Back in 1972 when I was still in junior high school the local IBM plant would open its office to a bunch of young geeks and let them play on the computers on Tuesday nights under the (very loose) supervision of a few engineers. We'd spread out and each take a seat in somebody's office cubicle and type on the 2741 Selectric terminals learning APL. I took to it with enthusiasm and mastered the language. It was my first experience using and programming a computer. I wrote a few games, and a graphics program for the Tektronix 4015 APL terminal. I used the STARMAP program for a school project in Earth Sciences to explain retrograde planetary motion. It was all great fun.
A few years later another IBMer heard about our group and took a couple of us under his wing. He was not part of the APL department; he used Fortran and offered to teach it to us. I'd heard of Fortran and thought it was something awesome that would run rings around APL. Wow, what a big disappointment. What you have to throw away a punch card every time you made a typing mistake? And you can't even add 2 arrays without writing a loop and doing them element at a time? Yet he, who knew nothing of APL, thought it the greatest ever. I learned then how you can get stuck in a rut doing things the hard way with a difficult tool if you don't look at others.
Later in college at MIT I was exposed to LISP, and ALGOL and some very different ways of thinking. But I was still surprised to see the lack of array processing in these languages. The early versions of LISP we used didn't have arrays as a data type at all.
I somehow learned about a fellow named Trenchard Moore working on improvements to APL and that he worked at an IBM office just a few blocks from campus. I went over to his office and introduced myself. He loaded me up with a bunch of papers he'd written and I devoured them. I also made friends with David Moon and Guy Steele who were grad students working on the LISP machine. I really wanted to get them and Trenchard together for Chinese food or something so that Trenchard could learn LISP and the other guys could learn APL. But it never happened. They saw IBM as the big, evil monopoly that made mediocre software (like Fortran.) And he probably saw them as a bunch of college hippies. I could never get them together and couldn't explain either language well enough to excite the other camp.
Years later I lost all involvement with APL and LISP but would occasionally think about them. Common LISP was standardized and I was relieved to read in Guy's book about lists and arrays and how a modern programming language needed both. He at least got it at some point. But the array functions in Common LISP are still just a few primitive map and loop constructs. I got a Macintosh and found out about MicroAPL. My excitement diminished as I played with it and found lots of bugs.
So today I'm very excited to stumble upon this forum and learning that APL is still admired, used, and being developed. I got NARS2000 running on my Mac under WINE and I'm catching up on heterogeneous arrays, nested arrays, and the new functions and operators added since APL 360. And when I read that NARS had indexed functions, like VECTOR+MATRIX, I cheered. I'd wished for and wondered why APL didn't do that in 1975.