Re: Languages: The Bigger the Uglier (was: Re: Aliasing in ISO C)

The moderator writes:


      [The usual justification for allowing different integer sizes and other
      such variations from one platform to another is that if you tie down too
      many details, you make it impractical for the language to run reasonable on
      machines with different word sizes. Now that everyone seems to have
      forgotten that there were ever machines with word sizes not powers of two,
      perhaps this argument is worth revisiting. -John]


I liked the version occam uses: If you don't care (e.g., for loop
counters and array indices), you specify INT which uses the native
integers, i.e., 16 bits on the 16 bit transputers and 32 bits on the
32 bit transputers. If you want a certain precision regardless of the
hardware you're running on, you specify INT16 or INT32 or..., as the
case may be. For sizes larger than the native size, the compiler will
automatically use the with-carry operations appropriately.


Jan
--

Return to the comp.compilers page.
Search the comp.compilers archives again.