A) (Int32)X | ((Int32)Y << 16);
B) (Int32)X + (Int32)Y * (Int32)Int16.MaxValue;
Shouldn't both be equivalent? I know from testing that the first works as expected, but for some reason the second doesn't. Both X and Y are shorts (Int16), and the return type is an integer (Int32).
Shouldn't Y << 16 <=> Y * Int16.MaxValue?
To get the desired behaviour, you need to multiply with 0x10000 (i.e. UInt16.MaxValue+1). Int16.MaxValue is 0x7fff.
5 << 16 327680 5 * 0x10000 327680 Compare to the decimal system: If you want to "shift" the number 5 to 500, you need to multiply with 100, not 99 :-)
There are 2 problems with your second approach:
Right-shift 16 bits = * 2^16 But:
Int16.MaxValue = 2^15-1 I think that you want an unsigned 16-bit max value + 1
Overlooking your MaxValue being one less than a power of two, and since you have a bigger problem to cover first:
The OR and SUM operations are not similar. When you are working with 32-bit integers and 16-bit shifts, there will be carries with your + operation and bit-wise OR'ing with the OR operation.
So, the two ways are quite different.
Then, of course, the MaxValue interpretation makes your two 'shift' attempts different. It should be (x * MaxValue + x) or (x * (MaxValue+1)).
