@barrycarter
line(1,0) : error AC0505: error (SYNTAX ERROR) executing line "⎕←2+\ 3 4⍴⍳12"
^
Complete: 1 error.
DOMAIN ERROR: There were errors processing the script
'#'⎕NS ⎕FIX'file:///home/runner/.bin.tio.dyalog'
∧
Real time: 0.032 s
User time: 0.007 s
Sys. time: 0.021 s
CPU share: 86.61 %
Exit code: 0
@EriktheOutgolfer No. (I actually just wrote a new operator for that yesterday.) But if your arrays are regular (non-ragged) you should be using high-rank arrays instead, and then you can use ⍤.
@barrycarter Most arithmetic functions are "scalar" meaning they penetrate to the very leaves of the arrays. ¨ is meaningless for scalar functions.
@Adám very sorry, but I have to off wifi now. @ThomasWard said he was trying to set up an instance of the bot, so hopefully that works out... I should be able to respond to question about it on my phone if necessary :)
In fact, you can have a monadic operator which takes an array as operand, and derives a monadic function. It will have an appearance very similar to a dyadic function.
@barrycarter
line(1,0) : error AC0505: error (VALUE ERROR) executing line "⎕←cos(x)⍣=1"
^
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: 0.009 s
Sys. time: 0.016 s
CPU share: 85.01 %
Exit code: 0
@barrycarter
line(1,0) : error AC0505: error (VALUE ERROR) executing line "⎕←2○(x)⍣=1"
^
Complete: 1 error.
DOMAIN ERROR: There were errors processing the script
'#'⎕NS ⎕FIX'file:///home/runner/.bin.tio.dyalog'
∧
Real time: 0.024 s
User time: 0.005 s
Sys. time: 0.017 s
CPU share: 87.37 %
Exit code: 0
@barrycarter
line(1,0) : error AC0505: error (VALUE ERROR) executing line "⎕←2○x⍣=1"
^
Complete: 1 error.
DOMAIN ERROR: There were errors processing the script
'#'⎕NS ⎕FIX'file:///home/runner/.bin.tio.dyalog'
∧
Real time: 0.028 s
User time: 0.006 s
Sys. time: 0.018 s
CPU share: 87.48 %
Exit code: 0
@barrycarter It doesn't return it. It calculates the sine when the left arg is 1. To get a "sine" function, you'd have to compose a fixed left argument of 1 with the ○ function as follows: sin← 1∘○
⍨ seems very simple, but it has some neat applications.
@Adám so basically it should be implemented with a single register for left operand/arg (and final result), temp register for right arg and a call stack for functions?
@Uriel That's a possibility if monadic functions are still prefix. If they would be suffix, it would be 9⍳×\⍨ if monadic operators would still be suffix, otherwise it would be 9⍳⍨\× i.e. just reversed of the current.
@Uriel
line(1,0) : error AC0505: error (SYNTAX ERROR) executing line "⎕←×\⍨⍳9"
^
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: 0.007 s
Sys. time: 0.019 s
CPU share: 88.25 %
Exit code: 0
@Adám that was my initial thought, then I realized it would look really weird with functions like log or shape, and practically, much similar to stack based langs. so I figure ×\⍨⍳9 would be better
All these autocorrect typos... I think I'm starting to miss T9
@Uriel Yes, and there is just no way around any APL (meaning no-precedence) having cases which seem strange to people at first.
@Uriel But remembering 1. operators, 2. monadic functions, 3. dyadic functions isn't so bad, especially if we go left-to-right, and there just isn't any other way than doing the monads first.
In A=⌽A we must evaluate ⌽A before A= even in lefty-APL, because = needs a right argument.
@Uriel It is totally confused about the directionality of the characters. I entered 1+2×3 but it changes the display order!
@Uriel For the last line, I entered 3×2+1
APL just looks at the logical order of characters. Chromium renders according to some obscure Unicode rules. I had set CSS {unicode-bidi:override;direction:rtl}
@Adám oh makes sense. yeah, I used to write Hebrew strings reversed in Python so they would show rtl in the editor and console. then my editor started supporting rtl, so I just entered them normally and reversed before printing
@Uriel The Windows IDE will display in logic order, but RIDE is just a webbrowser in disguise, so it does the whole Unicode magic, with sometimes humorous results.
@Uriel You already can change RIDE. The IDE doesn't use CSS.
@Uriel No, think you can use an old version of RIDE with 14.0, but there is no reason to. The old RIDE was pretty awful to look at. The new one looks cool, but sill doesn't have nearly as much functionality as the IDE.
@Adám well, but if it would be all Hebrew text, anyone would tell you that the right alignment surely is more intuitive, even if the right most item is at the top of the matrix
@Uriel I don't think so. It the programming language is LtR (reading order) then multiple Hebrew strings should appear LtR even if each one would be RtL internally.
Then ⊣/ would always pick the leftmost and ⊢/ the rightmost.
