The APL Orchard

« first day (631 days earlier) ← previous day next day → last day (2013 days later) »

12:28 AM

https://tio.run/##fZLPSsQwEMbveYo5JuzKJtWj4oKPsXhI/yBKqaVdkGXpRWXR1YLgI4i4Ny8Wpcc@Sl6kO5NEF8RakqFNZ37zzUd0nu7FC51envV9fF6UZvVkVs8HEJj6Q4GE7p2iBP/OmI4vTpI0LQEzYcnN3duk23Alzf1rKPCLy8NQiElIoLoZEbOC4Yfh1nmeLhwvS66okpu6xTXlEbaypBYKncWASCEwgHm4BZc0ppyxWV9TbfUL7ausVH6MM2lhHj@1RX7JI@zTLHHjmYNVottElBJhym5U/PGH8CLJEz33wjGqEaJm6tSK@1HezAJ34mcz6xs63N@l8cxZIOwEVjaUiQdzNcX2NPN/Ng5ZW8518Q1yck39EkjEtTzwXcls1w/NHWjCWN9z9BjvgRKeqiQt9BTkFg

@arcfide on the subject of mazes

rand finds random adjacent cells, apply @'s the next n to the board , newboards, current cell, and current n are then returned, so that a repeat function can be applied incrementing n

how can this approach be made more APL-ish

note, there is no backtracking, and it doesn't complete the maze as I've run out of time at the moment

so please don't spoil anything

if possible. Please comment on what I've done thus far so i can extricate the relevant learnings all up in the heezy

bit.ly/2P8uGk4 here's a commented version if that helps

PLEASE: no spoilers. I have an idea for backtracking, but I just don't have time atm

Quintec

12:49 AM

The set function seems extremely unnecessary to me, you should just put the set code into the start function.

Especially since that is quite literally what @ is supposed to do.

3 hours later…

nathan rogers

3:56 AM

bit.ly/2AfnZ7Q

it does a maze

through recursive backtracking

gonna modify it for any size board

but I have a question

in the repeat function, how do I phrase the power so that it continues until (∧/⍣2)0=⍵

where ⍵ is the board?

I can't quite get that one

https://bit.ly/2AgcblK
actually this is a better encoding. I had other data in there for debugging

nathan rogers

4:29 AM

https://bit.ly/2CRoDek
maze generator. call gen with an integer, it will create an n by n maze with random start and end points marked by the number 1 on the result

would love some feedback, this is probably horribly ineffecient, I'm going to try to refactor it, but I would love some thoughts, and an answer to the above power question

in the init function

nathan rogers

5:05 AM

https://bit.ly/2CQG3HY
I redid the backtrack to read a bit better. less golfed but more clear what's happening

H.PWiz

@nathanrogers If I understand what you want, you can do repeat⍣{∧/,0≠3⊃⍺}

nathan rogers

I'll try it in the morning

one other thing I would like for this to do is to backtrack the numbers instead of incrementing, so that it continues n+1 from the node it backtracks to instead of the next highest n

I'll try it tomorrow

10 hours later…

nathan rogers

3:22 PM

bit.ly/2QXLDew

so that solution isn't working for me. @H.PWiz. I'm not really sure what else to try. regarding the ⍣ boolean

Adám

4:20 PM

@nathanrogers (⍵{⍵⌷⍺⍺})¨cells‽ What's wrong with ⍵[cells]?

nathan rogers

4:58 PM

nothing I suppose :P

2 hours later…

nathan rogers

6:48 PM

https://bit.ly/2S59yds

Here's a refactored version. thanks to adam things got a lot smaller and more sensible. I still amn't quite sure what the repeat⍣(??) expression needs to be. Any advice there would be great

was able to remove a branch in backtrack therefore simplifying the rest of the code, eliminating passing the next value by inner operand to apply. was able to remove insert also

I know backtrack can be further reduced by simply taking the first instance of the value from instead of hokey sorting etc. I'll keep hacking

dzaima

7:14 PM

@nathanrogers little SE chat hack: prepend a long link with a newline (shift+enter) to be able to post it without the length limit affecting you (though markdown won't work)