Actually, even the DR releases were rock solid. I think eerything since DR8 has been pretty much at near UNIX stability levels. Though I think R3.0 was pretty flaky because it was the first release on x86.
I wonder why you Linux people insist that binary-only is bad. Sure you can't use it on different archs, but what does the vast majority of people running linux have? Maybe for server stuff, but you don't use DVD on a server. If a bunch of people can talk Redhat into getting an NDA and releasing binary-only drivers, then good for them. 99% of the people are happy, and those who whine about the binary-only drivers can just write their own. It still better than what you have now which is 100% of the people dissatisfied. You say that DVD on linux "mostly" works. Thats the problem. With a lot of linux drivers for new stuff, everything "mostly" works, but not much works really well.
I agree with most of your comments except DirectX and Office. DirectX is just not an accelerated graphics API. It is A) A consistant programming interface for the ENTIRE multimedia subsystem. That includeds graphics, sound, input, force feedback, networking, the works. B)It is a means for games to get the OS out of the way. You can write an entire game using very few calls to windows functions. All you have to deal with for windows is threads, interapp comm, and window handling. Everything else is just calls to DirectX. (Which are really fast because they don't have to be as concerned with interacting with other programs as normal API functions do. When you are in full screen DirectDraw, you have precedence over even the GDI and get some obscene amount of proc time.) So don't dis DirectX. And Office is really cool. It just has more features and usability stuff than other word processers. Plus you usually get it free.
Correct me if I'm wrong, didn't DirectDraw (the first direct something) come out of Intel? They then took that concept and started the GameSDK which eventually became DirectX. I don't thing it was purchased from outside, or else D3D wouldn't be as crappy as it is originally.
One word. DirectX. Yea, it may be a little buggy, but it is much better these days. Sure D3D is kinda flaky, but the other APIs are more or less solid. Most of all, it ONE high-performance interface to the system under ONE API. When you learn one DirectX interface, it take very little extra to learn the others. I don't know about some of the earlier OSes (maybe Amiga or NEXT had them) but no modern OS as of yet has this kind of feature. The API essentially pushes aside the OS and make a high performance layer between the app and the hardware. (How else do you think most games work with Windows? Just the latency in the normal sound API would make games sound like some bad Japanese movie.) Even the technology itself is not a big achievment. (DirectX is mainly 80% windows pusing aside code) but the concept.
Lack of shortcuts is inexcusable, but I take offense the the fact that you say, that it may be okay for Mac adherents. Make people hate taking their hands of the keyboard just as much as UNIX people. Believe it or not Mac people aren't all brainless idiots. So, yes it is considered a "bad thing" to have to take your hands of the keyboard, even for Mac users. I could make the same generalized comment about Linux users. (It amazing how many time you actually have to configure something with a SIMPLE dialog box!) But I'd be wrong.
You might be, but many people aren't. I have used the command line, and trust me, there is nothing as nice as the project concept. In Windows, if I were to use the commandline, I would have to type in this. "cl main.cpp libx.cpp liby.cpp etc.cpp morestuff.cpp" For my library, which has about 10 independant source files, it takes quite awhile to type that in. Then I have to type in link, etc.obj dxguid.lib ddraw.lib blah blah blah. In windows it is especially harsh since Win32 is spread out over numerous libraries. (About 10 or 15.) Sure, you can make a makefile, but you can in visual too. Say I add a file to my project. Can you open up the makefile in emacs, edit it to include the new file, go back to the command line, and type make etc before I can right click on "Add to project" and hit the build button? Even quitting out of your editor and typing make etc.cpp takes more time than hitting build (you can do it from a keyboard shortcut too.) So, for all but the simples projects, the IDE approach is faster. And you don't have to use all the built in objects. I write straight C/C++ code, but the integrated environment makes it easier for me to concentrate on the code. In addition, you don't get all the nifty features in Emacs. So far VC++ is the only coding tool I have used that does code completion. Sure you can say it is for weenies who can't keep their variables straight, but then you would be wrong. The Win32 API has functions that take many arguements, and instead of going to the help menu and looking up function XXX, I can just type in the name and as I'm typing a tooltip pops up next to the cursor saying the arguements the function takes. Also, if your style is correct, you should have long, discriptive variable and function names. It is so much easier to type the first few letters then hit enter, than it is to type out the whole thing. BTW, people do use codewarrior. On BeOS it is the interface to GCC and in fact works very well.
