but it's not immoral what they did, just ruthless business practice. Sucks that smaller companies are not as ruthless as MS, but that's the way it goes.
Anyone who tries to compete with MS knows the rules, they haven't changed them since the DAY they were founded. If you know your MS history, they have acted in a consistant manner.
It's the companies fault if they only made one product and never diversified what they made. BEOS was a great product, but it was not good enough to switch to. The drawbacks were more then the advantages. Netscape was an example of illegal but not immoral practices.
Sorry that you love computers and hate MS, it's a fact of life, deal with it. It's like loving cars and hating roads. If you think that because some people got fired through MSes actions makes them 'immoral' then I would venture to state that through their actions, more people then ever have jobs in the IT industry. If they were the number one server OS then almost everyone would have a job because it would take that many people to keep that infastructure going!
I added a turbo switch to my TI-85, fun stuff. I'm just saying that what he wants s/he can already get, and have it look and perform much better. Since s/he won't have to hack any hardware (only software) and won't have to reboot.
Personally I can't stand rebooting my ps2 from normal to linux mode to do programming with it. I really want to get another one but I'm not going to spend the money yet.
Can't you do that now if you buy a video card with a TV out? I don't get it. I agree it's neat to be able to use an xbox for more then it was intended, but come on.
What does morals have to do with anything? Why should that even enter into the question at all? Linux isn't moral or immoral it has nothing to do with that at all. You won't be judged by anyone on which operating system you chose to put on your xbox. You won't get rejected from the gates of heaven for modding your xbox.
Besides, how does this affect you except on a superficial level?
but don't forget that they want people who can do more then one thing. It's stupid to hire people that can do one and only one thing well in law enforcement. These are the people who need to be able to do a multitude of things to 'survive' at times.
They are just saying that you should be able to do many, things and may be required to be a 'normal' agent from time to time. If they actually get what they want is another matter altogether. Maybe the computer job pays really really well compared to a normal agent?
Java already does what you are asking for. Today it happens when you run a program in most modern jvms. Even the IE JVM compiles your class file to native code (JIT from symantec).
ARM has an extension for their chips that allows them to run many (not all) java bytecodes as native instructions.
Structs? Structs in C# are nothing like structs in C++. Structs in C++ are classes. The only difference is that all members are declared public in structs, private in classes (by default). C# structs are objects that live on the stack as opposed to the heap. Don't confuse the two, please.
C++ doesn't have unsafe code. It's just code. the JVM/CLR has a pair of special goggles that slows down their view of the world in order to make it 'safe'. You can turn off the array bounds checking in Java (with certain VMs) for example.
While I guess you can call C++ a layer about C, it's really not. It's really a new language, that supports all the features of C and compiles C code. A superset if you will, as it supports many many more as well. I think that it's marketing that has forced C++ to be percieved so close to C. (Not always a bad thing)
Also you cannot multiple inherit in C#. It's not supported nor should it be. It creates too many problems if you want to be able to always let people inherit your classes.
Garbage Collection is not only avaliable for C#/Java enviroments. You can easily use a GC in C/C++. Just do a search on Google and you will find a free implementation. Yes is it supported by the 'language', but I think that people confuse the language with the runtime.
What makes java so easy to debug and test is the fact that the runtime is not native. An emulated platform (CLR/JVM) is much easier to debug and see the state of since if you want more debugging imformation you just have to change the software and not the hardware. Many problems with debugging machine level code arise from the fact that the processors today throw away any debugging information because it would slow down the processor. A VM can run in a 'debug' mode or not.
Which shortcomings of java are addressed in C#? Can you bind to C/C++ methods without writing C/C++ code? I really don't know the answer to that question.
