen! No. It's not. This Isn't Acceleration.
Massive DUH! on your part. Read what the damn calls do. - Paradox Man of the C!!!
Well, voxels are great ways to model landscapes and such, but to make a voxel-based engine with the versatility and freedom of current polygon engines, and one that has similar memory requirements, is a pain. It's a big pain. And it's not even worth it, you don't stand to gain that much. In more limited (non 6dof) situations, voxels can be pretty easy to implement, though.
This is why Commanche used them. Remember Commanche didn't have a full 6dof, and as such it made the voxel engine design easier. - Paradox Man of the C!!!
Actually, I was very surprised the other day when I loaded a using java3d, and saw MS's java3d implementation seems to bind into direct3d! Maybe I'm wrong and the applet was just really well designed, but there was dynamic lighting and the thing looked great, even had perspective correct texture mapping and bilinear filtering. It was the coolest thing I've seen. I'm not an MSBoy, but I had to hand it to whever made that decision.
What we need is for someone to patch opengl into a java applet, for all platforms. Hardware acceleration for 3d web content is way too cool an idea. - Paradox Man of the C!!!
Problem with this approach is that the ray tree can get arbitrarily deep, in complex scenes. So you don't know with any certainty how long a frame will take to render, or even how long it MIGHT take without directly testing it. Sure, you can do it in parallel, but it's still a huge system of equasions that need to be solved. The benefit of polys and voxels is that they are relativly simple. A triangle, for instantce. Easy to draw a clip.
Everything gets much harder in ray tracing, much harder. And ray tracing means you have to have really accurate models (no blocky models) because you're doing real ray intersections, as such you can't fake things very well with a chunky model. With our current hardware, we can make 300 poly's look like 10000 polys, even under dynamic lighting conditions, in real time. There may be some degenerate cases, but hey, nothing's perfect:)
Slashdot Mirror
Yes. It does.
Also, it has an objective C++ runtime environment going by default as well. I think these are encapsualted in the Cocoa component. I could be wrong though.
- Paradox
Man of the C!!!
Err, unlike the GNU compiler which no one has been updating in the objc area, the Apple C Compiler uses Objective C++. C++ objects and syntax and items augment the static typing facilities of objective C in this situation. So, therefore, porting all C++ apps will be easy. Isn't the Carbon API C++? I thought only Cocoa was the the Java and Objective-C++ (two excellent languages btw) API set. And sure, those langauges are more flexible than the less dynamic C++. C'est la C++.
However, I _strongly_ encourage anyone open minded to look at Objective C. Just because it is old (and it's not THAT old!) dosen't mean it dosen't rule. Objective C is a really powerful and usefull language, based off of Smalltalk, which is famous for good reason.
One of the reasons Objective-C++ was discarded way back when was that it required garbage collection and the method calling is a tad slower (not much, mind you, and you can eliminate that delay with a line or two more of code). These days, it's not really true anymore. We've come a long way since the days of NeXT and the slow garbage collection and dynamic dispatch.
C++ merely dominates because MS pushes it. I'm not surprised, C++ seems like a Microsofty language to me, quite frankly, with it's not quite object oriented view.
- Paradox
Man of the C!!!
Well, Since OS X includes Cocoa, which is a fully OpenStep compliant development environment, then I can say with certainty, it rules.
OpenStep is a beatufiul, object oriented and well designed environment to work with! Better yet, you get to use Objective-C++, which makes development of OO programs much faster and cleaner because the language is so much nicer than straight C++ (imnsho, anyways, C++ is a turd compared to ObjC).
All the data structures you need to make a program are already made for you, and the GUI classes are intuitive and easy.
So yes, developing under Cocoa will be easy. Carbon apps are slightly harder from what I hear, but I haven't worked with them. However, it's much easier than the dark ages of the older OS 7-8 approach, which was an amazing pain.
- Paradox
Man of the C!!!
Hi. Check out IBM's website. You might actually begin to know what you are talking about. To be quite honest, now that OS X is out, even just a beta, to prove it's not a joke, I'm switching over. I am SICK SICK SICK of x86 Hardware that sucks, blows, and otherwise makes me pay sizable amounts of money for crap.
The G4 boxes (not the cubes) have come down in price. Just because they say "400 mhz" dosen't mean they suck. In fact, the G4's, at present speed, are very competetive on the modern market. Further, apple boxes cost more because they have more cool stuff in them. USB, Firewire, DVD RAM stuff is expensive, but apple makes them available for a reasonable price. Of course they overcharge for RAM, but hey no one is perfect.