Yo, moderators! Can I answer this jackass with some of his own level of crap? BeOS has a niche. Its users who want a powerful, fast, easy to use machine. Linux fills the first, nothing major as of yet fills the second, and macs fill the third. I believe that BeOS is on its 3rd year on Intel hardware, and already has a huge amount of mindshare. Question: Were hardware manufacturers making Linux drivers back in 1994? Don't count Be out just yet.
It is true that Direct3D is pretty bad, but it has gotten much better and AHH, is within the OpenGL ballpark now. And I take it you mean Direct3D don't you. DirectX is a fast set of libraries to access various hardware accelerated devices in a consistant way. Just what Linux DOESNT have. Even Quake 3 uses DirectInput for its input stuff.
Actually, it is not a winmodem situation at all. Shared memory is actually faster for doing somethings where there is a lot of memory traffic on the graphics bus but little on the memory bus. I know SGI uses it on all their machines. As for X support, it depends on AGP, and I don't know X's AGP support. But if it is good as Window's then it should work fine.
Well, you want data? Here is an article from NCES (National center for education statistics.) It is quite a good source of info, and a government organizatoin at that. So have at it.
http://nces.ed.gov/timss/twelfth/
Some particularly interesting exerpts. In mathematics general knowledge, students in the final year of secondary school in 14 countries scored above our twelfth graders (the Netherlands, Sweden, Denmark, Switzerland, Iceland, Norway, France, New Zealand, Australia, Canada, Austria, Slovenia, Germany, and Hungary). Students in 4 countries were not significantly different from ours (Italy, the Russian Federation, Lithuania, and the Czech Republic). Students in two countries (Cyprus and South Africa) performed significantly below students in the United States.
and
One explanation for our low performance that has been suggested in the past is that, because of our diverse population, there is a greater range of scores among U.S. students, and the difference between our lowest-scoring students and our typical student is greater than in many other countries. These low-scoring students, it has been argued, "bring down" the U.S. average. Available information suggests that this is not the case in TIMSS.A
and still more
For example, while a quarter of U.S. students scored 521 or higher, in many high-performing countries half or more of the students had scores that high. Furthermore, the scores of U.S. students at the 95th percentile were similar to those of students at the 75th percentile in some countries. (See Table A2.3 in Appendix 2 for percentiles for mathematics general knowledge; see Tables A2.4, A2.5, and A2.6 for percentiles for the other assessments.)
and some from the science tests.
On the assessment of science general knowledge, students at the end of secondary school in 11 countries (Sweden, the Netherlands, Iceland, Norway, Canada, New Zealand, Australia, Switzerland, Austria, Slovenia, and Denmark) outperformed U.S. twelfth graders. Students in 7 countries performed not significantly different from those in the United States (Germany, France, the Czech Republic, the Russian Federation, Italy, Hungary, and Lithuania). Students in Cyprus and South Africa performed below students in the United States.
So here you go, data. And take a look at some of the test questions. You'll find them to be quite reasonable.
Interesting. Whenever mindcraft comes up, linux users abandon their data and every post begins with, "well in my experience." During the numerous mindcraft posts, a single person had yet to give anything OTHER than anecdotal data. Oh well, just a thought.
Speaking of that, I think that they should remove MacOS from schools. Sure it has a nice GUI, and it is very simple to use even for the stupidest person, but in my opinion it sacrifices a lot of configurability and funcitonallity. I think the lack of a command prompt hurts is the most, because that's what it is simplest to program for >>>>>>>>>>>>>>>> No, thats the exact reason we HAVE mac machines in schools. They were prefrect until people realized they were to niche and expensive for school use. None of what you say makes sense. Get it through your head, PROGRAMMERS ARE NOT THE END ALL BE ALL!!!!!!!!!!! School is for varied experiences. If people want to learn about computers, sure they can. But they should not HAVE to. Unlike math and science and english, programming is not a basic skill. Math is something that is more or less mandetory for anything you do, English is necessary to be a normal human being, and science teache you a mindset which helps problem solving skills. What does programming teach you? How to overcome artificial roadblocks. Thats it. The kind of challanges solved in programming are very unlike the ones encountered in real life. School is for things that are widly applicable so one will know enough so when they get to college to figure out what to do in life, they can chose any field and have atleast a basic grounding in it. Second, computers are tools. Learning how to use a computer is increadibly important, learning how to program one is not. Learning how to drive a car is very important, learning how to service on is not. Third, public computers are supposed to be unconfigurable! The less one has to know about it to make it work the better. MacOS has every bit of functionality that Linux does unless you are running a server. But you aren't are you. What makes me mad is that people like you think that what they are interested in are so important, they should impose it on other people. Sure I like computers, and I like science. But I do not think it is right to impose it on people who don't want or need to know it. I have not gone through a day since I was 8 where I have not had to use english. I have yet to encounter a situation where my programming experience was useful other than when I was programming.
