The closest language to APL/J/K is NumPy, the numerics library for Python. It was deliberately designed in the spirit of an APL, but without all the squiggly characters.
I so liked the APL squiggly characters (mostly Greek letter based glyphs). Made for very succinct code. Combined with right-to-left precedence resulted in very low overhead to write and read code. Most people who complained about it APL as “write-only” seemed to be outside the community of serious APL programmers. Once you learned to read the language it was very easy to comprehend at regular reading speed...
In contrast NumPy seems absurdly wordy, and waiting for the completion menu in an IDE tends to derail my train of thought.
However I do tend to like NumPy for being APL-like, so it’s not a damning contrast as far as getting things done.
I get it, and I wrote that even though I'm a fairly big J enthusiast. It's fun to solve a tricky problem by cranking out two or three rows of line noise. :) Numpy is wordy, but you can still program in terms of arrays -- it might not feel like quick symbolic magic, but the underlying paradigm is (almost) the same.
I really don't understand why people dislike this. As long as the domain and input/ouputs are clear.. having a Δ is not worse than a `delta` or `computeDeltaOfPair` (actually the shorter the better to me..
But once you use grade up, it's easy to remember. I used it like a year ago when playing with APL and haven't touched it since then and still remember the symbol.
Yes, there are a handful of languages where you can operate on arrays (without looping), including Fortran. But this doesn’t make them true array languages like APL.
