Thread: Should I be smart about the types I use?

I know a value for a certain variable is going to be under or == 255...yet I always use ints... so it's:

Code:

typedef (unsigned char) byte;
byte hi[] = {1, 2};

VS

Code:

int hi[] = {1, 2};

I mean it's basically saying to myself if I want to use 1 byte, or 4 bytes. Which 3 bytes would be wasted, because all of the values would be between 1 and 255. In your opinion, would you go with the byte route, or using plain ol' ints?