A monolithic kernel DOES have some inherent advantages over a a microkernel, but the advantages of the microkernel outweigh its disatvantages. The whole concept of having servers respond to requests and communicating via IPC has the following advantages. 1. The are much more independant, which leads to more stability and easier coding. (You can make changes in one without chaning another, long as the interface remains the same.) 2. The are much more easily updated and maintained. 3. They are much more asynchronous since objects can make requests, then immediatly return and do some more work. This is especially good in something like a file system server or graphics server. For example, in the BeOS drawing kit, my application can make a request to draw a line. The line functions sends the message and returns immediatly. My app can continue its work while the line drawing occurs in the background. It helps even more in hardware accelerated things, since the server can have the hardware do some rendering, while the app continues to do some physics in the forground or something. All this leads to higher performance. Its true that IPC does incur some overhead, but it can be managed with. I don't know who Be does IPC in BeOS (and I don't think they are telling) but it obviously works well since BeOS apps are extremely fast and responsive. If the IPC overhead was really that bad I don't think BeOS would respond as well as it does in media apps. The other thing that bothers me is that you C programmers seem to think that Object Oriented programming incurs a huge amount of overhead. It does incur some, but it is negligable, and vastly outweighed by the fact that by using systems objects to represent the API, the system is not only easier to program, the API can evolve in time without adding a huge amount of weight to the system. I doubt if the performance hit is even 3 or 4%. And the time it saves can be put to good use optimizing the algorithms used. Finally, object oriented systems are much easier to make extensivly multithreaded, and even on a single processor machine, mulitple threads help because the processor does not stall on one task so long. Especially in an OS, which is mostly limited by the speed of the hard disk. By puting disk access into a seperate thread from the program, it stays MUCH more responsive. So not only CAN an object oriented, microkernel OS be fast, it already exists in the form of BeOS. Although it does bother me that HURD is slow at this point. Even BeOS developer releases were lightning fast, and if HURD is to be such a new OS, why is it slow at the beginning. Or is their a lot of overhead built in? Or maybe it is much farther from release than this article would have you believe. If all this sounds like some mind expanding thing, I urge you to go to Be's website and read the whitepaper on the MediaOS. It is heavily marketing based, but has some really nifty concepts.
Hello! That was "From the article," to qoute your heading. The magazine made that part up, not Casio. It think they were JOKING. J-O-K-I-N-G. Say it with me now. J-O-K-I-N-G.
If it weren't for games, Linux would not be in the place it is now. Hear me out first. What made x86 hardware even in the same ballpark as Suns and SGIs? Games. Games continued to push x86 hardware until one day, someone realized, "Hey I can run a server of this!" No x86 servers, no Linux. Hell, no x86 servers and the small guys would never make it onto the internet.
Becuase the Win32 API sucks. I was doing the low-level primatives for a DirectX drawing library I was doing, since it doesn't offer any drawing tools. Speed did matter because I use it in some of the games I make. (Plus not making the fastest possible code offends my sense of tideness.) I did use some ASM in the end, but only got about 10% out of it. Had I done the entire thing in ASM, A. I would not have been able to maintain it as easily. B. I would have had to learn a lot more ASM, and C. I could not have made it as a class.
No operating systems are not hefty no matter what. Unix people have this idea but it is false. The OS should be a thin abstraction and management layer around the hardware. The Be developers have the right idea, in that the OS should not be this behemoth thing that takes a 20-80% OS tax (in the word of Alex St. John) And closed source has nothing to do with it. BeOS is closed source and whips any other OS on x86 when it comes to speed. The problem is management and vision. Windows is not that much more bloated than Linux (at least not as much the 40M lines makes it look) Linux itself is lean, but the libraries above it are not. But where Linux developers concentrate on making the OS better and improving existing stuff then slowly adding new things, W2K developers just want to add features. In Windows you don't have major rewrites to a subsystem like you do in in Linux. That leads to workaround and general buggy code. Open Source wouldn't help it unless the philosophy changes. So no, W2K is not THAT much more bloated than Linux. And yes some systems are really good (like DirectX.) But due to the urge to add features, the systems are not refined, which leads to the general bugginess of the OS.
