« first day (2093 days earlier)      last day (564 days later) » 

B. Wilson
B. Wilson
12:51 AM
@Adám If you've been sufficiently brainwashed by group theory, under feels fairly natural: conjugacy classes blah blah.
 
 
2 hours later…
B. Wilson
B. Wilson
2:39 AM
@Adám These design principles sound interesting. I would guess that BQN might also have something to say on the matter.
Hrm. I would guess that the monadic forms would just be identity, like +, on the logic that they "default" to a left argument that's the identity element.
@LdBeth The biggest annoyance I see is when the left argument is computed.
 
 
2 hours later…
Adám
Adám
4:44 AM
@B.Wilson But the default left argument isn't ever the identity element (for that side). + has 0J¯2×11○Y as default left argument, if you want.
 
B. Wilson
B. Wilson
5:10 AM
@Adám Ha! Nice. I just thought + was identity, but it's actually complex conjugation. Cool.
Reciprocal can be thought of as having an implicit 1 identity argument, though, right?
Hrm... Currently, for +-×÷, the monadic forms are all their own inverses. *⍟ also have "obvious" default left arguments.
∨∧<> don't even have monadic forms.
> what would the monadic form do?
I took this somewhat as a logical riddle, but is it more of an open design question?
Naively, the current definitions seem perfectly natural even in conjunction with a redefined ÷ and -.
 
Adám
Adám
6:08 AM
@B.Wilson With swapped ÷ and -, one might want × and + to be reciprocal and negate, taking Y⁻² and −2Y as default left argument.
@B.Wilson It goes along with APL.
@B.Wilson That's a right-identity of ÷
@B.Wilson Probably nothing related, although I had the idea of all comparison functions (did you forget ≤≥=≠?) taking ⊃0⍴⊂ as default left argument, something I found actually useful, especially for golfing.
 
 
5 hours later…
Jeff Zeitlin
Jeff Zeitlin
10:59 AM
@Adám - Not a fan of repurposing in general, but unless/until tradfns are completely dropped from the language - at which point I'd be hesitant to call it APL as opposed to an APL-derived language - I would consider repurposing that would break compatibility to be a bad idea. (more...)
(...more) I'll grant that locked tradfns are probably arguably an edge case currently, and I know we've discussed the idea of 'binding' a workspace to the interpreter to (a) conceal source and (b) provide "single file" APL programs, which would essentially obsolete the locked fn - but IIRC it's still part of the defintion of the language.
 
 
4 hours later…
Jeremy gee
Jeremy gee
2:48 PM
@Adám Added to learning resources under Video Series.
 
 
1 hour later…
ArbIn
ArbIn
4:15 PM
5 5⍴25?10 is giving me trouble. I want to get a 5x5 matrix with 25 random numbers from 1 through 10. Anyone have any idea why this is generating a Domain Error?
 
dzaima
dzaima
@ArbIn a?b gives a distinct random numbers in the range 1 to b. you want ?5 5⍴10
 
ArbIn
ArbIn
Thank you, @dzaima. I understand
Is there any major difference between ?5 5⍴⍳10 and ?5 5⍴10
 
dzaima
dzaima
the distribution. e.g. for ?5 5⍴⍳10 the top-left number will always be 1, the one right of that - 1 or 2, etc; ?5 5⍴10 will be uniform from 1 to 10 for all result numbers
 

« first day (2093 days earlier)      last day (564 days later) »