BQN vs Uiua vs Kap

[@WilhelmDrake]

Congratulations Mr. Hoekstra on your recent engagement! All the best from a fellow Canadian! P.S. As a life-long Hamiltonian, I still haven't forgotten the dig (however true it may be) you made at my home town on ADSP! 😠

[@opposite342]

I can't really wrap my head well with n-trains stuffs, so Uiua being a stack language really aids readability for me (also makes me prefer it the most).

[@TankorSmash]

Yeah I'm the same way with trains, it's really hard for me to parse. Uiua I'm struggling with refactoring any small file due to how it stops you from 'compiling' when types don't line up. You can't see what the output would be, so you just have to pop and duplicate stuff until it compiles. Uiua is great but that flow is killing it for me otherwise.

[@THB192]

Trains confused me in J but they make a lot more sense in BQN because they didn't screw up the combinators so the rules were actually very consistent. Whereas in J the entire meaning of the train changes depending on if it has an even or odd number of functions in it and you end up having to cap the end of trains a lot more as a consequence.

[@matwadoesgames]

As a beginner to intermediate "classic" language programmer, its astonishing how you all talk abt your combinators and n trains like it were simple stuff? How do i learn this?

[@elnico5623]

I've been using uiua for about a year now and bqn for little over half that, and trains are still kinda mystifying to parse but genuinely not hard of a concept at all

[@lokedhs]

A small note: Kap does support structural under for monadic take. However, it uses the up arrow for this (since the language itself started with the APL2 syntax and not Dyalog). That said, it still wouldn't work since monadic take returns the first element and not the first row. That's the reason you need to use the dyadic version with the left argument being a 1

[@CrazyMan42DABOMB]

happy to see more Kap on the channel!! what a cool language

[@CrazyMan42DABOMB]

also, I would totally love to see some problems that showcase Kap's laziness semantics a little bit more

[@jacoblockwood4034]

The second Uiua solution can be one character shorter by writing +⇌°⊟ (add reverse uncouple) instead of /+⍜⊢⇌ (reduce add under first reverse). I do like the under way but I think uncouple is a bit more idiomatic for this case with an array of two rows.

I guess the nicest thing would be if there were min and max functions to get the min and max value respectively and create some kind of recursively defined function that always pops these two elements until the array is empty and at the same time construct a list of averages.

[@misterpopo3736]

Love how KZbin reccomends the same videos to me. I was confused for a sec edit: Kap has traditional if/else statements? I stll miss these

[@lokedhs]

Yes, it does. They have the same syntax as C or Java. The justification for this syntax is explained in more detail in the otger video. It's also worth highlighting that if is an expression and not a statement, so it returns a value that can be used in an expression (this is similar to Kotlin)

[@LukasSmith827]

🧢

[@zcubesdemoes2191]

a=[7, 8, 3, 4, 15, 13, 4, 1]; MIN((a⋰).nth(1..(a#/2),1,true).$$(AVG)) // result 5.5 // z^3 code. // a⋰ - a ascending // a# - length of array // .nth(nth, count, bothsides). here nth is from 1 to (a#/2) // count is 1, for each n // bothsides - extracts elements from both sides of the array, here min and max of the sorted array // .$$ applies AVG function to each row, and each row here contains min and max for the nth extract // if count is provided as an array for more interesting computation // .$$ can be given a array of functions such as AVG, SUM, and more.