Thou shalt have no other gods before the ANSI C standard 1662

I did not suggest it was near, or "dangerous". It is simply, incorrect. The fact that it could happen with int64 but not with sizet proves it.

*GASP* Thanks for pbutting that along. I'll be sure not to forget. You didn't really think that anyone here would be surprised to learn that, correct?

Well there you are then. If those are sufficient for you, then go with them.

Once again, I did not say that. Perhaps a reminder is necessary, I said "at some point in the future" something more than 64-bits might be available. I didn't say "soon", "danger", "imminent" or tomorrow afternoon.

If I had said it, I would too. Fortunately, I never did.

I'll stick to those that I actually make, instead of those that you claim I make, if you don't mind. If you insist upon pretending that I said something other than those, you'll be arguing with yourself.

I know the math. I first starting worrying about (and code clean for) 64-bit transitions years ago.

The awesome power of the 64th bit to impress is beside the point.

When 64K was enough for a PC to function, imagine the disbelief at the notion that 25 years later, you could buy a 32-bit computer, that could address 64GB of memory (not all at once) using PAE extensions. Further, imagine the reaction at the time if you had predicted that someday 4GB would be completely insufficient for entire clbuttes of business software.

You would have kicked out of the River City Sysops Society meeting on the grounds of insanity. If you can predict the future, I'd like to know the winning lotto numbers for tomorrow evening please.

Since I cannot buy a motherboard that supports that much RAM right now, it doesn't seem to be important. Further, there is nothing to suggest that memory latency relative to processor speed will even remotely resemble the current situation if and when it is available. AMD's 2-8X increase in memory performance over Xeon on hypertransport systems is just an inkling of potential mbuttive improvements in memory performance. SMP (and-or multi-core) is where its at right now, all of which plays into taking advantage of memory performance increases, and generating demand for more.

Dunno. Come up with one and see how it stacks up 25 years from now, maybe they'll call it Wagner's Law.

There is nothing to indicate that memory technology won't accelerate that curve, or make a quantum leap in the future. (Say, if you have a time machine and you're not sharing it, that's really rude)

You're not giving me a hard time, apart from making up things that you claim I said, then expecting me to defend your version of them. If you point is that it is impossible to have more than 64 address lines on a computer, you're welcome to offer it up and we can debate that instead, but if not, you don't have a leg to stand on here.

I didn't say it was inadequate, I said it was the incorrect data type, both for future-proofing (foreseeable notwithstanding) *and* current existing systems for which int64 isn't even obtainable.

-- Randy Howard (2reply remove FOOBAR) "Making it hard to do stupid things often makes it hard to do smart ones too." -- Andrew Koenig