Don't believe me. Do your own measurements. Crack open some Java code and count the fields.
I'm talking about the SPECjvm98 benchmark, which I instrumented and measured myself, as well as word-of-mouth about other Java code.
The only way that half of memory would be pointers is if your entire computer's memory had a tree or list of integers, and what use is that?
What are you talking about? What about a bunch of objects with 4 pointers and 4 ints? In fact, lots of classes have no actual fields at all; just a pointer to the class, and a pointer to the outer object.
Every pointer in every data structure now requires twice as much memory. I can't say for all programs, but in the Java world (where I work), about half of memory typically contains pointers. Therefore you expect to see a 50% increase in memory consumption.
CPU stacks now have 8-byte entries, so they are pretty much always twice as big.
AMD64 code is quite a bit bigger than IA32 code. Most estimates say 15%.
None of these double your memory requirements, but it's probably easier for them to prereq 2GB of ram than 1.4GB.
I'm going to go out on a limb and predict that we'll never, ever pull a comet into Earth orbit for the purpose of mining it. There's just no conceivable reason to do so: If you're going to take mining equipment into orbit, why not just take it to the comet?
Are you talking about our moon? There's no way it's easier. It takes 6 months by Hohman transfer orbit. If you're willing to spend that much time getting there, you could just go into a low orbit around Earth and use a really efficient booster to go in ever-wider circles until you get caught in the moon's gravity.
Ah, "just go into a low orbit around Earth"? I'm afraid that's the hard part. There is currently no "really efficient booster" with enough thrust to get a man from Earth's surface to LEO. It's much much cheaper to get from Mars surface to LMO.
Besides, who cares about going to the moon? This is about exploiting resources, which mostly means bringing stuff here.
I was responding to the claim that Mars resources are only useful on Mars or in Mars orbit. This is not true, because it's actually easier to get to almost anywhere in the solar system from Mars' surface than from Earth's.
You are asserting things about which you have no clue. For example, it's much easier to get to the moon or the asteroids from Mars than from Earth. And I'm not sure why you think Mars' minerals are any more "boring" than Earth's.
Take a look at Zubrin's "The Case for Mars" to get a clue.
I guess you're joking, but I don't really get it. A nonstop round-the world flight is not an orbit, any more than me sitting here stationary and going around the world in 24 hours is an orbit.
Well, true, though that depends on how much atmosphere. But yes, we agree that the atmosphere makes an orbit much, much harder -- not easier as the gp seems to claim.
That is false. To get around the globe just once requires practically the same velocity as a stable orbit. Doing it while you're in the atmosphere would require even more speed, because you need to have extra speed to make up for atmospheric grag.
Slashdot Mirror
I'm talking about the SPECjvm98 benchmark, which I instrumented and measured myself, as well as word-of-mouth about other Java code.
What are you talking about? What about a bunch of objects with 4 pointers and 4 ints? In fact, lots of classes have no actual fields at all; just a pointer to the class, and a pointer to the outer object.CPU stacks now have 8-byte entries, so they are pretty much always twice as big.
AMD64 code is quite a bit bigger than IA32 code. Most estimates say 15%.
None of these double your memory requirements, but it's probably easier for them to prereq 2GB of ram than 1.4GB.
GC also protects against dangling references, which is a more serious problem than a pure memory leak.
The sad part is that I knew all these things already. I guess I just couldn't pass up the opportunity to be a smartass.
I think that's a transit of Mercury.
I'm going to go out on a limb and predict that we'll never, ever pull a comet into Earth orbit for the purpose of mining it. There's just no conceivable reason to do so: If you're going to take mining equipment into orbit, why not just take it to the comet?
You just keep asserting things. If you have some kind of reasoning, evidence, or citations to back them up, I'd love to learn about it.
Wait, you were joking, right?
Take a look at Zubrin's "The Case for Mars" to get a clue.
What is 8*50*20*12? Why not just multiply by 2000 hrs/year? (You don't work 20 days every month, do you?)
I don't get it. How did you get from this book review to Tolkein?
It said "supernova event". Maybe that's vague enough to be considered synonymous with "supernova observation".
Yes. And maybe nobody wants an unpredictable cost structure. Rather than $0.01 per click, maybe users want to pay $2 per month and not worry about it.
That's four words, Mr. Quayle.
I had to look this up... General Products.
I guess you're joking, but I don't really get it. A nonstop round-the world flight is not an orbit, any more than me sitting here stationary and going around the world in 24 hours is an orbit.
I'm honoured to be the first to add you to my foes list.
You're talking with the benefit of hindsight. Where does the "if you've made $3 million, walk away" argument end? $1 million? $10,000?
Well, true, though that depends on how much atmosphere. But yes, we agree that the atmosphere makes an orbit much, much harder -- not easier as the gp seems to claim.
That is false. To get around the globe just once requires practically the same velocity as a stable orbit. Doing it while you're in the atmosphere would require even more speed, because you need to have extra speed to make up for atmospheric grag.