Wrong. Case in point. I just implemented a complex GUI with a hundred or so wigets, including a JTree with over a thousand nodes on average, and the UI is *extremely* responsive on a PIII 600, and usable on a PII Celeron 450. Both systems are considered slow these days. It's frightening how pervasive the "Java is slow" really is. Do a Google on the subject, and you'll see that in many cases, C/C++ is as fast or slower than Java with Hotspot. I didn't say that Swing is ALWAYS slow, I said that Swing is not as responsive as just using AWT and coding everything yourself or building an application in the native widget set of the platform. Java is not slow, the APIs that you use to access the underlying machine are. Swing is slower (then native) because it uses no native platform widgets, it draws to a buffer that is then drawn to the screen.
As far as the standard library is concerned, being big is both a blessing and a curse. It's great because you have everything, and it sucks because you have EVERYTHING. Corba is nice, but should there be a way to take it out if I don't want it? Same with the SQL packages. Now they include the XML parser by default, what if I don't like that either. This is not a big deal if you expect the client to have a JVM installed already on their system. But it is much easier to package a JRE with your code to deploy it. Unfortunatly, you can't delete all the classes that you don't need easily (AFAIK). Abstraction is good, but sometimes I wonder if they went a little overboard.
The speed issue is crazy because there are so many things to talk about. I'd say that in every task except GUI, java is equal to a similar C++ program. However, when dealing with the AWT/Swing libraries, you will get reduced performace compared to a C++ program. You will probably be able to write a bug free program faster with Swing then with the MFC or Win32 apis. But if you are developing a large application and responsiveness is a requirement, then I'd say that C++ is better. (Speed and responsiveness are not the same thing)
Native extensions: JNI works great, jni works fine. But making a quality desktop application requires that you integrate with the desktop sometimes. Unless you have JNI Java can't do lots of things on the desktop, it can't put little icons on the Windows toolbar, it can't integrate with the shell to provide services, it can't be a COM object. It can't run as a service without JNI. It can't see extended file attributes like ownership, security permissions and under NTFS the other streams (a file can have multiple data streams in NTFS, they are not well used, but they exist). All of these features require you to write a JNI extension. Java can't go and modify the registry, ( though I believe that in 1.4 they store preferences in the reg for you automagically).
I don't mean to bash so much, but saying that Java doesn't have problems is just wrong. It has problems with responsiveness (anyone who has had their weblogic server garbage collect for hours at a time knows this or used forte), it has problems with bloat ( you can't reduce the size easily) and it has problems with integration ( yes, integration is a 'good thing'). Java on the client is not dead but it has a long way to go before it can be called a replacement for C/C++.
But having written a large app in Foo, does prove that it can be done in Foo. Which I think is the point. Has anyone ever written a large scale DB in Perl? I dunno, but until someone does, it's unproven. Looking at what other people have done and how hard it was for them to implement it gives you an idea of what toolset to use when creating something new. (Or just reimplementing something)
(disclamer, I've not actually read the article) If there is a reason that java died on the client it is becuase still to this day, if you wish to make a program that is competive with a normal C/C++ client, it requires the same amount of work. There is no advantage to using Java, in fact many times you end up doing more work to get the same job done.
If you want speed, then you can't use Swing. Swing is nowhere near as responsive to the user as a AWT or even faster, a native API application. Though it is much faster to develop with then Win32 API. So unless you want your application to be percieved as 'slow' you have to do many tricks that are difficult and hack-like.
Java's strengths are that it's easier to use thanks to GC, an OO api (that is usually extendable) and a large standard API. Some weaknesses are bad memory model, API is too large and abstract (in some places) and lack of support for native extensions. No matter what you say, it sucks that you have to get out a C compiler anytime you want to interface with a part of the system that the Java API designers didn't design in. Another thing that I don't like is that I think that they should do a code audit and get rid of all depreciated methods.
Java has killed itself on the client, not anyone else. Applets were shoehorned into the community of non-developers. Programmers may have made the tools that make the web possiable, but they did not make the web what it is today. Flash is made for tech-artists, those people who don't really want to program, but can get enough done to get what they imagine on the screen. Java never filled that niche, it's a 'real' programming language that can display stuff in a browser.
