« first day (1608 days earlier)      last day (1047 days later) » 
01:00 - 21:0021:00 - 00:00

9:04 PM
I remember writing sin⁻¹(abs(sin(x and stuff like that.
@0xACE Well, that's Even←{⍵/⍨0=2|⍵}
I've though about assigning a keyboard combination for <Home>(<End>)
 
i have a hotkey for reformatting foo bar baz to: foo(bar, baz) // though it also takes optional commas, but then it requires commas on each argument
 
@Adám or maybe <shift>+<home> ( <right> <right>
 
It might be nice if typing an unpaired ) (anywhere but at the beginning of a line) inserted a ( at the far left.
'LL' '(' 'RL' ')'⎕PFKEY 4 works under Windows; makes F4 do <Home>(<End>)
 
@Adám my android app has a button for that (that also handles putting the parenthesis inside a dfn (but apparently not multiple statements :| ))
 
@dzaima Once parens can have diamonds, it doesn't make sense to handle it anyway. Is that the …) button?
 
9:11 PM
@Adám it's probably what you still want though; yes, it's ...)
 
^ does the trick under RIDE.
 
@dzaima (oh right, that idea was completely stolen from calculator++)
 
that seems to () the entire line... but i can figure out how ot make it ) from where the cursor is
Haha
 
@0xACE Oh. Maybe it'll work to select to the beginning of the line, delete that, type ( then paste. Let me try…
 
wouldn't that hijack the clipboard?
i just used: )<LL>(
 
9:16 PM
@0xACE But then you end up at the far left.
 
OK, my first attempt at solving a problem in APL. The problem I chose is finding the first character that is duplicate in a string. Now that I have the problem solved, I would like feedback (as I am a total beginner) as to whether I am doing something the hard way:
```
CountChars←{u←∪⍵ ⋄ s←⍵ ⋄ 2(≢u)⍴u,{+/s∊⍵}¨u}
FirstDup←{1↑(⍵[2;]>1)/⍵[1;]}
FindFirstDup←FirstDup∘CountChars
```
Then:
```
FindFirstDup 'the amazing thing'
```
 
@GreenSaguaro Put your code in a separate message, and press Ctrl+K before Enter.
 
OK, will do
 
@Adám you don't need to delete+paste - when something's selected, ( will just wrap the selected thing
 
You can edit your message (press Up)
@dzaima In RIDE?
 
9:17 PM
@GreenSaguaro have you seen ≠?
 
@Adám yeah
 
@dzaima Oh wow, TIL.
 
CountChars←{u←∪⍵ ⋄ s←⍵ ⋄ 2(≢u)⍴u,{+/s∊⍵}¨u}
FirstDup←{1↑(⍵[2;]>1)/⍵[1;]}
FindFirstDup←FirstDup∘CountChars
 
Thank you, that's easier to read.
 
@dzaima (apparently not from the keyboard shortcut thing though)
 
9:19 PM
@GreenSaguaro OK, so maybe I can show you a different approach?
 
@Adám Sure
 
@Adám the link for ≠Y is broken on aplcart, for some reason it is 1atest not latest
 
@rak1507 Thanks. Fixed.
@GreenSaguaro Here:
      {{≢⍵}⌸⍵}'the amazing thing'
2 2 1 2 2 1 1 2 2 2
      {2≤{≢⍵}⌸⍵}'the amazing thing'
1 1 0 1 1 0 0 1 1 1
      {⍸2≤{≢⍵}⌸⍵}'the amazing thing'
1 2 4 5 8 9 10
      {⊃⍸2≤{≢⍵}⌸⍵}'the amazing thing'
1
      {⍵[⊃⍸2≤{≢⍵}⌸⍵]}'the amazing thing'
t
@GreenSaguaro rak1507 mentioned :
      ≠'the amazing thing'
1 1 1 1 1 1 0 1 1 1 1 0 0 0 0 0 0
      ~≠'the amazing thing'
0 0 0 0 0 0 1 0 0 0 0 1 1 1 1 1 1
      {⍵/⍨~≠⍵}'the amazing thing'
a thing
      {⍵∊⍵/⍨~≠⍵}'the amazing thing'
1 1 0 1 1 0 1 0 1 1 1 1 1 1 1 1 1
      {⍵[⊃⍸⍵∊⍵/⍨~≠⍵]}'the amazing thing'
t
 
@Adám On tryapl.org, from the hover hint, it is not exactly clear what ⌸ does. Can you explain that one?
 