Do you know WHY ATI released specs to the Rage Pro? Cuz no one buys it anymore! It has been replaced by the current generation! Why did Matrox release the G400 drivers? Because matrox is a distant competitor in the consumer 3D market and wanted to do something to get the mindshare up. (BTW. I love the G400MAX) The cold reality is that nVidia is a market leader and thus has no real incentive to try to ride the Linux hype. (BTW. Believe it or not, 3Dfx is nowhere near as popular as nVidia. It may top retail sales charts, but most graphics card buying is done through OEM to smart individuals and system manufacturers. nVidia has a HUGE lead here, particularly OEMs like to be able to trumpet all the advanced features it has. Plus Voodoo 3 is basically the same card as Voodoo 1 so it is in the too old to matter bin.) Lastly, maybe nVidia has a good reason not to give register info. If you hadn't noticed, it give a good look into the cards internal workings, and why should nVidia give its superior technology out? Just get X4 done and make binary drivers work well. BeOS does and in its 3rd year on intel hardware, it already has manufacturers making drivers for it. Not only do manufacturer made drivers increase their quality, more drivers get made for different hardware. If you look at windows, the OS sux, but the drivers are really much better polished than Linux drivers usually are. Getting access to driver sources rarely does any good because A) If you are such a manly kernel hacker that you can beat people whose only job is to write drivers for one particular piece of hardware, you'd be working for them. B) You have to weigh the consequences. For every 1 person who ever hacked his own driver to fix a problem, I bet there are 10,000 who wished they had drivers for their card.
Actually, since he is playing Quake, which only gets about a 40% boost with dual, a 700 MHz athlon will be faster. And even more so in non SMP games (which means all except Quake3)
There is so much that bothers me in this and succeding posts I don't even know where to start! First. 1. Linux OpenGL support is crap, since on TNT it is done trought GLX and even on Voodoo it is wrappered through Glide. Glide, however, is pure, unadultered direct rendering. So don't fault nVidia for faults in the OS. 2. GeForce whoops anything 3Dfx has so far. Even a TNT2 Ultra will beat V3 3500, especially if you turn up the detail. 16 bit color and 256 X 256 textures just don't cut it. 3. Sure, dual Voodoo 2s beat a TNT, but not by much. In Windows, only by 30% even in quake, and even less in other games. 4. Geforce is about a full 50% faster in quake 3 than any 3Dfx card, and will only get faster in newer games that use T&L 5. Voodoo 5 6000 will put out 1.3 gigapixels, yes but at what cost? 4 chips and 128 megs of RAM. And thas only 32 meg effective since all textures must be stored multiple times for each chip. I don't think they will fix that problem since it seems to be inherent in SLI (Rage Fury MAXX only has 32 meg effective out of the 64 on the board) If you put 4 GeForces together you get 2 gigapixels per second, and the GeForce Pro (which will be out by the time of Voodoo 5) will be even faster, (probably about 70% like TNT 2 was). 6. In Windows (the only OS that has good 3D, yet) nVidia has a full OpenGL ICD while 3Dfx is lopping along with its lame miniGL driver. (I think they might have an Alpha as hell ICD though) I could go on, but I think I have made myself clear.
Hey buddy, Linux ain't exactly light itself. If you put together the kernel, X, GNOME, GLIBC, GLIB, GTK+, one of the document formatting packages, Qt (for the C++ API and internet services), the Loki game library, all the protocols including IPX, and all the utilities and libraries you need just to get Linux working, you get pretty close to said 40 million lines. You guys don't realize that this is 40 million for the entire OS, not just the kernel. It includes DirectX, Internet explorer, MFC, COM, the database access stuff, plus all the services that need alternate libraries in Linux, but are built into Win32, all the configuration utilities, etc. So yes, windows is fat, but so is Linux. (although to a slightly lesser degree).
Hey buddy, Linux ain't exactly light itself. If you put together the kernel, X, GNOME, GLIBC, GLIB, GTK+, one of the document formatting packages, Qt (for the C++ API and internet services), the Loki game library, all the protocols including IPX, and all the utilities and libraries you need just to get Linux working, you get pretty close to said 40 million lines. You guys don't realize that this is 40 million for the entire OS, not just the kernel. It includes DirectX, Internet explorer, MFC, COM, the database access stuff, all the configuratin utilities, etc. So yes, windows is fat, but so is Linux. (although to a slightly lesser degree).
Slashdot Mirror
Actually, even the DR releases were rock solid. I think eerything since DR8 has been pretty much at near UNIX stability levels. Though I think R3.0 was pretty flaky because it was the first release on x86.