I use Java every day, and I like it but I know its pitfalls all too well. Having to make up for them is what we Java developers do.
Damn straight, why go with a full EJB container when all you need is a servlet engine. Tomcat is the best engine. Resin is supposed to be faster, but it's not. Under heavy heavy load, both cope slightly differently, with a slight edge to resin. Though resin does give weird errors sometimes under heavy load.
I think that it comes down to more of the old fashioned economy of scale.
If there were no patents, then a larger firm with a higher efficiency would be able to steal a smaller firm's design, and produce it at a lower cost. It's much cheaper to make a million of something if you have the ability to do so.
With patents, the larger firm is forced to license the design from the smaller firm to make it.
Actually the idea of a patent is to grant a temporary monopoly to someone with the idea of luring people to do things who would otherwise not spend their time and enegry on such pursuits. Yes, it's about greed and money, but the end result is supposed to be that people spend their time doing things like advancing technology. Before people had to hide their ideas from other people because they would lose their competive edge if anyone found out the secret. Now people publish their ideas without fear of people stealing them since they are protected by the law.
Many ideas were lost because the only person who knew them died without writing them down or passing them on. They might have written them down if they knew that they were protected by the gov.
Intel will make sure that they realease a binary that is optimized for the P4 with one of their compilers. The 'real' intel C/C++ compilers are always faster then the MS or GCC compilers when optimizing for their chips. AMD AFAIK doesn't have a compiler suite that optimizes for their chips.
Total paperless is just not a possiability yet. The thing that we need to go total paperless is a digital system that we can use to verify everyone's vote that is tamper proof.
Say that everyone carries around a disk of somesort that contains all the reciepts of everything we buy, and do. It is a digital replacement for the reciepts that we all carry around today. Whenever we buy something at the store or vote or do whatever we can put this in and it will copy it to the disk. Keeping a copy there for our later investigation. Also the company that we interact with keeps a copy and there is some sort of security key that can check to see if the 2 copies are the same. So basically there should be a way to see if you change your copy or if they change their copy. So somehow we can prove forgery at either end. AFAIK, this problem has yet to be solved but if we did have a way to detect forgery then I would trust in a totatly digital solution. It's already too easy to forge the current system.
Though they blame the mistake on the Windows OS for crashing, it's stupid to believe them. Basically whomever decided to push this out there didn't test their product enough. Everyone who has used computers for any length of time realize that a closed system like this should have zero problems if properly tested no matter what the underlying software is, be it windows, linux, HP-UX, or mac.
What does this really mean? That the voting system should go back for yet more testing. QAing software is probably the most boring part of the job, but it's also the more important. If we are to even pretend that we live in a fair society then any voting system should work and work fairly. Be it paper or computer based.
Is America ready for a computer based system? I think a computer based system should be able to replace a paper based system. I think that possiably we should also use paper in addition to the computer system, meaning that they should actually print a reciept of your vote so that in the case of a recount, they have physical proof that you voted for (Gore and not Bush?) the person you said that you did.
I wasn't aware that you could extend Direct3D, I guess that you'd need a seperate API from the card manufacturer or something right? OpenGL assumes that each card will have it's own extensions and thus provides a common extension mechanism.
The reason that DirectX has been able to grow so quickly is that MS has kept complete control over the platform, only with versions 8 and higher have they actually let the other companies (read nVidia and ATI) have a real say in the api it seems like. I know that they always consulted the card manufacturers, but they still had the final say.
Contrasted to OpenGL where it has taken forever to get version 2.0 out because there are so many people with so many opinions to consider. Open apis are just slower to develop, but in the end, hopefully better since the requirements are more diverse.
I think that you and the parent are missing the real reason that OpenGL is king on the workstations and Direct3D never will be. OpenGL is extendable by everyone and Direct3D is completly controlled by MS.
