Friday, April 21, 2006, 4:02:03 AM, Johnny Andersson wrote:
JA> Actually I really don't understand the point in having a limit at all. Well, there sort of isn't one, so you're about right. The problem is that folks in this business are both optimistic (Sure, I can have that done by, oh, six pm) and pessimistic (no one will ever need more than 64K of memory, and they couldn't afford it if they did.) If you fiddle with the numbers for a while, you'll discover that 256 X 64K is a reasonable limit because those are the maximum unsigned values in the number of bits allocated for the indices to the table of cells: 256 is the maximum number of values one can fit in two hexadecimal digits (FF=255, and there's zero), and 65535 is the corresponding max for four bytes, 0xFFFF. Adding the 65537th row requires that every instance of the row index in the code be redefined to be twice as big (FFFFFFFF) and perhaps the same for the column limit (FFFF). 4,294,967,296 (FFFFFFFF+1) would be the new limit for rows. And that's only slightly bigger than the number of code changes required to make it happen ;-) Hope that helps. Chuck --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