Actually, I belive the Sega logo is much older than the Debian one.
I wonder why you Linux people insist that binary-only is bad. Sure you can't use it on different archs, but what does the vast majority of people running linux have? Maybe for server stuff, but you don't use DVD on a server. If a bunch of people can talk Redhat into getting an NDA and releasing binary-only drivers, then good for them. 99% of the people are happy, and those who whine about the binary-only drivers can just write their own. It still better than what you have now which is 100% of the people dissatisfied. You say that DVD on linux "mostly" works. Thats the problem. With a lot of linux drivers for new stuff, everything "mostly" works, but not much works really well.
I agree with most of your comments except DirectX and Office. DirectX is just not an accelerated graphics API. It is A) A consistant programming interface for the ENTIRE multimedia subsystem. That includeds graphics, sound, input, force feedback, networking, the works. B)It is a means for games to get the OS out of the way. You can write an entire game using very few calls to windows functions. All you have to deal with for windows is threads, interapp comm, and window handling. Everything else is just calls to DirectX. (Which are really fast because they don't have to be as concerned with interacting with other programs as normal API functions do. When you are in full screen DirectDraw, you have precedence over even the GDI and get some obscene amount of proc time.) So don't dis DirectX.
And Office is really cool. It just has more features and usability stuff than other word processers. Plus you usually get it free.
Correct me if I'm wrong, didn't DirectDraw (the first direct something) come out of Intel? They then took that concept and started the GameSDK which eventually became DirectX. I don't thing it was purchased from outside, or else D3D wouldn't be as crappy as it is originally.
One word. DirectX. Yea, it may be a little buggy, but it is much better these days. Sure D3D is kinda flaky, but the other APIs are more or less solid. Most of all, it ONE high-performance interface to the system under ONE API. When you learn one DirectX interface, it take very little extra to learn the others. I don't know about some of the earlier OSes (maybe Amiga or NEXT had them) but no modern OS as of yet has this kind of feature. The API essentially pushes aside the OS and make a high performance layer between the app and the hardware. (How else do you think most games work with Windows? Just the latency in the normal sound API would make games sound like some bad Japanese movie.) Even the technology itself is not a big achievment. (DirectX is mainly 80% windows pusing aside code) but the concept.
Lack of shortcuts is inexcusable, but I take offense the the fact that you say, that it may be okay for Mac adherents. Make people hate taking their hands of the keyboard just as much as UNIX people. Believe it or not Mac people aren't all brainless idiots. So, yes it is considered a "bad thing" to have to take your hands of the keyboard, even for Mac users. I could make the same generalized comment about Linux users. (It amazing how many time you actually have to configure something with a SIMPLE dialog box!) But I'd be wrong.
You might be, but many people aren't. I have used the command line, and trust me, there is nothing as nice as the project concept. In Windows, if I were to use the commandline, I would have to type in this.
"cl main.cpp libx.cpp liby.cpp etc.cpp morestuff.cpp" For my library, which has about 10 independant source files, it takes quite awhile to type that in. Then I have to type in link, etc.obj dxguid.lib ddraw.lib blah blah blah. In windows it is especially harsh since Win32 is spread out over numerous libraries. (About 10 or 15.) Sure, you can make a makefile, but you can in visual too. Say I add a file to my project. Can you open up the makefile in emacs, edit it to include the new file, go back to the command line, and type make etc before I can right click on "Add to project" and hit the build button? Even quitting out of your editor and typing make etc.cpp takes more time than hitting build (you can do it from a keyboard shortcut too.) So, for all but the simples projects, the IDE approach is faster. And you don't have to use all the built in objects. I write straight C/C++ code, but the integrated environment makes it easier for me to concentrate on the code. In addition, you don't get all the nifty features in Emacs. So far VC++ is the only coding tool I have used that does code completion. Sure you can say it is for weenies who can't keep their variables straight, but then you would be wrong. The Win32 API has functions that take many arguements, and instead of going to the help menu and looking up function XXX, I can just type in the name and as I'm typing a tooltip pops up next to the cursor saying the arguements the function takes. Also, if your style is correct, you should have long, discriptive variable and function names. It is so much easier to type the first few letters then hit enter, than it is to type out the whole thing. BTW, people do use codewarrior. On BeOS it is the interface to GCC and in fact works very well.