You can't go and make your own version of Direct3D with a hypercube (or whatever) extension that draws a super widget because Direct3D is closed. OpenGL is completly open and a vendor of 3-D hardware can add whatever they want to the API. The programmers have to then use the extension, but for workstations where the constraints are defined this is usually not a problem.
PC Games on the other hand have to run on the most diverse set of hardware ever. From the lowest 90Mhz Pentium to the newest 2600+ Athlon, with a range of video cards to make a driver writer weak at the knees. (they don't have to run well, just run) D3D solves this by making all functions static and creating software implementations of them. So you program to a 'version' of the API and you know exactly what the machine is capable of and if you can't get an accelerated version of the functionality, then at least you don't have to create your own. Great for game writers, but not as useful for workstation class applications, since they are more concerned with accuracy then games are (in general).
OpenGL is better when the HW that you want to support is smaller, since you can get better performance out of it. Direct3D is better when you just want it to run without having to write a ton of code to emulate the top level functionality.
Solid State Hard Drives exist, with IDE interfaces. They are just expensive for what you get. What are you going to do? You can't use DRAM without a large battery to back it up when the system is off. You can't have a Buffer the size of the drive, that's silly as it's better to just add more system RAM. Basically, it's always better to add more system ram rather then make the HD really fast and solid state. System ram is used to buffer the HD when usage is low and can actually be used by the programs (you know, what you want to use the computer for) when it needs to be. I've got a gig and a half of ram in my machine, I could buffer an entire CD in ram but I don't have to.
Just another slow news day at Tom's. It's not like you can actually go and buy these things yet. Even if you could, would you unless you were going to buy a whole new computer anyway? Does anyone want to throw out their old drives and replace them all with new ones just because they exist? It will be nice to get rid of those huge cables though.
Slashdot Mirror
yeah, his music is one step removed from musak. Any more and it'll be in elevators for old people everywhere. (if it's not already).
but it's not immoral what they did, just ruthless business practice. Sucks that smaller companies are not as ruthless as MS, but that's the way it goes.
Anyone who tries to compete with MS knows the rules, they haven't changed them since the DAY they were founded. If you know your MS history, they have acted in a consistant manner.
It's the companies fault if they only made one product and never diversified what they made.
BEOS was a great product, but it was not good enough to switch to. The drawbacks were more then the advantages.
Netscape was an example of illegal but not immoral practices.
Sorry that you love computers and hate MS, it's a fact of life, deal with it. It's like loving cars and hating roads.
If you think that because some people got fired through MSes actions makes them 'immoral' then I would venture to state that through their actions, more people then ever have jobs in the IT industry. If they were the number one server OS then almost everyone would have a job because it would take that many people to keep that infastructure going!
I added a turbo switch to my TI-85, fun stuff. I'm just saying that what he wants s/he can already get, and have it look and perform much better. Since s/he won't have to hack any hardware (only software) and won't have to reboot.
Personally I can't stand rebooting my ps2 from normal to linux mode to do programming with it. I really want to get another one but I'm not going to spend the money yet.
Can't you do that now if you buy a video card with a TV out? I don't get it. I agree it's neat to be able to use an xbox for more then it was intended, but come on.
Strawmen left and right.
What does morals have to do with anything? Why should that even enter into the question at all? Linux isn't moral or immoral it has nothing to do with that at all. You won't be judged by anyone on which operating system you chose to put on your xbox. You won't get rejected from the gates of heaven for modding your xbox.
Besides, how does this affect you except on a superficial level?
Because they can.
but don't forget that they want people who can do more then one thing. It's stupid to hire people that can do one and only one thing well in law enforcement. These are the people who need to be able to do a multitude of things to 'survive' at times.
They are just saying that you should be able to do many, things and may be required to be a 'normal' agent from time to time. If they actually get what they want is another matter altogether.
Maybe the computer job pays really really well compared to a normal agent?
Java already does what you are asking for. Today it happens when you run a program in most modern jvms. Even the IE JVM compiles your class file to native code (JIT from symantec).