You're just upset because you want something for your x86 box. Why? To avoid spending money? Fat chance of that working. To avoid losing linux maybe.. ok, I could see that being a concern. A false one, LinuxPPC is in great shape.
Apple can afford to bide a bit of time here, when everyone comes time to buy their next computer, they'll find for less than $300 more, they can get a far superior piece of hardware and monitor, with a terrific GeekMeetsTrendy OS. I don't see how they can go wrong NOT porting it.
- Paradox
Man of the C!!!
You can make that argument for anything, and it's meaningless. If Hitler hadn't existed, some other tyrant would have taken his place. Riight.
RMS helped to organize the FSF. This has led to more free software under an organizational scheme which is important. I dislike RMS's principle of trying to force people to conform. He is by no means my saint, but he has helped people Accomplish Things. I am compelled therefore to give him some respect. Things I like to use, such as emacs, exist in part because the FSF helped to make them. You can't toss me a hypothetical argument and expect it to hold water.
Of course free software existed before RMS. And of course it will exist without RMS. And yes, a lot of FSF software sucks, I've ranted about this. But we work with what we have. The gcc is an excellent example. It is NOT a cruddy program. Especially since it is free, it is actually a stellar deal.
There may be better compilers, but they are placed out of my reach by prohibitive cost. Lots of developers forget that, being backed by large companies who can afford the liscense fee. What do you want me to do? Go out and buy an SGI and IRIX and their devtools? They're pretty good, better than the gcc anyways. Too bad it costs over $1k just for a year of use.
I'll get right on it. Riiight.
- Paradox
Man of the C!!!
I apologize, I had it wrong. It's been awhile since I used mac os. The last version I touched was 8.0.
- Paradox
Man of the C!!!
The biggest problem with the flood of new developers publishing under the GPL, for GNU, and floods of submissions to Freshmeat.net, is thats a lot of them SUCK. I don't care if fr3d wrote a Perl script to sort his mp3's. Oooh, it's released under the GPL, how trendy.
Has anyone heard a pro developer complain saying, "It's so GNU, it's practically worthless!" ? There is a undercurrent I've discovered of people tired of being told the virtues of software freedom, and trying it only to see that it, quite frankly, sucks sometimes. All these worthless, or neat but poorly written programs, give Free software (by any definition) bad reps. For instance. Look at GNU Emacs. Despite people who dislike emacs in favor of VI, one thing you need to admit is that barring lisp hacking, Emacs dosen't crash. Heck, I've written some pretty wierd stuff in elisp, and I've never once seen it crash, or lost more than a few characters of data, with emacs (with stable versions of course).
Linux is another example. It works, it works well, and fixes it's problems quickly. But unfortunatly in the growing crowd of GPL'ed, BSD'ed,and public domain software, they are exceptions. No one will respect the ideals of GNU, the ideals of the BSD liscense, or the idea that software should be in the hands of and in the control of the USERS, not the developers. Developers only exist to fill the users needs.
I don't care if someone says "Well I'm not getting paid, so don't expect so much." This is a bad attitude, and people like this should STOP MAKING PROGRAMS. Don't release something unless it's usefull. And feel free to release COMPONENT code!!! Libraries are plentiful and easy to modify. Fully working, poorly written programs are less usefull to me than a well written reusable librarly (the GMA library comes to mind).
Also, programmers of Free, Open Sourced, or Public Domain software should APPLY FOR PATENTS and then hold them. Then retire the patent. This way we legally hold work in the public's grasp.
I do not know a service established to help us, lone developers or small project groups, to accomplish this. If there is one, point more people to it!
This is just my personal rant.
- Paradox
Man of the C!!!
I fully agree. Apple has finally gotten off their ass and "Thought Different" for the first time since their shining pinnacle of glory back in the beginning.
This OS is basically what Be strove to be, UNIX for the rest of us.. but it's not just for the rest of us. It's the real deal here folks. At the same time, we get a nice (themeable, did you know that? Macs went themeing many years ago with Kagi Software's Kaleidoscope) GUI and some good graphics performance. Ask anyone whos serious about *NIX, "What's wrong with *NIX?" and they'll say "Xwindows" right off the bat.
