@gwa000 It looks like a tutorial on using ⎕SM and friends for controlling the CLI windows and stuff. As such, I'm not too surprised it's stripped out of the Windows version.
BTW, if you run )lib you can see a list off all available workspaces sitting under your WSPATH.
@B.Wilson I had to wait until I was just over a year before attending my first APL conference. Rosa is definitely coming to one when she's just over 9 months :-)
It's actually something I wrote! Vaguely remember reading about this algorithm before knowing APL, and then today it just popped out of my hands naturally :D
Didn't realize I had Aaron-ed until a couple seconds later.
Trying to solve the generalized problem, where depths can increase by arbitrary integer amounts.
There's an ugly linear scan solution to first normalize the depths: ¯1+≢∘∊¨{∪⍺[⍳1+⍺⍸⍵],⍵}⍥,⍨⌿U⌽¨,⍀d. However, I don't like it.
The garbage at the tail is just trying to do a left-associative scan, really.
(Actually, left-associative scan is something I want from time to time. Maybe some primitive that improves this could be on the WANT list?)
Yeah, my conceptualization is that Scan has 2 (implicit) parameters at play: associativity and sublist-direction. My naming abilitiy on the latter sucks, but each has 2 options, giving your total of 4 possible combinations.
Given that left-associativity on reductions is simple enough with Reverse and Commute, I feel like the design space permits something that lets us leverage that as well. Maybe?
Apologies, I'm not being crisp in my expression here. We have the associativity of the (dyadic) function application. In 2 and 4 above, it's right-associative, just like normal APL, with 1 and 3 applying in a left-associative fashion.
@B.Wilson Yes, sure, but you can define all of these in terms of ⍨ and ⌽. The idea here is that you won't need to, and that it can be more efficient not to.
Both ⎕CT and ⎕DCT work the same way, right? Just the difference in default value and both work for different datatypes... Otherwise, everything is exactly the same?
@Adám Yes yes, that's what I meant, apologies. Basically with ⎕FR←1287 and ⎕FR←645 for DCT and CT. Otherwise, do both work the same or is there a difference?
> Given an integer scalar or vector representing the number of sides on each of a set of dice, return a histogram showing the distribution curve for the possible totals that can be rolled using those dice.
@B.Wilson I think it's the key ⌸ giving back a matrix with spaces in it, rather than a column of vectors?
@Richard The last thing to note, which is already being shown here, is the relationship between indices of a multidimensional array (in ⎕IO←0) and the use of encode ⍺⊤⍵ and decode ⍺⊥⍵
@B.Wilson I think most commercial applications use default, and then people specify in the code if they are doing something different (rather than using DEFAULT_IO), but that's my hunch I don't actually know
A valid point well made. I do think though it's a reasonably common thing to want to do, and that whilst ,⌿⊢⊤∘⍳×/ is a clever solution, ,⍳X I'd wager is the natural spelling most would arrive at.
I'm getting DELETE PENDING errors intermittently when my program deletes a file and then quickly thereafter creates a new file with the same path as the deleted file. In the Dyalog APL session window, it's reported as FILE ACCESS ERROR. I can re-create the error with Dyalog APL 16.0.35389.0 32 Cl...
