Re: Understanding Booleans

[dewar@gnat.com (Robert Dewar)]

rod@praxis-cs.co.uk (Rod Chapman) wrote in message news:<ba18d5cb.0111220751.3ba7797a@posting.google.com>...
> >    A := (b = c) or (e = f);
> > 
> > some people prefer to write
> > 
> >    if (b = c) or (e = f) then
> >       A := True;
> >    else
> >       A := False;
> >    end if;
> 
> The latter form suffers slightly from introducing two 
> paths,
> whereas the former only has one (ignoring the possibility
> for the moment of the compiler transforming the "or" into
> an "or else".) This could have a negative
> effect on coverage analysis and the volume (but
> not the difficulty) of code proof.
>  - Rod

If this is harder to prove, then it is just an indication
of weaknesses in the techniques used, since the two forms
are obviously semantically identical. Yes, just as we sometimes
distort code to deal with incomplete optimization, we may have to
distort code to deal with
incomplete proof tools :-)