So now we have the best of both worlds here. I for one, am shelling out the cash for dual 500mhz G4 as soon as I get my next paycheck. The hardware is better, and now the OS is better too. What more reason do you want? You don't even need to settle for a 1 button mouse!
Just my rant.
- Paradox
Man of the C!!!
"Coming along" meaning what? That one day we might have quantum computers? 'sides. You can make current encryption algorithms work even when there are quantum computers that can do a whole lot of stuff at once in constant time, you just jack up keysizes, prime sizes (in the case of RSA) and the like. At least, this is how I understand it. Imagine if you could generate primes on a real working quantum computer. You could make primes so huge they would foil quantum computers, couldn't you?
But right now quantum computers, with reasonable certainty, don't exist in any useful form.
- Paradox
Man of the C!!!
You know, everyone loves to hate RMS.
But without him, a lot of software we
take for granted today wouldn't be around.
So, maybe sometimes RMS goes a bit off the deep
end, and goes a bit to far in his quest for
software to be totally free. Maybe he has a bad
singing voice, and maybe he's a little scary.
Without him GNU and the GPL (and it's (better
imho) younger sibling the LGPL) woudln't exist.
These things are Good Things that most people
can agree have caused some very nice software
to be free of charge and with source distributed.
So give him a break folks. No one is perfect.
- Paradox
Man of the C!!!
Hi.
Go out and buy a G4, people. Do you have any idea how crappy our beige boxes are, even brand new ones, in comparison with these G4s?
"But Paradox, you're wrong, the G4 only runs at 500mhz!" You say. Please try and remember that mhz dosen't mean all that much. Look at the MIPS, FLOPS, and practical performance. Biiiiig difference. The G4 is such a cool machine I can't imagine why people are buying them en masse. The hardware is just too good to shrug it off with the phrase, "It's a mac." One of apple's new G4 cubes makes a damn sexy linux box. And sans monitor they are really quite cost effective for what you are getting. Linux is in a comparable state for the newer motorolla chips, and all your favorite programs will work so long as you have the source, with a few Loki game ports excepted.
Apple should be hardware company. Their machines kick ass once you unshackle them from MacOS. So stop poking at the big guy (Intel) and rooting for the underdog (AMD) and actally buy a really nice computer based on a very nice RISC design.
- Paradox
Man of the C!!!
Hmm.
Debian needs to fix this problem. Debian has given me nothing but problems, which is why I switched to slackware (and am happy now). I actually left debian because it took me 2, count them 2 weeks to figure out how to shore up debian so X would compile. Every other system I know of, it just works. It's ridiculous.
Compiling X4.0 on any distro but debian is a matter of chaning ProjectRoot (as specified clearly in the INSTALL file, one of the first things it says!) and typing make World &> world.log ; make Install &> install.log ; echo ALL DONE!
You _can_ use the binary distro, but it break's debians overly fragile package system, which means you can't use the cool features apt-get offers you. Anyone using debian is screwed here until SOMEONE working with potato fixes this problem.
Oh, and people who use Linux should expect compilaitions sometimes. It's just The Way Things Are. Probably the way they always will be, and that's not a bad thing, even for end-users who don't code. automake, autoconf.. these make life too easy.
Good luck getting X to work. And remember that Nvidia's drivers fail if you load the DRI modules in your XF86Config file, Nvidia uses DRI through their own modules, because they didn't want to wait for linux to catch up (everyone was complaining).
- Paradox
Man of the C!!!
Well, you could always run opengl apps and see how they handle. Duh!
Maybe it's not the best benchmark, but running simple linux games like quake3, heavy gear 2 and the like are reasonable judges of how well a 3d card will perform under linux.
I have now tried almost every major brand of card on my machine, and without a question, for games and standard opengl, Nvidia's drivers, in the case of the TNT2 and the Geforce2 I've tried, kick butt.
Try running the program "mtri" under your mesa's demo dir. Make sure to link it to the right opengl libs. Like magic, you can see how many triangles your card can push in-window (which is slower than full screen btw).
In window, the Matrox G400 I tried scored about 1m triangles. The Nvidia geforce score more than that, by almost 3.5 times! (taken at 32-bit color 1280x1024 on an athalon t-bird 800 w/ 256 megs pc100 ram, for those who care). The TNT2 scores at just below 2mtri. For reference, the Voodoo3 3000 scored just below the TNT2, at about 1.8 mtri (this test was done in 16-bit color tho, so ymmv).
