John W. Backus, RIP

I have a growing backlog of entries to post, but this gets an immediate bump to the top.

Link to Backus at Wikipedia, where I got most of the details.

Backus led the team that developed FORTRAN. While I may have a scathing, burning hatred for FORTRAN whenever I have to deal with it, when you put it in historical context my opinion immediately inverts. This isn't the place for me to rant about why it's wrong, this is the place for me to rant about why it was right.

It's development was a Good Thing (tm), and a definite step in the right direction. First used in 1957, FORTRAN was the first high-level language that people could take seriously -- it was an optimizing compiler from the start, providing the first ease of programming/efficiency of runtime tradeoff that was actually worth considering.

In 1958 FORTRAN II extended the language to a procedural one, with new keywords like "CALL" and "RETURN", providing a much-superior alternative to the limited "GOTO". That's 10 years before "Go To Considered Harmful", the classic Dijkstra letter.

To this day, I still see people writing code in FORTRAN 77.

His work also made it easier for us to describe, and thus develop computer languages in general. He's the "B" in BNF, the language for describing languages. You'll find BNF in the documentation for any programming language worth its weight in horse piss.

* * *

As a guy who swings a hammer all day, I've got to tip my hat to the one who first tied a rock to the end of a stick.