@GreenSaguaro I once gave a lesson on it.
There's a lot of jargon in what I write there. You can look up all the unfamilliar terms on apl.wiki
@GreenSaguaro Also, on TryAPL's Primer, you can click on a glyph to get some example uses, and you can type ]help ⌸ to open the full documentation in a different tab.
 
9:32 PM
@Adám Cool thank you. I can work through your examples fairly simply as long as I can read docs for a given glyph.
Oh wow! ..I didn't know that I can do that
 
I still wish tryapl had the docs ride + the windows ide have on hover
 
@rak1507 And thank you for mentioning the monadic not equals ...I will have to look at the docs to understand what that is doing
 
@Adám bit overcomplicated no? {⊃⍵/⍨~≠⍵} does the trick
 
@rak1507 That's exactly what you get when clicking. Is it the lack of hover-to-float-popup that bothers you?
 
@Adám when I click on it it just types the symbol
 
9:34 PM
@rak1507 Sure, but thought [⊃⍸ would be easier to understand.
@rak1507 On the Primer.
 
oh right, didn't know about that, that's good
I still think being able to hover over symbols and see definitions is super useful as a newbie
 
We got rid of the old hover-to-reveal because it was notoriously finicky and to be more inclusive of touch devices.
 
Wow ...the ]help (glyph) worked great. I had NO IDEA from the hover tip that monadic not equals was a unique mask
 
I find that both RIDE's and the IDE's language bar tips end up obscuring my view a lot.
@GreenSaguaro Oh dear, I see now that it wasn't updated. I'll fix that. Thank you!
 
@GreenSaguaro this is what you get if you hover over it locally
 
9:37 PM
Yes, and in RIDE too, but not on TryAPL.
 
yes
 
Huh, and on the stand-alone browser language bar too. Just not TryAPL.
 
Because that derives from the the browser language bar.
 
yeah
 
9:41 PM
@rak1507 Maybe we should advertise the Primer a little on the Intro page?
Like where it says
 
maybe
 
> Move to the Learn tab and get started!
 
yep
 
It could say
> Move to the Learn tab and get started or visit the Primer tab to explore on your own!
 
yeah, that's definitely an improvement
 
RGS
10:17 PM
@0xACE I've been working on the final videos :) If you have any sort of feedback on what you have watched, please do let me know
@rak1507 I'll take that also as a compliment! Which ones? And the thing is, I don't really know how/where :P
 
RGS Great :D, honestly majority of the videos were fine as is. but I think there were 2 videos where you implemented/talked about algorithms that were completely unfamiliar to me as a listener. I think I need to research the problems first before i give you feedback. And i'll try to write notes from now on... But really, you are doing great
 
@RGS all, and here/dyalog website/twitter/whatever
 
(honestly the biggest issue is probably me looking at the videos while outdoors. I only have access to a phone and using RIDE from it is a bit cumbersome...)
 
RGS
@0xACE oh wow, that is some hardcore commitment
(Wasn't there a "safe executor" logo somewhere?)
 
RGS
10:25 PM
Your reply looks weird but I found one in your GH repo
 
@RGS Why?
 
RGS
@Adám potential twitter user logo for the bot :P
 
I don't think the fact that it is sandboxed is very interesting. I'd use this guy's profile pic.
 
RGS
I think I agree :P
 
10:49 PM
Aha, ShapeCatcher has a simple API.
 
@Adám you probably don't want to spam it with every character you type though
 
No :-(
 
RGS
Alright, alright, I'll finish the neural nets series
And then I'll do one on CNNs
And one of you can take that and implement a CNN that recognises the APL glyphs :P
 
I'm curious how that will fare compared to the tech ShapeCatcher is using.
 
RGS
11:16 PM
How are we going to fix the monospaced font issues?
 
Hm.
 
RGS
(Also, does anyone know what was the pastebin-like service that the bot here used to paste large results?)
 
0x0.st?
 