ARM has an extension for their chips that allows them to run many (not all) java bytecodes as native instructions.
The future is now!
Structs? Structs in C# are nothing like structs in C++. Structs in C++ are classes. The only difference is that all members are declared public in structs, private in classes (by default).
C# structs are objects that live on the stack as opposed to the heap. Don't confuse the two, please.
C++ doesn't have unsafe code. It's just code. the JVM/CLR has a pair of special goggles that slows down their view of the world in order to make it 'safe'. You can turn off the array bounds checking in Java (with certain VMs) for example.
While I guess you can call C++ a layer about C, it's really not. It's really a new language, that supports all the features of C and compiles C code. A superset if you will, as it supports many many more as well. I think that it's marketing that has forced C++ to be percieved so close to C. (Not always a bad thing)
Also you cannot multiple inherit in C#. It's not supported nor should it be. It creates too many problems if you want to be able to always let people inherit your classes.
Garbage Collection is not only avaliable for C#/Java enviroments. You can easily use a GC in C/C++. Just do a search on Google and you will find a free implementation. Yes is it supported by the 'language', but I think that people confuse the language with the runtime.
What makes java so easy to debug and test is the fact that the runtime is not native. An emulated platform (CLR/JVM) is much easier to debug and see the state of since if you want more debugging imformation you just have to change the software and not the hardware. Many problems with debugging machine level code arise from the fact that the processors today throw away any debugging information because it would slow down the processor. A VM can run in a 'debug' mode or not.
Which shortcomings of java are addressed in C#? Can you bind to C/C++ methods without writing C/C++ code? I really don't know the answer to that question.
is the in jokes. The discussion on here is nowhere near as good as the jokes. Too bad nobody would get it.
Wrong. Case in point. I just implemented a complex GUI with a hundred or so wigets, including a JTree with over a thousand nodes on average, and the UI is *extremely* responsive on a PIII 600, and usable on a PII Celeron 450. Both systems are considered slow these days. It's frightening how pervasive the "Java is slow" really is. Do a Google on the subject, and you'll see that in many cases, C/C++ is as fast or slower than Java with Hotspot.
I didn't say that Swing is ALWAYS slow, I said that Swing is not as responsive as just using AWT and coding everything yourself or building an application in the native widget set of the platform. Java is not slow, the APIs that you use to access the underlying machine are. Swing is slower (then native) because it uses no native platform widgets, it draws to a buffer that is then drawn to the screen.
As far as the standard library is concerned, being big is both a blessing and a curse. It's great because you have everything, and it sucks because you have EVERYTHING. Corba is nice, but should there be a way to take it out if I don't want it? Same with the SQL packages. Now they include the XML parser by default, what if I don't like that either. This is not a big deal if you expect the client to have a JVM installed already on their system. But it is much easier to package a JRE with your code to deploy it. Unfortunatly, you can't delete all the classes that you don't need easily (AFAIK). Abstraction is good, but sometimes I wonder if they went a little overboard.
The speed issue is crazy because there are so many things to talk about. I'd say that in every task except GUI, java is equal to a similar C++ program. However, when dealing with the AWT/Swing libraries, you will get reduced performace compared to a C++ program. You will probably be able to write a bug free program faster with Swing then with the MFC or Win32 apis. But if you are developing a large application and responsiveness is a requirement, then I'd say that C++ is better. (Speed and responsiveness are not the same thing)
Native extensions:
JNI works great, jni works fine. But making a quality desktop application requires that you integrate with the desktop sometimes. Unless you have JNI Java can't do lots of things on the desktop, it can't put little icons on the Windows toolbar, it can't integrate with the shell to provide services, it can't be a COM object. It can't run as a service without JNI. It can't see extended file attributes like ownership, security permissions and under NTFS the other streams (a file can have multiple data streams in NTFS, they are not well used, but they exist). All of these features require you to write a JNI extension. Java can't go and modify the registry, ( though I believe that in 1.4 they store preferences in the reg for you automagically).
