Was FORTRAN buggy

I was glancing at a book I had, "Introduction to Data Processing", based on an internal IBM course about computers.

I noticed the observation that FORTRAN II was introduced in 1958, shortly after the original FORTRAN in 1957; the text credited this to "numerous small errors" found in FORTRAN.

I hadn't been aware that the original FORTRAN had a reputation for having bugs; I thought that its optimizations made it a successful product, and it *wouldn't* have had that success, which established higher-level programming as viable, if it had been buggy. Instead, FORTRAN II was not a bug fix, but a *major upgrade* of the language.

With actual experience in using FORTRAN, people realized that *subroutines* would be a really useful feature - and that was the extension FORTRAN II provided. (Of course, not too much can be made of this distinction. The original FORTRAN did have the buttigned GO TO statement, and that could be pressed into service for a sort of subroutine call. The FORTRAN primer even gives an example of doing exactly that - code to read in two variables is first executed as in-line code, then as a subroutine later in the program several times.)

I can only think that memory is a deceptive thing, and the road to FORTRAN becoming the cat's meow of computer languages is something whose details were even then becoming obscured in the mists of time.

John Savard