These things are sold at a loss and all profit comes from the games. If people buy the PS2 as a DVD player, Sony looses money. That's one of the reasons Microsoft is resisting adding DVD playback to the XBox.
I've been a TF/TA for several computer science classes and we always required student submissions to compile and run on our primary environment. In our case that was Digital Unix using gcc or the JDK. When you are grading the assignments you have to be able to test them to give a fair assessment. I use Linux on my personal machine so accepting assignments that will only compile under Visual Studio on Win32 isn't a realistic situation. I'm sure/. readers wouldn't have as much problem with requiring Unix, but it's the same principle.
That said, many of my students did they work on Windows using several of the popular IDEs for C, C++, and Java. For intro classes, the focus is generally on the standard language sans platform-specific extensions. Once you've written and tested your code, it's easy to copy it onto the target machine and recompile to make sure it still works. The only problems I encountered were with students who didn't know how to get their source files out of the projects that most IDEs create.
If everyone handed in code that worked perfectly there would be no problem, but unfortunately, grading often means a certain amount of debugging to see how serious the problem really was, and that means the code needs to be in a usable format.
don carnage said:
How many of us were walking around saying, "You know, I wish I had a game where I could run around and shoot Nazi's..."?
I see your point, but I thought it interesting that Wolfenstein 3D was the game that really spawned this genre, and it was based on an old Apple II game called Wolfenstein. Carmack was an Apple II game programmer before he moved to PCs, so I assume that's where he came across Wolfenstein. The basic idea (in crude 2D graphics) was that you were a prisoner breaking out of a Nazi stronghold. Wolfenstein 3D (and its successor, Doom) certainly deserve credit for creating the first person shooter genre, but even they were based on old material.
One source of potential problems is that people download binaries and/or source and install them without checking them out too closely. Sure, you feel more secure after downloading 200k of source code and compiling it yourself, but how often do you actually read through the code before using it? RPMs are often available (or.tar.gz files with install scripts) and we happily run all associated scripts as root, as well as installing any arbitrary code that happens to be in a package. So far, I think our main assurance has been that nothing too bad has happened (at least nothing malicious) in the past, and we count on communities like/. to warn us if there is anything we should watch out for.
As projects get bigger and the user community gets bigger, it will become easier and easier to download a trusted package, patch a nice trojan into the source code in some obscure place, package a few RPMs, and stick it in a public place. Even most people who download source packages don't examine the source too closely (until it crashes). Checksums? Sure, I can make a checksum if that will make you feel better. The point is, we don't always download from trusted sites, and even if we do it's doubtful that the packages we download have been scrutinized. Someone could slip a trojan into emacs and upload it to a public download site and thousands of people would probably download & use it without ever knowing until it's too late. If you were subtle, you might even get one into a widely used distribution. You think the RedHat folks have read through all the hundreds of megabytes of code on a source RPM CD? It seems unlikely to me.
So what's the point? Some kind of trusted authority system might be helpful. Verisign, etc., have their problems, but at least you have some idea of who you are trusting when you use a verification system. If there's a system already in place, I don't use it, and I suspect most others don't use it either. I usually ignore checksums and signatures that some sites provide with packages. We need a system that is more integrated. Every time I do rpm -ivh.rpm I'd love to see a message that says "package verified by Linuxone.com." Okay, maybe not that exact message, but something like that where the verification was a process I could trust. We run too many install scripts as root, so the Linux security model doesn't really offer as much protection as we'd like.
This industry has a short memory. The Pentium was called the P5 for a long time, and the Pentium Pro core was generally called the P6. P4 seems like about the worst choice of things to call the Pentium IV.
A study of AltaVista's logs for August 1998 showed that "yahoo" was the 7th most popular query over a period of a few weeks. It seems kind of dumb for Lycos to do that, but on the other hand they are probably catching a lot of hits that way. There are quite a few people new to the web who don't know any better.
Slashdot Mirror
These things are sold at a loss and all profit comes from the games. If people buy the PS2 as a DVD player, Sony looses money. That's one of the reasons Microsoft is resisting adding DVD playback to the XBox.
- Russ
I've been a TF/TA for several computer science classes and we always required student submissions to compile and run on our primary environment. In our case that was Digital Unix using gcc or the JDK. When you are grading the assignments you have to be able to test them to give a fair assessment. I use Linux on my personal machine so accepting assignments that will only compile under Visual Studio on Win32 isn't a realistic situation. I'm sure /. readers wouldn't have as much problem with requiring Unix, but it's the same principle.
That said, many of my students did they work on Windows using several of the popular IDEs for C, C++, and Java. For intro classes, the focus is generally on the standard language sans platform-specific extensions. Once you've written and tested your code, it's easy to copy it onto the target machine and recompile to make sure it still works. The only problems I encountered were with students who didn't know how to get their source files out of the projects that most IDEs create.
If everyone handed in code that worked perfectly there would be no problem, but unfortunately, grading often means a certain amount of debugging to see how serious the problem really was, and that means the code needs to be in a usable format.
- Russ
I see your point, but I thought it interesting that Wolfenstein 3D was the game that really spawned this genre, and it was based on an old Apple II game called Wolfenstein. Carmack was an Apple II game programmer before he moved to PCs, so I assume that's where he came across Wolfenstein. The basic idea (in crude 2D graphics) was that you were a prisoner breaking out of a Nazi stronghold. Wolfenstein 3D (and its successor, Doom) certainly deserve credit for creating the first person shooter genre, but even they were based on old material.
As projects get bigger and the user community gets bigger, it will become easier and easier to download a trusted package, patch a nice trojan into the source code in some obscure place, package a few RPMs, and stick it in a public place. Even most people who download source packages don't examine the source too closely (until it crashes). Checksums? Sure, I can make a checksum if that will make you feel better. The point is, we don't always download from trusted sites, and even if we do it's doubtful that the packages we download have been scrutinized. Someone could slip a trojan into emacs and upload it to a public download site and thousands of people would probably download & use it without ever knowing until it's too late. If you were subtle, you might even get one into a widely used distribution. You think the RedHat folks have read through all the hundreds of megabytes of code on a source RPM CD? It seems unlikely to me.
So what's the point? Some kind of trusted authority system might be helpful. Verisign, etc., have their problems, but at least you have some idea of who you are trusting when you use a verification system. If there's a system already in place, I don't use it, and I suspect most others don't use it either. I usually ignore checksums and signatures that some sites provide with packages. We need a system that is more integrated. Every time I do rpm -ivh .rpm I'd love to see a message that says "package verified by Linuxone.com." Okay, maybe not that exact message, but something like that where the verification was a process I could trust. We run too many install scripts as root, so the Linux security model doesn't really offer as much protection as we'd like.
- Russ
This industry has a short memory. The Pentium was called the P5 for a long time, and the Pentium Pro core was generally called the P6. P4 seems like about the worst choice of things to call the Pentium IV.
A study of AltaVista's logs for August 1998 showed that "yahoo" was the 7th most popular query over a period of a few weeks. It seems kind of dumb for Lycos to do that, but on the other hand they are probably catching a lot of hits that way. There are quite a few people new to the web who don't know any better.
- Russ