See the post above yours. As for why the XBox specifically: why not? It's annoying to some of us if there's a platform in use that Linux doesn't support. That alone would be a sufficient reason to me, were I a kernel/hardware hacker.
For another example, my roommate has a bullet proof vest. Another friend makes plate mail armor. The plate costs ~$2500 and takes months to build. A.45 will put a hole in one side, a bigger hole in the soft squishy thing in the middle, and a bigger hole in the back. The bulletproof vest will stop it, and probably only leave you with a big briuse and possibly a cracked rib or 3. The vest cost less than 1/10th of the armor, weighs about 1/100th of the weight, and is more readily available.
Nononononono. Better = costs more. Didn't you know that?
That's bullshit. The two arguments are not equivalent. Obviously, there are a limited number of good seats that can be physically produced. Obviously, there is no limit to the number of Office XP Professional disks that can be pressed.
No, you're wrong. Go read some economics. Capitalist systems are supposed to maximise the number of transactions; the number of transactions increases as the price decreases towards marginal cost. Draw your own conclusions.
As with much of capitalism, paying more gets you more.
This is ok most of the time, because providing more service actually costs more money. That isn't the case here---the code exists and costed the same to develop whether 10 people use it or 100 people.
Actually it's not hard to get printing working if your tech people are thinking about what they're doing. Now, if you're too cheap to hire tech people with an IQ of 100, that's another story...
I see your point, sort of, but why bother running the check at run-time if it's obviously not needed (and you can prove this at compile time)? I just don't see the point.
In any case, you really shouldn't be using arrays unless performance matters anyway, so arrays should receive special performance attention for that reason.
A third arm of the solution would be to generate two entry points for each function with a pre-condition, one which verifies the pre-condition and jumps to the other to do the real work. Then the proof (whether supplied or derived) becomes an optimization (which makes heuristics more acceptable as well---the program won't fail because I used ghc and you used nhc, and nhc can prove the correctness of usages ghc can't).
You don't have to provide formal proofs of all code. I would estimate 90% of all array accesses could be done without any run-time checking if the compiler could supply simple arguments. (I.e., in Haskell you could say: pre-condition for a ! i = i `elem` range (bounds a), then when you access an array with mapM (\i -> ) (range (bounds a)), the compiler knows a ! i within is legal. Simple stuff like that.)
Please explain to me one damn thing that can be imagined more clearly using character-strings-that-aren't-lists than with lists-of-characters. Haskell and ML define String = List of Char, and, as far as I've noticed, nobody's ever complained.
So if a future language has the same kind of GC and exception handling that both Modula-3 and Java has, is it decending from Java, or from Modula-3?
It's descended from Modula-3. Seriously, go read the Java docs some time. The design philosophy behind Java was not to do anything innovative, just stick to really good implimentations of established ideas from other languages. Probably good for a working language, but that also establishes Java pretty firmly as an evolutionary dead-end.
You're wrong, and here's why: there are definite advantages to thought, definite advantages to machine computation, and definite differences between them. So, neither can replace the other (i.e., no AI) and you can't really merge them (at least not without a fundamental re-thinking that's not going to happen because of some technology).
Well, considering the only idea the UN has had for dealing with Iraq in the last 12 years was ``let's chase them around the desert with inspections some more'', I'm not that thrilled about the idea of letting them pick the Iraqi government. The last thing we want is an Iraqi government weak enough that organized crime ends up really running the country.
That has got to be the dumbest statement I've ever heard. (And yes, I know who Einstein was. I don't think every thing he ever thought was intelligent; take his idea of Quantum Physics, for example.)
This is all true, and all in all a wonderful point. But, it doesn't add up to ``user: programmer in training'', which is the argument I was responding to.
I love how evil fixing your mistakes is... Oh, and: we have two parties here. The Democrats had the president (Carter) that put Saddam in power. Funny thing is, they're still the party that wants him there. Or maybe not so funny.
Slashdot Mirror
What does GPL have to do with this? gzip isn't GPL.
