Was FORTRAN buggy 4356

As with the question of the storage medium for the compiler (elsewhere in this thread) bugginess is usually more related to an implementation than a language, unless the language specifications result in a functional ambiguity...a problem that I've not seen any documentation for in the original FORTRAN specs. (Caveat: I've not looked for them either...)

And sometimes a "bug" is in the eye of the beholder. Some small-system compilers, for example, took shortcuts that allowed technically invalid source to compile; one of the better-known examples was the trick that accepted any 9-character string at the beginning of a line which began with "D" and ended with "N" as "DIMENSION". (A common use of this in student decks was "DAMNATION".) As systems began to acquire more memory to allow more exacting syntax checks, at the same time new revisions to the FORTRAN standard were being developed, which may have resulted in the erroneous buttociation of better compilers with the newer standards for the language.

Of course, there were always the poorly-considered deliberate design decisions for some compilers and libraries. One that I heard of but don't recall the system involved resulted in the ability of control to fall through an arithmetic IF to the next sequential instruction if an arithmetic overflow occurred while evaluating the expression.

Slightly off-topic question: has anyone here ever used the FREQUENCY statement in a FORTRAN program, or know anyone who did? Or, for that matter, know any FORTRAN compiler that actually implemented it? (The FORTRAN II compiler under FMS accepted it with a syntax check, but otherwise ignored it.)

Joe Morris