Yo, moderators! Can I answer this jackass with some of his own level of crap? BeOS has a niche. Its users who want a powerful, fast, easy to use machine. Linux fills the first, nothing major as of yet fills the second, and macs fill the third. I believe that BeOS is on its 3rd year on Intel hardware, and already has a huge amount of mindshare. Question: Were hardware manufacturers making Linux drivers back in 1994? Don't count Be out just yet.
It is true that Direct3D is pretty bad, but it has gotten much better and AHH, is within the OpenGL ballpark now. And I take it you mean Direct3D don't you. DirectX is a fast set of libraries to access various hardware accelerated devices in a consistant way. Just what Linux DOESNT have. Even Quake 3 uses DirectInput for its input stuff.
Actually, it is not a winmodem situation at all. Shared memory is actually faster for doing somethings where there is a lot of memory traffic on the graphics bus but little on the memory bus. I know SGI uses it on all their machines. As for X support, it depends on AGP, and I don't know X's AGP support. But if it is good as Window's then it should work fine.
Well, you want data? Here is an article from NCES (National center for education statistics.) It is quite a good source of info, and a government organizatoin at that. So have at it.
http://nces.ed.gov/timss/twelfth/
Some particularly interesting exerpts.
In mathematics general knowledge, students in the final year of secondary school in 14 countries scored above our twelfth graders (the Netherlands, Sweden, Denmark, Switzerland, Iceland, Norway, France, New Zealand, Australia, Canada, Austria, Slovenia, Germany, and Hungary). Students in 4 countries were not significantly different from ours (Italy, the Russian Federation, Lithuania, and the Czech Republic). Students in two countries (Cyprus and South Africa) performed significantly below students in the United States.
and
One explanation for our low performance that has been suggested in the past is that, because of our diverse population, there is a greater range of scores among U.S. students, and the difference between our lowest-scoring students and our typical student is greater than in many other countries. These low-scoring students, it has been argued, "bring down" the U.S. average. Available information suggests that this is not the case in TIMSS.A
and still more
For example, while a quarter of U.S. students scored 521 or higher, in many high-performing countries half or more of the students had scores that high. Furthermore, the scores of U.S. students at the 95th percentile were similar to those of students at the 75th percentile in some countries. (See Table A2.3 in Appendix 2 for percentiles for mathematics general knowledge; see Tables A2.4, A2.5, and A2.6 for percentiles for the other assessments.)
and some from the science tests.
On the assessment of science general knowledge, students at the end of secondary school in 11 countries (Sweden, the Netherlands, Iceland, Norway, Canada, New Zealand, Australia, Switzerland, Austria, Slovenia, and Denmark) outperformed U.S. twelfth graders. Students in 7 countries performed not significantly different from those in the United States (Germany, France, the Czech Republic, the Russian Federation, Italy, Hungary, and Lithuania). Students in Cyprus and South Africa performed below students in the United States.
So here you go, data. And take a look at some of the test questions. You'll find them to be quite reasonable.
Interesting. Whenever mindcraft comes up, linux users abandon their data and every post begins with, "well in my experience." During the numerous mindcraft posts, a single person had yet to give anything OTHER than anecdotal data. Oh well, just a thought.
Speaking of that, I think that they should remove MacOS from schools. Sure it has a nice GUI, and it is very simple to use even for the stupidest person, but in my opinion it sacrifices a lot of configurability and funcitonallity. I think the lack of a command prompt hurts is the most, because that's what it is simplest to program for
>>>>>>>>>>>>>>>>
No, thats the exact reason we HAVE mac machines in schools. They were prefrect until people realized they were to niche and expensive for school use. None of what you say makes sense. Get it through your head, PROGRAMMERS ARE NOT THE END ALL BE ALL!!!!!!!!!!! School is for varied experiences. If people want to learn about computers, sure they can. But they should not HAVE to. Unlike math and science and english, programming is not a basic skill. Math is something that is more or less mandetory for anything you do, English is necessary to be a normal human being, and science teache you a mindset which helps problem solving skills. What does programming teach you? How to overcome artificial roadblocks. Thats it. The kind of challanges solved in programming are very unlike the ones encountered in real life. School is for things that are widly applicable so one will know enough so when they get to college to figure out what to do in life, they can chose any field and have atleast a basic grounding in it. Second, computers are tools. Learning how to use a computer is increadibly important, learning how to program one is not. Learning how to drive a car is very important, learning how to service on is not. Third, public computers are supposed to be unconfigurable! The less one has to know about it to make it work the better. MacOS has every bit of functionality that Linux does unless you are running a server. But you aren't are you. What makes me mad is that people like you think that what they are interested in are so important, they should impose it on other people. Sure I like computers, and I like science. But I do not think it is right to impose it on people who don't want or need to know it. I have not gone through a day since I was 8 where I have not had to use english. I have yet to encounter a situation where my programming experience was useful other than when I was programming.
