Short and sweet

// Any number larger than a small number will do
final double A_LARGE_NUMBER = 600000.0;

I dunno. Any finite number is pretty small, if you ask me. I would have gone with Double.POSITIVE_INFINITY.

THAT IS NOT A SMALL NUMBER! THAT IS A BIG NUMBER!

It's finite, though. By definition it's smaller than most natural numbers. Almost all natural numbers, in fact. How can something so small be considered large?

That was not directed at you, sorry if it came out that way. It was supposed to be a rather lame inside joke. You're obviously right.

That's the problem with inside jokes, I guess. They're risky business.
Anyway, I kinda want to know what a "number larger than a small number" is. Obviously 600 000 is such a number, and by extension all numbers larger than 600 000 should be in the same class. However, what is the smallest number larger than a small number?

I don't think you can answer that question. The condition is just messed up. Because the only thing we know about a "small number" is that it's not a "large number", it's a circle definition. You could just as well write "This function works well for all inputs except for those where it will break". Which makes the thing so hilarious.

I don't think it's as circular as you think it is. Since there must be logic that uses this socalled "large number", then the smallest number for which the logic still works as desired is, in fact, the smallest large number, and the largest number for which the logic does not work is, in fact, the largest small number. The question, though, is whether the smallest larger number is larger than the largest small number.

Not necessarily. The number could be used in an algorithm where it influences the runtime. Then the algorithm would work with small numbers as well, it would just need ages to finish. Or, what I think is most likely, the number is used to initialize some buffer that the programmer was to lazy/incompetent to calculate the real size of. In that case the number would influence the time the program runs until it crashes spectacularly.

That raises another question, just WTF is A_LARGE_NUMBER used for in this program? I would think that it's a buffer or some kind of maxloop iteration test to make sure that a potentiallyinfinite loop doesn't go too far, but then why is it a double? Nearly all cases I can imagine needing A_LARGE_NUMBER would require ints instead of doubles. Off the top of my head I can't think of any uses for large doubles. We should ask the submitter.

I guess it's the boundery of space.

For all you guys know the previous line is:
final double A_SMALL_NUMBER = 599 999.9;
Then the whole thing makes perfect sense. Sort of.

Here's the thing, though: 600000 isn't even that large for a double. Hell, it's not large for a float, or even an int.

@Welbog said:
Anyway, I kinda want to know what a "number larger than a small number" is. Obviously 600 000 is such a number, and by extension all numbers larger than 600 000 should be in the same class. However, what is the smallest number larger than a small number?
2.

@Welbog said:
I don't think it's as circular as you think it is. Since there must be logic that uses this socalled "large number", then the smallest number for which the logic still works as desired is, in fact, the smallest large number, and the largest number for which the logic does not work is, in fact, the largest small number. The question, though, is whether the smallest larger number is larger than the largest small number.
Say that out loud. It kind of made my brain hang up for a minute. Almost needed a reboot.

Almost all natural numbers are very, very, very large.

@Welbog said:
Almost all natural numbers are very, very, very large.
Exept for the ones that are < A_LARGE_NUMBER.
final double A_VERY_VERY_VERY_LARGE_NUMBER = A_LARGE_NUMBER + A_VERY_LARGE_NUMBER + A_VERY_VERY_LARGE_NUMBER;




so what about final double A_LARGE_NUMBER = 800813.5 ?

@Welbog said:
That raises another question, just WTF is A_LARGE_NUMBER used for in this program? I would think that it's a buffer or some kind of maxloop iteration test to make sure that a potentiallyinfinite loop doesn't go too far, but then why is it a double? Nearly all cases I can imagine needing A_LARGE_NUMBER would require ints instead of doubles. Off the top of my head I can't think of any uses for large doubles. We should ask the submitter.
In raytracing, it's quite useful to define the "largest reasonable number" and the "smallest reasonable number". Since the vast majority of raytracers use floating point, these two numbers are often doubles.

I found another large number:
Scroll a bit down for:
math.huge
The value
HUGE_VAL
, a value larger than or equal to any other numerical value.