Through random wanderings watching videos on APL and J, I encountered a short segment on J's definition of the 'nub sieve' function (~:)
https://aplwiki.com/wiki/Nub_Sieve In J (and, according to the "APL wiki" which I presume is for other APL variants, https://aplwiki.com/wiki/Nub_Sieve) this is shorthand for ⎕cr 'Q' ┌→──────────────┐ ↓λ←λ1 ⍵ │ │λ←((⍵⍳⍵)=⍳⍴⍵)/⍵│ └───────────────┘ Apparently GNU APL currently has no binding for monadic ≠. I presume this is because it wasn't historically part of standard APL/2 syntax. However, if the monadic use of ≠ is 'free', would there be a good reason *not* to define it as such (ie., are there other contexts which would break due to a monadic form being defined)? This use is admittedly trivial and merely a convenience if it were implemented; but would it enable other more succinct phrases without violating APL/2 compatibility or have other useful applications if defined natively? -Russ