A monolithic kernel DOES have some inherent advantages over a a microkernel, but the advantages of the microkernel outweigh its disatvantages. The whole concept of having servers respond to requests and communicating via IPC has the following advantages. 1. The are much more independant, which leads to more stability and easier coding. (You can make changes in one without chaning another, long as the interface remains the same.) 2. The are much more easily updated and maintained. 3. They are much more asynchronous since objects can make requests, then immediatly return and do some more work. This is especially good in something like a file system server or graphics server. For example, in the BeOS drawing kit, my application can make a request to draw a line. The line functions sends the message and returns immediatly. My app can continue its work while the line drawing occurs in the background. It helps even more in hardware accelerated things, since the server can have the hardware do some rendering, while the app continues to do some physics in the forground or something. All this leads to higher performance. Its true that IPC does incur some overhead, but it can be managed with. I don't know who Be does IPC in BeOS (and I don't think they are telling) but it obviously works well since BeOS apps are extremely fast and responsive. If the IPC overhead was really that bad I don't think BeOS would respond as well as it does in media apps. The other thing that bothers me is that you C programmers seem to think that Object Oriented programming incurs a huge amount of overhead. It does incur some, but it is negligable, and vastly outweighed by the fact that by using systems objects to represent the API, the system is not only easier to program, the API can evolve in time without adding a huge amount of weight to the system. I doubt if the performance hit is even 3 or 4%. And the time it saves can be put to good use optimizing the algorithms used. Finally, object oriented systems are much easier to make extensivly multithreaded, and even on a single processor machine, mulitple threads help because the processor does not stall on one task so long. Especially in an OS, which is mostly limited by the speed of the hard disk. By puting disk access into a seperate thread from the program, it stays MUCH more responsive. So not only CAN an object oriented, microkernel OS be fast, it already exists in the form of BeOS. Although it does bother me that HURD is slow at this point. Even BeOS developer releases were lightning fast, and if HURD is to be such a new OS, why is it slow at the beginning. Or is their a lot of overhead built in? Or maybe it is much farther from release than this article would have you believe. If all this sounds like some mind expanding thing, I urge you to go to Be's website and read the whitepaper on the MediaOS. It is heavily marketing based, but has some really nifty concepts.
Hello! That was "From the article," to qoute your heading. The magazine made that part up, not Casio. It think they were JOKING. J-O-K-I-N-G. Say it with me now. J-O-K-I-N-G.
If it weren't for games, Linux would not be in the place it is now. Hear me out first. What made x86 hardware even in the same ballpark as Suns and SGIs? Games. Games continued to push x86 hardware until one day, someone realized, "Hey I can run a server of this!" No x86 servers, no Linux. Hell, no x86 servers and the small guys would never make it onto the internet.
Becuase the Win32 API sucks. I was doing the low-level primatives for a DirectX drawing library I was doing, since it doesn't offer any drawing tools. Speed did matter because I use it in some of the games I make. (Plus not making the fastest possible code offends my sense of tideness.) I did use some ASM in the end, but only got about 10% out of it. Had I done the entire thing in ASM, A. I would not have been able to maintain it as easily. B. I would have had to learn a lot more ASM, and C. I could not have made it as a class.
No operating systems are not hefty no matter what. Unix people have this idea but it is false. The OS should be a thin abstraction and management layer around the hardware. The Be developers have the right idea, in that the OS should not be this behemoth thing that takes a 20-80% OS tax (in the word of Alex St. John) And closed source has nothing to do with it. BeOS is closed source and whips any other OS on x86 when it comes to speed. The problem is management and vision. Windows is not that much more bloated than Linux (at least not as much the 40M lines makes it look) Linux itself is lean, but the libraries above it are not. But where Linux developers concentrate on making the OS better and improving existing stuff then slowly adding new things, W2K developers just want to add features. In Windows you don't have major rewrites to a subsystem like you do in in Linux. That leads to workaround and general buggy code. Open Source wouldn't help it unless the philosophy changes. So no, W2K is not THAT much more bloated than Linux. And yes some systems are really good (like DirectX.) But due to the urge to add features, the systems are not refined, which leads to the general bugginess of the OS.
