« first day (2031 days earlier)      last day (630 days later) » 

3:07 AM
@AviFS Of course, but I'm pretty sure you've seen it already: xpqz.github.io/learnapl
 
 
5 hours later…
8:37 AM
@Adám Maybe as this?
supply ← {⍺←⊢ ⋄ ⍺(⊣∘⍺⍺⊣⊢)⍵}

⍝     (f supply) Y
⍝ f Y

⍝     X (f supply) Y
⍝ X

⍝     (A supply) Y
⍝ A

⍝     X (A supply) Y
⍝ X
Oh, actually that ⊣∘A case is incorrect, you should understand the idea though
 
@awagga How would you describe this?
 
supply or compute default / supply or compute overridable default
 
I don't follow. Maybe that ⊣∘A talk throws me of.
 
supply ← {⍺←⍺⍺ ⋄ ⍺⊣⍵}
If function apply, if array constant, unless left argument is present, in which case use left argument as constant
 
Can you think of a single usecase?
 
8:50 AM
Yes (f g h supply)
 
I don't get it.
 
+⌿÷≢ supply
 
??
 
(+⌿÷≢ supply) Y ←→ (+⌿Y)÷(≢Y)
X(+⌿÷≢ supply) Y ←→ (+⌿Y)÷X
I used it yesterday, in my definition of insert as well... which is a good example
      1 2 3 ('QWER' insert) 3 3⍴⎕A
⍝ Q
⍝ ABC
⍝ W
⍝ DEF
⍝ E
⍝ GHI

      1 2 3 (÷ insert) 3 3⍴⍳9
⍝ 1            0.5   0.3333333333
⍝ 1            2     3
⍝ 0.25         0.2   0.1666666667
⍝ 4            5     6
⍝ 0.1428571429 0.125 0.1111111111
⍝ 7            8     9
So, the use case for supply is providing default values or allowing for different semantics depending on the nameclass of an operand
 
I guess my head just isn't working, because I still don't get it.
 
8:56 AM
Maybe we should talk sometime?
 
Sure.
 
This makes me think tacit.help should have a feature where you can define functions/operators as part of a header. Easy way to implement f g h too... :)
 
 
2 hours later…
10:41 AM
Anyone here a Linux user who relies a lot on key compositions (e.g. <compose key> <'> <e> gives é? How annoying would it be if either those were overridden so that <compose key> e gives ∊, and/or changing the existing compose sequences to something longer?
OR more importantly, do you think there's a demand for a prefix-style input method similar to what you get on TryAPL? where instead of holding down a "shifting" key to get APL glyphs, you type one key followed by another?
 
@RikedyP compose sequences don't necessarily need to start with the compose key
 
@dzaima no, but then it sort of becomes choosing what is an appropriate key I think - although I'll have a think maybe that's something the user can configure
 
right; it's not hard to change though, just replace the leading <...> of each line
unfortunately, while <grave> <w> : "⍵" for backtick prefix works, <Alt_R> <w> : "⍵" doesn't
 
hmmmm
you can set right alt to be Multi_key through various means
then you can have that, but then you could just have the shifting layout as normal
 
right, but I wanted to see if there'd be a way to have both a regular compose key and an APL compose key
 
10:57 AM
Funny idea: If you made a hybrid with the preference of being {⍺←⍺⍺ ⋄ ⍺⊣⍵} usual would be represented as ⊣⊣
Super left tack!
 
 
2 hours later…
12:37 PM
@dzaima that's what I want to find out as well, it may be possible
 
1:14 PM
@RikedyP from my experience with my key composition base APL input method, doing anything other than writing APL code gets annoying easily.
 
 
4 hours later…
4:48 PM
@Adám You're not relying on the Array notation wiki page, are you? Think I'm going to move that to Array notation in Dyalog APL and make something less language-specific and more condensed.
 
5:34 PM
@Marshall I was going to do exactly that (both edit pages open in tabs) as evidenced by my navbox edit, but a home move came in the way. Please do go ahead and do the work!
 
Okay, will go for it when I get the chance. Hope your(?) move is going well!
 
It is indeed both mine and going well. The daily bike ride to the (Danish) office is about the same as from the Dyalog house to Bramley, so I'm getting my exercise too.
 

« first day (2031 days earlier)      last day (630 days later) »