I thought I wasn't going to respond to you, but then I noted that you were the forty-second post and couldn't ignore that deep of sign. Anyways, I never meant to say the project shouldn't have been done. As another poster pointed out, having fun and playing with the code is what it's all about for many, if not most of the people who work with Linux's code. I'd also point out that I did say there was some chance of good learning \ technology coming out of the project.
It just hit me how much work goes into projects that don't really accomplish much more than bragging rights, fun and some maybe learning. While there is nothing wrong with this, it doesn't push Linux much closer towards being some awesome OS which is better than the rest. Its why Apple could bring a quality Unix platform to the common user when Linux gui's just aren't quite there for the average Joe.
Setting the record straight:
1) I like linux
2) I have discovered things by playing with code
3) I code for fun (not kernels - so sue me)
4) I don't think everyone should have their noses to the grindstone all the time.
Microkernals are so cute, you just have to love them. Then you realize that sometimes being small means you have to go a long way to talk to those you need. Ok, bad analogy but that's the problem. The Microkernal concept relies on messanging techniques to get done what normal kernals do. You take a decent performance hit there. That's why Apple had to add tons of functionality back into the kernal when it tried to use Mach for OSX.
We know some people lost data and there was a general week long downtime. Sounds like there was some sort of catastrophic hardware failure and the redunancy / backup was... shall we say, less than ideal.
One way or another, this is just another example of why we don't want to trust Microsoft with EVERYONE's personal data. However, it may only be problems like this that save companies from becoming.NET partners which Microsoft will later exploit. Just a thought: Let's say I'm a website which sells sea monkeys, in fact I'm the leading sea monkey seller. I want to make life easy for my customers so I let them use Passport at my site. Now Microsoft has a list of sea monkey buyers, and because they use passport, their email addresses. Microsoft creates seamonkeys.microsoft.com advertises to all my customers and offers prices so low I can't compete. I lose. If Microsoft doesn't screw up like they just did with Messenger, how many web companies could they do this to?
Another decent way to get a fairly cryptic password that is easy to remember is to make some sort of pattern on the keyboard. Make sure you hit a number or two and some punctuation and you've got a nice, cryptic easy to remember password. I wouldn't guard the pentagon with it though.
My personal bet: By the time SSE2 use becomes widespread enough for it to be important, AMD will have it on all their chips. AMD is bright enough to see when instruction sets have to be implemented to stay competitive.
Slashdot Mirror
I thought I wasn't going to respond to you, but then I noted that you were the forty-second post and couldn't ignore that deep of sign. Anyways, I never meant to say the project shouldn't have been done. As another poster pointed out, having fun and playing with the code is what it's all about for many, if not most of the people who work with Linux's code. I'd also point out that I did say there was some chance of good learning \ technology coming out of the project. It just hit me how much work goes into projects that don't really accomplish much more than bragging rights, fun and some maybe learning. While there is nothing wrong with this, it doesn't push Linux much closer towards being some awesome OS which is better than the rest. Its why Apple could bring a quality Unix platform to the common user when Linux gui's just aren't quite there for the average Joe. Setting the record straight: 1) I like linux 2) I have discovered things by playing with code 3) I code for fun (not kernels - so sue me) 4) I don't think everyone should have their noses to the grindstone all the time.
Microkernals are so cute, you just have to love them. Then you realize that sometimes being small means you have to go a long way to talk to those you need. Ok, bad analogy but that's the problem. The Microkernal concept relies on messanging techniques to get done what normal kernals do. You take a decent performance hit there. That's why Apple had to add tons of functionality back into the kernal when it tried to use Mach for OSX.
We know some people lost data and there was a general week long downtime. Sounds like there was some sort of catastrophic hardware failure and the redunancy / backup was... shall we say, less than ideal. One way or another, this is just another example of why we don't want to trust Microsoft with EVERYONE's personal data. However, it may only be problems like this that save companies from becoming .NET partners which Microsoft will later exploit. Just a thought: Let's say I'm a website which sells sea monkeys, in fact I'm the leading sea monkey seller. I want to make life easy for my customers so I let them use Passport at my site. Now Microsoft has a list of sea monkey buyers, and because they use passport, their email addresses. Microsoft creates seamonkeys.microsoft.com advertises to all my customers and offers prices so low I can't compete. I lose. If Microsoft doesn't screw up like they just did with Messenger, how many web companies could they do this to?
Another decent way to get a fairly cryptic password that is easy to remember is to make some sort of pattern on the keyboard. Make sure you hit a number or two and some punctuation and you've got a nice, cryptic easy to remember password. I wouldn't guard the pentagon with it though.
My personal bet: By the time SSE2 use becomes widespread enough for it to be important, AMD will have it on all their chips. AMD is bright enough to see when instruction sets have to be implemented to stay competitive.