I don't mean to bash so much, but saying that Java doesn't have problems is just wrong. It has problems with responsiveness (anyone who has had their weblogic server garbage collect for hours at a time knows this or used forte), it has problems with bloat ( you can't reduce the size easily) and it has problems with integration ( yes, integration is a 'good thing'). Java on the client is not dead but it has a long way to go before it can be called a replacement for C/C++.
But having written a large app in Foo, does prove that it can be done in Foo. Which I think is the point. Has anyone ever written a large scale DB in Perl? I dunno, but until someone does, it's unproven. Looking at what other people have done and how hard it was for them to implement it gives you an idea of what toolset to use when creating something new. (Or just reimplementing something)
(disclamer, I've not actually read the article)
If there is a reason that java died on the client it is becuase still to this day, if you wish to make a program that is competive with a normal C/C++ client, it requires the same amount of work. There is no advantage to using Java, in fact many times you end up doing more work to get the same job done.
If you want speed, then you can't use Swing. Swing is nowhere near as responsive to the user as a AWT or even faster, a native API application. Though it is much faster to develop with then Win32 API. So unless you want your application to be percieved as 'slow' you have to do many tricks that are difficult and hack-like.
Java's strengths are that it's easier to use thanks to GC, an OO api (that is usually extendable) and a large standard API. Some weaknesses are bad memory model, API is too large and abstract (in some places) and lack of support for native extensions. No matter what you say, it sucks that you have to get out a C compiler anytime you want to interface with a part of the system that the Java API designers didn't design in. Another thing that I don't like is that I think that they should do a code audit and get rid of all depreciated methods.
Java has killed itself on the client, not anyone else. Applets were shoehorned into the community of non-developers. Programmers may have made the tools that make the web possiable, but they did not make the web what it is today. Flash is made for tech-artists, those people who don't really want to program, but can get enough done to get what they imagine on the screen. Java never filled that niche, it's a 'real' programming language that can display stuff in a browser.
I use Java every day, and I like it but I know its pitfalls all too well. Having to make up for them is what we Java developers do.
Damn straight, why go with a full EJB container when all you need is a servlet engine.
Tomcat is the best engine. Resin is supposed to be faster, but it's not. Under heavy heavy load, both cope slightly differently, with a slight edge to resin. Though resin does give weird errors sometimes under heavy load.
Are you management? What's a manager doing on /.?
How in the world are you useing // in C? C doesn't support anything but /* */ comments...
(But really I joke, I joke.)
I think that it comes down to more of the old fashioned economy of scale.
If there were no patents, then a larger firm with a higher efficiency would be able to steal a smaller firm's design, and produce it at a lower cost. It's much cheaper to make a million of something if you have the ability to do so.
With patents, the larger firm is forced to license the design from the smaller firm to make it.
Actually the idea of a patent is to grant a temporary monopoly to someone with the idea of luring people to do things who would otherwise not spend their time and enegry on such pursuits. Yes, it's about greed and money, but the end result is supposed to be that people spend their time doing things like advancing technology. Before people had to hide their ideas from other people because they would lose their competive edge if anyone found out the secret. Now people publish their ideas without fear of people stealing them since they are protected by the law.
Many ideas were lost because the only person who knew them died without writing them down or passing them on. They might have written them down if they knew that they were protected by the gov.
Intel will make sure that they realease a binary that is optimized for the P4 with one of their compilers. The 'real' intel C/C++ compilers are always faster then the MS or GCC compilers when optimizing for their chips. AMD AFAIK doesn't have a compiler suite that optimizes for their chips.
Total paperless is just not a possiability yet. The thing that we need to go total paperless is a digital system that we can use to verify everyone's vote that is tamper proof.