Ah... You're my AC troll.
When was the last time I had mod points? About a week ago. HTH. Btw., I think you're stalking me. I'm not sure whether to be flattered.
Why do I get the feeling I'm being stalked?
See the post above yours. As for why the XBox specifically: why not? It's annoying to some of us if there's a platform in use that Linux doesn't support. That alone would be a sufficient reason to me, were I a kernel/hardware hacker.
Nononononono. Better = costs more. Didn't you know that?
That's bullshit. The two arguments are not equivalent. Obviously, there are a limited number of good seats that can be physically produced. Obviously, there is no limit to the number of Office XP Professional disks that can be pressed.
Explain to me how it's ``freeloading'' to refuse to get a free account? I thought freeloading was getting something for free you should have paid for.
In any case: the guy you're replying to had real accounts and played by the rules all the way. The accounts were deleted. Why?
Furthermore, you're advocating giving (essentially) fake information as an alternative to ``freeloading''? What?!?
No, you're wrong. Go read some economics. Capitalist systems are supposed to maximise the number of transactions; the number of transactions increases as the price decreases towards marginal cost. Draw your own conclusions.
This is ok most of the time, because providing more service actually costs more money. That isn't the case here---the code exists and costed the same to develop whether 10 people use it or 100 people.
Actually it's not hard to get printing working if your tech people are thinking about what they're doing. Now, if you're too cheap to hire tech people with an IQ of 100, that's another story...
It's also a slashdotting at its best!
I see your point, sort of, but why bother running the check at run-time if it's obviously not needed (and you can prove this at compile time)? I just don't see the point.
In any case, you really shouldn't be using arrays unless performance matters anyway, so arrays should receive special performance attention for that reason.
Python doesn't belong on this list, due to the presence of lambda.
A third arm of the solution would be to generate two entry points for each function with a pre-condition, one which verifies the pre-condition and jumps to the other to do the real work. Then the proof (whether supplied or derived) becomes an optimization (which makes heuristics more acceptable as well---the program won't fail because I used ghc and you used nhc, and nhc can prove the correctness of usages ghc can't).
You don't have to provide formal proofs of all code. I would estimate 90% of all array accesses could be done without any run-time checking if the compiler could supply simple arguments. (I.e., in Haskell you could say: pre-condition for a ! i = i `elem` range (bounds a), then when you access an array with mapM (\i -> ) (range (bounds a)), the compiler knows a ! i within is legal. Simple stuff like that.)
Please explain to me one damn thing that can be imagined more clearly using character-strings-that-aren't-lists than with lists-of-characters. Haskell and ML define String = List of Char, and, as far as I've noticed, nobody's ever complained.
It's descended from Modula-3. Seriously, go read the Java docs some time. The design philosophy behind Java was not to do anything innovative, just stick to really good implimentations of established ideas from other languages. Probably good for a working language, but that also establishes Java pretty firmly as an evolutionary dead-end.
You're wrong, and here's why: there are definite advantages to thought, definite advantages to machine computation, and definite differences between them. So, neither can replace the other (i.e., no AI) and you can't really merge them (at least not without a fundamental re-thinking that's not going to happen because of some technology).
Well, considering the only idea the UN has had for dealing with Iraq in the last 12 years was ``let's chase them around the desert with inspections some more'', I'm not that thrilled about the idea of letting them pick the Iraqi government. The last thing we want is an Iraqi government weak enough that organized crime ends up really running the country.
That has got to be the dumbest statement I've ever heard. (And yes, I know who Einstein was. I don't think every thing he ever thought was intelligent; take his idea of Quantum Physics, for example.)
This is all true, and all in all a wonderful point. But, it doesn't add up to ``user: programmer in training'', which is the argument I was responding to.
I love how evil fixing your mistakes is... Oh, and: we have two parties here. The Democrats had the president (Carter) that put Saddam in power. Funny thing is, they're still the party that wants him there. Or maybe not so funny.
Hey! Be careful how you speak of my birthday!
And this is different from other established news sources how?