Ross, I tend to use Integer for everything and Int64 only when I know Integer does not have enough bits. I almost never bother with the lower precision integers. There is hardly any speed difference. Fetching an aligned 16-bit integer from memory is just as fast as fetching an aligned 32-bit integer. Furthermore, you are not going to save any memory using 16-bit integers, as Delphi automatically generates code that aligns variables to 32-bit boundary. The only case when you would save memory is if you have a packed array.
If you need to explicitly limit the value, it is better to use a sub-range type and turn range checking on. Dennis. ----- Original Message ----- From: "Ross Levis" <[EMAIL PROTECTED]> I've tended to use ShortInt and SmallInt rather than Integer to not waste memory when values are never going to exceed their maximum, but I read somewhere that Integer should be used when possible for speed purposes. I've found it slightly useful when reading the source to know that a particular variable doesn't ever exceed 127 or 32767. What sort of speed differences are we talking about? Is it worth doing a global find/replace? Cheers, Ross Levis. _______________________________________________ Delphi mailing list [EMAIL PROTECTED] http://ns3.123.co.nz/mailman/listinfo/delphi