Say that everyone carries around a disk of somesort that contains all the reciepts of everything we buy, and do. It is a digital replacement for the reciepts that we all carry around today. Whenever we buy something at the store or vote or do whatever we can put this in and it will copy it to the disk. Keeping a copy there for our later investigation. Also the company that we interact with keeps a copy and there is some sort of security key that can check to see if the 2 copies are the same.
So basically there should be a way to see if you change your copy or if they change their copy. So somehow we can prove forgery at either end.
AFAIK, this problem has yet to be solved but if we did have a way to detect forgery then I would trust in a totatly digital solution. It's already too easy to forge the current system.
Though they blame the mistake on the Windows OS for crashing, it's stupid to believe them. Basically whomever decided to push this out there didn't test their product enough. Everyone who has used computers for any length of time realize that a closed system like this should have zero problems if properly tested no matter what the underlying software is, be it windows, linux, HP-UX, or mac.
What does this really mean? That the voting system should go back for yet more testing. QAing software is probably the most boring part of the job, but it's also the more important. If we are to even pretend that we live in a fair society then any voting system should work and work fairly. Be it paper or computer based.
Is America ready for a computer based system? I think a computer based system should be able to replace a paper based system. I think that possiably we should also use paper in addition to the computer system, meaning that they should actually print a reciept of your vote so that in the case of a recount, they have physical proof that you voted for (Gore and not Bush?) the person you said that you did.
I wasn't aware that you could extend Direct3D, I guess that you'd need a seperate API from the card manufacturer or something right?
OpenGL assumes that each card will have it's own extensions and thus provides a common extension mechanism.
The reason that DirectX has been able to grow so quickly is that MS has kept complete control over the platform, only with versions 8 and higher have they actually let the other companies (read nVidia and ATI) have a real say in the api it seems like. I know that they always consulted the card manufacturers, but they still had the final say.
Contrasted to OpenGL where it has taken forever to get version 2.0 out because there are so many people with so many opinions to consider. Open apis are just slower to develop, but in the end, hopefully better since the requirements are more diverse.
I think that you and the parent are missing the real reason that OpenGL is king on the workstations and Direct3D never will be. OpenGL is extendable by everyone and Direct3D is completly controlled by MS.
You can't go and make your own version of Direct3D with a hypercube (or whatever) extension that draws a super widget because Direct3D is closed. OpenGL is completly open and a vendor of 3-D hardware can add whatever they want to the API. The programmers have to then use the extension, but for workstations where the constraints are defined this is usually not a problem.
PC Games on the other hand have to run on the most diverse set of hardware ever. From the lowest 90Mhz Pentium to the newest 2600+ Athlon, with a range of video cards to make a driver writer weak at the knees. (they don't have to run well, just run) D3D solves this by making all functions static and creating software implementations of them. So you program to a 'version' of the API and you know exactly what the machine is capable of and if you can't get an accelerated version of the functionality, then at least you don't have to create your own. Great for game writers, but not as useful for workstation class applications, since they are more concerned with accuracy then games are (in general).
OpenGL is better when the HW that you want to support is smaller, since you can get better performance out of it. Direct3D is better when you just want it to run without having to write a ton of code to emulate the top level functionality.
Solid State Hard Drives exist, with IDE interfaces. They are just expensive for what you get. What are you going to do? You can't use DRAM without a large battery to back it up when the system is off. You can't have a Buffer the size of the drive, that's silly as it's better to just add more system RAM. Basically, it's always better to add more system ram rather then make the HD really fast and solid state. System ram is used to buffer the HD when usage is low and can actually be used by the programs (you know, what you want to use the computer for) when it needs to be.
I've got a gig and a half of ram in my machine, I could buffer an entire CD in ram but I don't have to.
Just another slow news day at Tom's. It's not like you can actually go and buy these things yet. Even if you could, would you unless you were going to buy a whole new computer anyway? Does anyone want to throw out their old drives and replace them all with new ones just because they exist?
It will be nice to get rid of those huge cables though.