That's plenty enough benchmarking for me. Code I write runs well with one card, and at 1/4 speed on a Matrox card. Now, don't get me wrong. I love Matrox. I think it's admirable what they are doing, and I plan to buy a g450 w/ dual heading for my workstatin computer because dual monitors make coding easier. But for raw 3d performace, both in games and simulations, nvidia cards give you the best performance you are going to get on an x86 machine, open source or closed.
- Paradox
Man of the C!!!
Only if you have shitty level designers and you don't use curves. The Q3 engine, when used right, makes for some very nice detail, because of the
bezier curve stuff. (It is bezier curves, right?).
'sides.. for man-made structures, usually the wall IS a small little sharp corner. I look at the corner in my office and I don't see a smoothly blending curve, I see a very tight corner. If I look very close I see the curve, but it has to be pretty close!
- Paradox
Man of the C!!!
Fine. Ray Tracing. :P
You still cast rays in ray tracing. Ray casting is just a special case of ray tracing. I hate that stupid semantic separation, logically it makes little sense.
- Paradox
Man of the C!!!
First let me address the old open-driver issue.
Nvidia doesn't make all the tech they use. They purchase rights to use it. These rights do not include the right to tell everyone how all these things work. So Nvidia, whether it would like to open drivers or not, CANNOT. It's not an issue of Nvidia being dicks, it's an issue of Nvidia being legally bound.
Nvidia is going to try and get drivers out for every market they feel is viable. They're a very busy company, so we'll see about the Be thing. Porting it to BSD shouldn't be too hard, so if Nvidia gets enough email saying, "I use FreeBSD, I'll buy your card if you give me a driver" then thats great. But I doubt many freebsd users (I'm one) use freebsd or openbsd or netbsd for games. It just isn't the focus of the OS. FreeBSD is for high performance stuff on PC hardware. OpenBSD is security, NetBSD is just about propagating. Of all of them, FreeBSD x86 is the most promising. But they need a reason.
I, being a graphics programmer who is devout in his belief OpenGL rules, love Nvidia. They make great drivers for the platforms they DO support, and their linux support is, without question, on par with windows support. Futher, as I've said in other posts, their drivers are very "complete." If you're not an opengl programmer, you may know know about opengl extensions and how much better they can make a given image look. OpenGL drivers with a bunch of offical (SGI recognizes them) extensions are easier to use to make fast, realistic effects.
Since these drivers are complete, Nvidia cards can be used for more than just games. They can be used for simulation, medical imaging, satellite image display, a bunch of cool stuff!
Does 3dfx do this? No. And no one makes a technically better card, with better drivers. I wish they could open source and port to every platform, but I'll take performance over an ethical pat on the shoulder any day.
- Paradox
Man of the C!!!
Get off your ass and compile Xfree 4.0. /use/local/X11R4 or something like that. Then adjust your libpaths and PATH variable to look at the new libraries and binaries.
:)
Set the target dir of the install to
Taadaa! Finished.
It works, I've got it working right now.
It's really worth it.
Oh, the Nvidia drivers for Xfree 4.0 are:
A) Very fast. As fast as the detanator2 series when they came out for fullscreen stuff, and almost as fast in windowing. For those who love quake3 framerates at full detail 800x600 I never drop below 80, and peak at 150. At optimized detail with low geometry this lower bound goes to 100, similar peak, and I can do it at 1280x1024!
(w/ an 800mhz athalon thunderbird to handle all the culling, portals, and BSP stuff)
I've coded some old opengl demos that ran like crap on a TNT2, because I didn't know what I was doing and coded poorly. The Geforce2 gts pumps them out smoothly, despite my poor old code. Optimized code runs even faster.
B) Very stable. I've not had one crash yet. The older nvidia drivers, the ones for Xfree3 were buggy and frequently put your videocard in a state that required a full hard reboot. Virtual consoles still mess it up, but only when you have a 3d accelerated window running. This is more than I could say for the Voodoo3 drivers I ran for awhile.
C) Being updated! There are new versions coming out with speed improvements and support for new versions of Xfree4.0. As soon as the 2.4 series becomes the real real reality and settles down a bit, Nvidia promises to have those drivers out.