@RGS One (not so nice) possibility would be to have the bot render a picture and tweet that.
 
RGS
@dzaima yes! thanks
@Adám I could also turn ]box off... but that wouldn't be great as well, I think
 
11:25 PM
@RGS Maybe replace all characters with fullwidth forms? The line drawings would have to use +-| like in the old days.
 
RGS
I'll think about it tomorrow. I'll turn the thing off (the bot isn't hosted yet; I'll accept suggestions)
Yeah that's an alternative.
 
Fullwidth even handles negative APL numbers:  ̄1.9E234
 
Apparently disp-family functions don't support alternative boxing chars (it's natively supported in J as one of system settings)
 
RGS
wdym?
@Bubbler one of the dfns.xxx one does
Don't know by heart which one, though.
 
@Bubbler They used to. We removed the option since "everyone" supports line drawing chars these days.
 
RGS
11:32 PM
Wait, I was pretty sure one of them did, but now I can't find it :P
 
Anyway, using +-| wouldn't help at all, since they have very unequal width, and space is generally super thin too.
 
The only one left now seems to be dfns.box, which isn't quite useful for displaying arrays
 
RGS
ah right, so there was one left
 
@RGS 3 additional options: 1) Return a 1-liner expression for the result. 2) Use JSON. 3) Use APLAN.
 
Also agreed to the point that it's useless outside of monospace context
 
RGS
11:34 PM
 
Spaces are not only super thin, runs of spaces are often stripped down to a single space
 
Yeah, but that can usually be mitigated by nbsp
 
RGS
Well, I'll have some sleep now
I'll think about it tomorrow
 
Also to think about is the "active" sequence. `⋄ seems a bit OTT.
 
RGS
About the one-liner, not sure if that would be helpful if a user types ⍳3 3 and I reply with ]repr, which gives something like 3 3⍴(2⍴0) (0 1) (0 2) (1 0) (2⍴1) (1 2) (2 0) (2 1) (2⍴2)
@Adám that's a trivial change in the beginning, it's just a matter of tweaking my basic regex.
 
11:38 PM
@RGS True, but how about [(1 1 ⋄ 1 2 ⋄ 1 3 ⋄) ⋄ (2 1 ⋄ 2 2 ⋄ 2 3 ⋄) ⋄]?
 
RGS
@Adám I'd 100% go with that if APLAN were already a thing
 
Or [[[1,1],[1,2],[1,3]],[[2,1],[2,2],[2,3]]]
 
@Adám I think a picture would be better than that
 
Maybe it can return a one-liner expression and a picture?
 
RGS
@dzaima maybe that + picture? the picture is good to look at but the APL result/repr/APLAN is good for copying/pasting
 
11:39 PM
Array notation would be great if it is a common knowledge across APLers already. Not right now I guess
 
@RGS some hybrid is probably best. Also the image can be omitted if the pretty-printed version is a single line
 
Or simply restrict it to results that can fit on a line.
@Bubbler Though fairly intuitive (and whitespace doesn't matter) if written as:
↑⎕SE.Dyalog.Array.Serialise ⍳2 3
[
(
1 1
1 2
1 3
)
(
2 1
2 2
2 3
)
]
 
RGS
Alright, I do have to go now. I'll pick your suggestions up "tomorrow"
 
But of course, at least leading spaces would help:
[
⠀⠀(
⠀⠀⠀⠀1⠀1
⠀⠀⠀⠀1⠀2
⠀⠀⠀⠀1⠀3
⠀⠀)
⠀⠀(
⠀⠀⠀⠀2⠀1
⠀⠀⠀⠀2⠀2
⠀⠀⠀⠀2⠀3
⠀⠀)
]
 
11:56 PM
afaik 0x0.st has a 1 year retention policy, after 365 days it's not guaranteed that your data will remain...
 
Seems enough for tweets.
 
out of curiousty does anyone know if python or any other language has a nice REPL like RIDE/Dyalog? it find it very convenient to not only be able to CTRL-Shift-Backspace, but also be able to up to a previous line and edit it in place and re-evaluate it
Previously tinkered with lisp and emacs, but I feel like ride makes better use of it
 
01:00 - 21:0021:00 - 00:00

« first day (1608 days earlier)      last day (1047 days later) »