« first day (1220 days earlier)      last day (1420 days later) » 

Bubbler
Bubbler
7:04 AM
It's so satisfying to resolve red links, but the problem is that the total of red links will increase for a while as the pages with APL built-in template increase.
 
 
4 hours later…
dzaima
dzaima
10:57 AM
new behavior of @ & ⊇ with ⎕VI←1:
https://dzaima.github.io/paste/#0bVJLS8NAGLx/v2JuqSiabTZt8CCtj@Jb8e0xqSsGNYF2pYh4EVFRIxTx6KV49qKe/S3uL/HbREGr7GlnZueb2d0xCJTN9aW5e6hTfZzGUDLXZ4wN9IHZ819QmJtzRuFb3IdvstcvdiqnLWUl/XRjvaDtKvP4wsWD/MdpYmaWWO2EUdOp5ZaS9TZJDz@F4cTkFDWiaZbPzS8sLtFyc2V1jdY3Nre2rcGJuegKF8J1Bz@6V@bi3mRvp7WS5PFskb0E7FpsOEnuHvCk7FWa23dLw2SPnLlwMNmTGEAnPTrYQaSQxE01hOhII0lZ91THsdIYHSEuKADbk8rg6MLluuRBeIx68EiCSwtXQpIP4TPKnagCVBitoEJViCqjVVQpAEcSboCAiGwavafaCrqTYidNHI22UofQKeJEq9auail0Yr2HNFEIk5TVrSG0meaDcbKPWDtt7MaJGh7OL9ip5V@hxw1eJIUI7Yv8ZsrftyW5WshF5Cc#APL18
 
 
2 hours later…
dzaima
dzaima
12:28 PM
err, ⎕VI←0 ⋄ a←100 100⍴⍳10000 ⋄ i←⍳⍴a ⋄ 10000⎕htime'i⊇a' is actually more like 0.1464 → 0.1954 :/
0.1464 → 0.06986 now :)
 
 
3 hours later…
Avi F. S.
Avi F. S.
3:49 PM
It's bothered me for a long time that one can't simply add another row to a vector with ⍪. I understand why, or so I think, one needs to do ⍪[.5] the first time around, and then one can continue doing ⍪ plainly. But the inconsistent format makes it hard to, for instance, extend functions which are intended to create a matrix from an empty vector on row at a time.
Of course, one can create a simple vector one row at a time and then reshape, but that requires knowing the shape after the fact which obfuscates the underlying whatevers again. But it occurs to me that if I find it so bothersome, there's likely another way known to the more experienced APLer!

TLDR: What would you guys do to extend a matrix row-by-row, from nothing?
 
dzaima
dzaima
@AviF.S. start with nothing, and ⍪ the rows. :P
 
Avi F. S.
Avi F. S.
But that doesn't work if one starts with nothing, in my experience
 
dzaima
dzaima
@AviF.S. you have to pick a specific nothing
 
Avi F. S.
Avi F. S.
In that case, ⍪ acts synonymously with ,...
 
dzaima
dzaima
@AviF.S. wrong link?
 
Avi F. S.
Avi F. S.
3:54 PM
Sorry, fixed it!
 
dzaima
dzaima
@AviF.S. yeah, i tried that first too.. a←⍪⍬ might make more sense, but doesn't work either
if you're building row-by-row, there's probably the the potential of 0 rows, so at some point you'd need to define what no rows looks like
 
Avi F. S.
Avi F. S.
Took me a second to figure out what you were doing, you sly fox! Where did you see ⍴ 0 n that such mischievous magic would've ever occurred to you?!?!
 
dzaima
dzaima
@AviF.S. there have been conversations here about "no rows n columns" matrices here, alongside talk about prototypes, it seemed relevant to try
 
Avi F. S.
Avi F. S.
@dzaima That makes sense as a potential explanation, but can't be right. Pretty sure it's about the default axis. The right axis seems to be assumed default, for some odd reason, because the following works:
Ah, thanks! Link being created, btw

Also, lost WiFi; sorry for delay
Wifi very poor :(
 
dzaima
dzaima
@AviF.S. it choosing "right axis" is just what ⍪ does. , chooses "left axis", and you can specify the axis manually
 
Avi F. S.
Avi F. S.
4:05 PM
So pretty sure it's an axis issue, but really wish [0.5] were default when one uses ⍪. Would be super curious to hear an explanation of why, on a vector by default, ⍪ assumes the axis that would make it behave identically to ,
@dzaima Super neat github.io, hadn't seen it!
And I agree with that, but I'm not sure I understand how it relates to the behavior demonstrated in the link...
 
dzaima
dzaima
@AviF.S. by changing the rank of a (by "adding more shape items to the right" for the lack of me knowing proper terms) you changed what is that "right axis" that ⍪ chooses to use
 
Avi F. S.
Avi F. S.
@dzaima Curious how to use your github.io! Figured out how to edit, but how do you get a line to execute, and how do you indent the next line?
Anything else I'm missing with using it?
 
dzaima
dzaima
@AviF.S. it's pretty much a pastebin but the contents are stored in the link, APL syntax highlighting is just a fancy thing i make for purposes of prettiness
 
Avi F. S.
Avi F. S.
@dzaima Uh oh, still not sure... Where would more info be to read?
@dzaima Under 'axis' on the wiki perhaps?
 
dzaima
dzaima
 
Avi F. S.
Avi F. S.
4:11 PM
@dzaima Ah, I see! So did you simply copy/paste from your actual repl (RIDE, or whatever)?
 
dzaima
dzaima
@AviF.S. yep. might add a more REPLy display at some point, but evaluation of any code is definitely not its purpose
(will have to go in a bit)
 
Avi F. S.
Avi F. S.
@dzaima Thanks, will give those a shot! Though, tbh, I'm not sure I understand axes well enough in the first place to understand the catenate-specific docs. I skipped those bits in Legrand for some reason...
So might have to go back to those first, is what I meant!
@dzaima No worries! Thanks a bunch for answering my questions!! Super appreciate it :)
Even if I don't understand the bit, ⍴ 0 n is a really elegant solution!
 
dzaima
dzaima
(might have messed up some of my previous messages) so ⍪ seems to be equal to ,[1]. If you'd want 1 2 3⍪4 5 6 to be equal to adding making a matrix out of 2 rows, should (2 2⍴1)⍪(2 2⍴2) be equal to making a 3D array out of 2 layers (which would defeat the purpose of catenating 2 matrices vertically)?
anyways gtg
 

« first day (1220 days earlier)      last day (1420 days later) »