Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!texsun!texsun.central.sun.com!convex!authorplaceholder From: hutchson@convex.UUCP Newsgroups: comp.lang.c Subject: Re: C Style (Was Re: == vs =) Message-ID: <64400009@convex> Date: 6 May 88 17:31:00 GMT References: <126@atpal.UUCP> Lines: 15 Nf-ID: #R:atpal.UUCP:-12600:convex:64400009:000:755 Nf-From: convex.UUCP!hutchson May 6 12:31:00 1988 If you must take Fortran in a c newsgroup, at least take Fortran, not some brain-damaged non-error-checking dialect. The current ANSI standard(x3.9-1978,"fortran-77") does not allow one to know the representation of the logical values .true. and .false. (except that it fits in the same space in memory as an integer value), so each implementor chooses his own. I have seen at least these: - sign bit==1 is true, sign bit==0 is false. - lsb==1 is true, lsb==0 is false. - any non-zero value is true, zero is false. - any non-zero is considered true, .true. is canonically all one bits - any non-zero is considered true, .true. is canonically==(int) 1 Reversing the meaning and representation of each of these would be equally valid.