The early 1970's were a very interesting time in computer sciences. By mid decade, the argument over structured programming had given way to acceptance. But those of us who programmed in FORTRAN were left without structure naturally defined in the programming language. We were forced into using the dreaded GOTO.

Two colleagues, Tim Johnson and Ben Consilvio, and I set out to correct this failing. We defined a set of macros that executed within the University of Maryland Text Editor (med). The set of macros was named B4 (so called because it needed preprocessing before being submitted to the FORTRAN compiler). B4 and Rational FORTRAN had a lot in common. Code that was written in either dialect required preprocessing. The resulting transformed code was not related to what the programmer originally wrote (making debugging very difficult). But the advantages were significant.

Tim has corrected my recollection:

B4 did generate a rather odd looking FORTRAN source, but the two were certainly related, and the correspondence was not too hard to follow (assuming you wanted to). I wouldn't think the executable code would be any more unrelated to B4 than it was to a typical FORTRAN program. As I recall, debugging the stuff was usually done with WRITE statements anyway.

In 1977, all three of us were delighted when the ANSI FORTRAN Technical Committee announced the forthcoming American National Standard (ANS) FORTRAN, nicknamed "FORTRAN 77". The technical committee scheduled two public hearings, one on the West coast, the other on the East coast (at the National Institutes of Health campus, outside Washington DC). All three of us obtained a copy of the draft and set out reading about our new language. Almost immediately, it became apparent that FORTRAN 77 was nothing more than the old FORTRAN. There were no new structured constructs. GOTOs were inevitable. I can't describe the let down we all felt.

Because the East coast meeting was only an hour from our work, we decided to attend the public hearing. The morning was spent going over the draft ANS. Having read the draft, there were no surprises except to note that the West coast meeting had not addressed the lack of structured constructs in the draft. I was somewhat surprised that no one had raised the issue.

Come the afternoon and the question and answer period. Some attendees asked some technical questions. I waited at a balcony microphone. When the Chair recognized me, I identified myself and asked "Were the Technical Committee members unaware of or did they just ignore the current literature regarding structured programming?" The Chair was obviously not pleased and immediately recognized a gentleman at another balcony microphone. I was delighted when that questioner said "I want to hear your answer to his question." What followed was a rehash of the previous seven years of structured vs. non-structured programming. Finally, the Chair indicated that, because there were prior commitments for the auditorium, we would adjourn to the cafeteria. The discussion continued until early evening. But finally, the Chair indicated that a new draft would be distributed and that our comments were solicited.

The new draft arrived. It included the Block IF statement. It was acceptable although we may have wanted a slightly different set of keywords and a few more structures. So, that's why FORTRAN 77 wasn't official until late in 1978.

The B4 macros were dependent upon the University of Maryland Text Editor. In turn, the text editor was dependent upon the Univac computer (the editor was wholly written in Univac assembler). One of the B4 team, Ben Consilvio, decided to implement a FORTRAN preprocessor in FORTRAN. He told the story that ForMel, thought to mean "FORtran Made ELegant" really meant "For Mel". Mel was his manager who believed that such a beast could not be built. That ForMel was successful just demonstrates that renaissance men, like Ben, succeed at most of what they do.