Do you know WHY ATI released specs to the Rage Pro? Cuz no one buys it anymore! It has been replaced by the current generation! Why did Matrox release the G400 drivers? Because matrox is a distant competitor in the consumer 3D market and wanted to do something to get the mindshare up. (BTW. I love the G400MAX) The cold reality is that nVidia is a market leader and thus has no real incentive to try to ride the Linux hype. (BTW. Believe it or not, 3Dfx is nowhere near as popular as nVidia. It may top retail sales charts, but most graphics card buying is done through OEM to smart individuals and system manufacturers. nVidia has a HUGE lead here, particularly OEMs like to be able to trumpet all the advanced features it has. Plus Voodoo 3 is basically the same card as Voodoo 1 so it is in the too old to matter bin.) Lastly, maybe nVidia has a good reason not to give register info. If you hadn't noticed, it give a good look into the cards internal workings, and why should nVidia give its superior technology out? Just get X4 done and make binary drivers work well. BeOS does and in its 3rd year on intel hardware, it already has manufacturers making drivers for it. Not only do manufacturer made drivers increase their quality, more drivers get made for different hardware. If you look at windows, the OS sux, but the drivers are really much better polished than Linux drivers usually are. Getting access to driver sources rarely does any good because A) If you are such a manly kernel hacker that you can beat people whose only job is to write drivers for one particular piece of hardware, you'd be working for them. B) You have to weigh the consequences. For every 1 person who ever hacked his own driver to fix a problem, I bet there are 10,000 who wished they had drivers for their card.
Actually, since he is playing Quake, which only gets about a 40% boost with dual, a 700 MHz athlon will be faster. And even more so in non SMP games (which means all except Quake3)
There is so much that bothers me in this and succeding posts I don't even know where to start! First. 1. Linux OpenGL support is crap, since on TNT it is done trought GLX and even on Voodoo it is wrappered through Glide. Glide, however, is pure, unadultered direct rendering. So don't fault nVidia for faults in the OS. 2. GeForce whoops anything 3Dfx has so far. Even a TNT2 Ultra will beat V3 3500, especially if you turn up the detail. 16 bit color and 256 X 256 textures just don't cut it. 3. Sure, dual Voodoo 2s beat a TNT, but not by much. In Windows, only by 30% even in quake, and even less in other games. 4. Geforce is about a full 50% faster in quake 3 than any 3Dfx card, and will only get faster in newer games that use T&L 5. Voodoo 5 6000 will put out 1.3 gigapixels, yes but at what cost? 4 chips and 128 megs of RAM. And thas only 32 meg effective since all textures must be stored multiple times for each chip. I don't think they will fix that problem since it seems to be inherent in SLI (Rage Fury MAXX only has 32 meg effective out of the 64 on the board) If you put 4 GeForces together you get 2 gigapixels per second, and the GeForce Pro (which will be out by the time of Voodoo 5) will be even faster, (probably about 70% like TNT 2 was). 6. In Windows (the only OS that has good 3D, yet) nVidia has a full OpenGL ICD while 3Dfx is lopping along with its lame miniGL driver. (I think they might have an Alpha as hell ICD though) I could go on, but I think I have made myself clear.
Hey buddy, Linux ain't exactly light itself. If you put together the kernel, X, GNOME, GLIBC, GLIB, GTK+, one of the document formatting packages, Qt (for the C++ API and internet services), the Loki game library, all the protocols including IPX, and all the utilities and libraries you need just to get Linux working, you get pretty close to said 40 million lines. You guys don't realize that this is 40 million for the entire OS, not just the kernel. It includes DirectX, Internet explorer, MFC, COM, the database access stuff, plus all the services that need alternate libraries in Linux, but are built into Win32, all the configuration utilities, etc. So yes, windows is fat, but so is Linux. (although to a slightly lesser degree).
Hey buddy, Linux ain't exactly light itself. If you put together the kernel, X, GNOME, GLIBC, GLIB, GTK+, one of the document formatting packages, Qt (for the C++ API and internet services), the Loki game library, all the protocols including IPX, and all the utilities and libraries you need just to get Linux working, you get pretty close to said 40 million lines. You guys don't realize that this is 40 million for the entire OS, not just the kernel. It includes DirectX, Internet explorer, MFC, COM, the database access stuff, all the configuratin utilities, etc. So yes, windows is fat, but so is Linux. (although to a slightly lesser degree).
As I recall PC66 RAM is like 12 or 10 ns, and PC100 goes down to like 8 or 9 and PC133 goes all the way to 7ns.