🚀 go-pugleaf

jkut@diku.dk (Jonatan Kutchinsky) writes:
> We are writing a program in C as a computer science project, where we
> simulate a Motorola 68000 processor. We decided decided to typedef the
> built in types char, short, and long to new names, BYTE, WORD, and LWORD.
> 
> The reason why we want to do this ...
[reasons omitted: that there *are* reasons is enough.]
> 
> The problem is that our teacher has the opinion that this use of typedef is
> a bad thing, and that it makes the program less readable when you do not
> know that LWORD really stands for long, without looking in the header file.
> ...

IMO your teacher is too much wedded to knowing the internal
representation of everything.  After many years of assembly language
programming one can get into that state, and one can see C as assembly
language written horizontally rather than vertically; but there are
other ways of using C.  C can actually be useful in hiding
implementation details where they're not needed.

I'm not sure I'd do it the way you have, but ten years of using the
language have moved me toward the approach you're using and away from
flaunting implementation details wherever possible.          

Besides, it's possible that your teacher's advice is actually Bad
Practice.  Who says the LWORD ``really stands for'' long, or BYTE for
char?  If you ported your simulator to another machine, you might have
to define BYTE as signed char, or something, since the characteristics
of the 68000 are constant, while those of the simulating machine are
not.  If this kind of thing is a practical possibility, then your
practice is Right, and the teacher's is Wrong.  I.e., your philosophical
reasons for the declarations could turn out to be entirely practical; no
one, least of all a professor, should be surprised when that happens.

--
Dan Drake                 In software, the *complete* design *is* the software!
drake@Autodesk.com        --Hugh Lamaster