garry@batcomputer.tn.cornell.edu (Garry Wiegand) (07/08/87)
In a recent article firth@bd.sei.cmu.edu.UUCP (PUT YOUR NAME HERE) wrote: >In article <371@nikhefh.UUCP> i91@nikhefh.UUCP (Fons Rademakers) writes: >>Does somebody know what the ANSI FORTRAN 77 standard says about >>list directed I/O from internal files? >... >The Apollo compiler is right. The others are non-compliant, >but you might check for some invocation switch or option that >would request compliance. I've seen this usage before, and it strikes me as very odd. In common English, "compliant" means something like "permissive, cooperative, forgiving". "Non-compliant" means that someone is not complying with something. In the case at hand, the compiler is going out its way to allow the programmer to do something she wants to do. It is the *Standard* that the compiler is not complying with, in the section where the standard says "this activity is not to be permitted, cooperated with, or forgiven". Thus - by not cooperating with an instruction not to cooperate - the forgiving compiler gets labelled as "ick, nasty, non-compliant!" It's a bit of double-speak worthy of the Government. How about we banish it before Justice Rehnquist discovers it? garry wiegand (garry@oak.cadif.cornell.edu - ARPA) (garry@crnlthry - BITNET)
ken@rochester.arpa (Ken Yap) (07/08/87)
Do the standards documents really use the word "compliant"? I thought the usual term was "conforming". Ken
drw@cullvax.UUCP (07/09/87)
garry@batcomputer.tn.cornell.edu (Garry Wiegand) writes: > I've seen this usage before, and it strikes me as very odd. In common > English, "compliant" means something like "permissive, cooperative, > forgiving". > > "Non-compliant" means that someone is not complying with something. In > the case at hand, the compiler is going out its way to allow the > programmer to do something she wants to do. It is the *Standard* > that the compiler is not complying with, in the section where the > standard says "this activity is not to be permitted, cooperated with, > or forgiven". > > Thus - by not cooperating with an instruction not to cooperate - the > forgiving compiler gets labelled as "ick, nasty, non-compliant!" The problem here is that you want to equate "compliant" with "being nice", and think that the more the compiler lets the programmer get away with, the "nicer" it is being. The trouble is that the truth is almost opposite -- the more the compiler forces the code writer to stick to an exact, formally defined language, the nicer it is being *for the programming project as a whole*. Even ANSI C is remarkably lax in this regard. We've written a few zillion lines of code here that will break when we try to port this stuff to a machine with a different length int. Dale -- Dale Worley Cullinet Software ARPA: cullvax!drw@eddie.mit.edu UUCP: ...!seismo!harvard!mit-eddie!cullvax!drw Back spewing nonsense again!