@EriktheOutgolfer Just like a one-liner dfn For byte counting purposes, it would be neat if TIO would merge the last line of Header with the first line of Code (possibly on condition that the last line of Header ended with ←)
So for a b c d e f only c and e are always called dyadically. b, d, and f are called in whichever way the whole function is called. a is always called monadically.
So 10∘⊥2∘⊥⍣¯1⎕UCS is (10∘⊥) (2∘⊥⍣¯1) (⎕UCS) and the middle one is (failing to be) called dyadically. 10⊥2⊥⍣¯1⎕UCS is (10) (⊥) (2) (⊥⍣¯1) (⎕UCS) so ⊥⍣¯1 and ⊥ are called dyadically, which is exactly what you want.
@EriktheOutgolfer (10∘⊥) (2∘⊥⍣¯1) (⎕UCS) tries (and fails) to call 2∘⊥⍣¯1 with the result of 10∘⊥ as left argument and the result of ⎕UCS as right argument.
⍺(f g h)⍵ ←→ (⍺ f ⍵) g (⍺ h ⍵) ⍝ dyadic (fgh) fork
⍺(A g h)⍵ ←→ A g (⍺ h ⍵) ⍝ dyadic (Agh) fork
⍺( g h)⍵ ←→ g (⍺ h ⍵) ⍝ dyadic atop
(f g h)⍵ ←→ ( f ⍵) g ( h ⍵) ⍝ monadic (fgh) fork
(A g h)⍵ ←→ A g ( h ⍵) ⍝ monadic (Agh) fork
( g h)⍵ ←→ g ( h ⍵) ⍝ monadic atop
@Adám I did that because, well, outsiders who didn't participate may want to review those lessons too, and that would make them more accessible (note how I included the appropriate key words to each lesson, that's important)
Not long until Morten Kromberg, Roger Hui and Aaron Hsu head off to @FnConf in Bangalore. For the full schedule see https://functionalconf.com/schedule.html
@Adám is it possible to enlist every possible combination of a given string while keeping it's original order? so for 'abcd' I get 'a' 'b' 'c' 'd' 'ab' 'bc' 'cd' 'abc' 'bcd' 'abcd'
@Adám I imagined so. I'll need to get clever for that. I'm currently trying to use ⍷ to find where the combinations of vowels are, but I'm having trouble making that function you gave me work properly in a train
If I do f←⊃,/⌽(,\⌽)¨,\⌽ I get a syntax error if I call f 'abcde' D:
@Adám I'm trying to check whether the string 'aa' ∊ <anotherString> but that returns 1 1 (because I think it's checking for the scalar 'a' 2 times?), how can I check for the entire string 'aa'?
@Adám good, good. I'm also trying to do something (I don't even know if it makes sense tbqh) like: zip 'aeiou' to itself n times where n = ⍴⍵ (to hopefully get something like 'a' 'e' 'i' 'o' 'u' 'aa' 'ae' 'ai'...). I tried {(,'aeiou')⍣⍴⍵}'abcde' but it doesnt work
I imagine ≢⍵ instead of ⍴⍵ would be the same result btw?
@Adám f here being ,? {,⍣(⍴⍵)⊢'aeiou'}'abcdef' raises a rank error. IDK if I made it clear enough, but what I wanted to accomplish is zipping 'aeiou' to itself ⍴<input> times. The logic I'm using here is that if I have every vowel combination up to input size I'll have all combinations that could be in that input?
@Adám well in this case, I mean to take 'aeiou' and getting 'a' 'e' 'i' 'o' 'u' 'aa' 'ae' 'ai'... until I have all possible combinations up to input size.
@EriktheOutgolfer When nesting was added to APL (in the beginning there was only rank, no depth) there were two schools. One held that 4≡⊂4 while the other allowed it to be nested. IBM went with the former, and most vendors followed IBM. IPSA (and J) went with the latter. I agree with you that that 4≡⊂4 is a mistake.
@EriktheOutgolfer The problem arose since IBM introduced stranding (the ability to juxtapose things to create a vector) as an extension of old APL's vector notation: If 'abc' 'def' is (⊂'abc'),(⊂'def) then shouldn't 'a' 'b' be (⊂'a'),(⊂'b') too? Yes, you say. OK, but if (1 2 3)(4 5 6) is (⊂1 2 3),(⊂4 5 6) shouldn't 1 2 be (⊂1),(⊂2) too? (!)
IPSA, on the other hand, did not extend the vector notation to, but instead introduced a "link" primitive (we'll denote it here with ;) so that 1 2 3;4 5 6 is (⊂1 2 3),(⊂4 5 6). The definition of link is {(⊂⍺),⊆⍵}.
However, "link" doesn't work for all argument the way the juxtaposition conveniently does. Thus (1 2;3 4);(5 6;7 8) gives (using stand notation) ((1 2)(3 4))(5 6)(7 8) and not ((1 2)(3 4))((5 6)(7 8)) like you might expect.
@Adám I tried to add it to the sandbox to mess with it there but then it stopped working... I guess it can't handle multiple rooms for some reason. Should be back now.
Yes. +/ is a derived ambivalent function. The monadic function is plus-reduction (i.e. sum) and the dyadic function is windowed sum, as in sliding windows of size ⍺ (shorthand for "left argument")
@barrycarter
line(1,0) : error AC0505: error (VALUE ERROR) executing line "⎕←{'(',⍺,⍵,')'}/ h e l l o"
^
Complete: 1 error.
DOMAIN ERROR: There were errors processing the script
'#'⎕NS ⎕FIX'file:///home/runner/.bin.tio.dyalog'
∧
Real time: 0.030 s
User time: 0.004 s
Sys. time: 0.022 s
CPU share: 87.13 %
Exit code: 0
@barrycarter No, N-wise reduce does not lower the rank.
@barrycarter Right. This may be handy in golfing…
Notice that / goes along the trailing axis, i.e. the it reduced the rows of the matrix. It has a twin, ⌿ which goes along the first axis, i.e. the columns of a matrix.