D) Very Complete! Take a look at the opengl extensins support when the geforce drivers run. Try it! Use quake3 if you don't know the gl call. Under driver info. It has a ton of stuff, including Nvidia's cool register combiners and a bunch of SGIS stuff. Stuff opengl programmers love to see, I assure you.
Get a good tech job and buy the damn card. You know you want to
- Paradox
Man of the C!!!
Make some money and buy a nice video card.
Nice video cards with more features mean my opengl programming produces better results, faster. Screw gaming. With Nvidia's great opengl
drivers EVERYTHING works better. Not just my q3 framerate.
I fail to see any penis issues here. Other than the fact you're being dick.
- Paradox
Man of the C!!!
Just because the japanese use that grammar structure dosen't mean it would affect things. One of the reasons a function header is in front is for indentification purposes. Makes it conveniant to have a nice tag at the start of a block.
Besides, wouldn't compilers have a problem with languages that required that much lookahead? Fast compilers and interpreters usually use predictive parsing and similar stuff for translation. I don't think it would work very well.
Watashi wa Perl ga sukurimas (I use Perl) (I think it's sukao for use, isn't it?) probably isn't the best example anyways. If you look it actually follows a regular pattern that reminds me, at least, of programming. You specify what something is, then what is being done with it.
- Paradox
Man of the C!!!
It dosen't take that much CPU time, you know.
:)
'sides, just breaking a voxel model appart would look ugly! Things don't shatter in nice little pieces that way, they explode in an irregular fashion. You don't need to segment an existing polygonal model, you just add a few more triangles to begin with, or a different model entirely, and billboard up an explosion while you switch the models and send some tank-colored particles everywhere.
It may sound hackier, but it looks pretty good, and it's not very expensive to do. That's what's important in real time rendering.
Nice name btw
- Paradox
Man of the C!!!
Ramp-up-down terrain is a thing of the past or an artifact of bad terrain programming. These days even a mediocre OpenGL or D3D programmer can make a very smooth heightmap. It's easy when 2000 triangles is only a small fraction of what your card can push.
Some algorithms you could look at if you are interested: Quadtree LOD, ROAM, Triangle Bintrees w/out ROAM.
- Paradox
Man of the C!!!
It tessalates them for you, doof. Those aren't openGL calls you are making, they are calls to the GLU which reduces them to polygons for the machine to render.
This isn't acceleration. When cards can be FED a set of NURBS knots and control points, and do it all itself, then it's accelerated.
Look! My card draws quake models in an accelerated fashion. I call myGGLURenderModel() on a string "Ranger" and it draws the ranger model to screen! No. It's not. This Isn't Acceleration.
Massive DUH! on your part. Read what the damn calls do.
- Paradox
Man of the C!!!
Well, voxels are great ways to model landscapes and such, but to make a voxel-based engine with the versatility and freedom of current polygon engines, and one that has similar memory requirements, is a pain. It's a big pain. And it's not even worth it, you don't stand to gain that much. In more limited (non 6dof) situations, voxels can be pretty easy to implement, though.
This is why Commanche used them. Remember Commanche didn't have a full 6dof, and as such it made the voxel engine design easier.
- Paradox
Man of the C!!!
Actually, I was very surprised the other day when I loaded a using java3d, and saw MS's java3d implementation seems to bind into direct3d! Maybe I'm wrong and the applet was just really well designed, but there was dynamic lighting and the thing looked great, even had perspective correct texture mapping and bilinear filtering. It was the coolest thing I've seen. I'm not an MSBoy, but I had to hand it to whever made that decision.
What we need is for someone to patch opengl into a java applet, for all platforms. Hardware acceleration for 3d web content is way too cool an idea.
- Paradox
Man of the C!!!
Problem with this approach is that the ray tree can get arbitrarily deep, in complex scenes. So you don't know with any certainty how long a frame will take to render, or even how long it MIGHT take without directly testing it. Sure, you can do it in parallel, but it's still a huge system of equasions that need to be solved. The benefit of polys and voxels is that they are relativly simple. A triangle, for instantce. Easy to draw a clip.
:)
Everything gets much harder in ray tracing, much harder. And ray tracing means you have to have really accurate models (no blocky models) because you're doing real ray intersections, as such you can't fake things very well with a chunky model. With our current hardware, we can make 300 poly's look like 10000 polys, even under dynamic lighting conditions, in real time. There may be some degenerate cases, but hey, nothing's perfect
- Paradox
Man of the C!!!