« first day (1538 days earlier)      last day (11 days later) » 

10:52 AM
in stq golflang, 37 mins ago, by emanresu A
I might make a (not-backwards-compatible) fork of vyxal 2 at some point, it's currently better for golfing than v3 even without vyncode due to just having so much more time and careful design put into it, but at the same time there's a lot of lingering jank
@emanresuA in your opinion, how could v3 be made better for golfing?
I ask because I too think that something is making it not as good as v2
But I can't quite figure out what
I can't really say much because I haven't used v3 much
But it does feel a bit off
Ikr
But then I look at the list of elements and it doesn't seem like too much is missing
So why
Why does it feel not right?
Is there an actual problem or is this some sort of bias from experience?
I'm not really sure, trying to look through vyxal 3 answers and getting every 3-byte vyxal answer in existence
SE search be like that
3
A: Unable to search for words inside and outside code blocks together

lyxalA little bit of experimentation suggests it could be because the pascal div query is searching for a stemmed term not in the stemmed version of the answer. SE uses stemming when conducting searches, which very roughly means that it shortens terms to the absolute minimum word needed for a family o...

I looked into it once
It's not even that, it's just that I can't distinguish "vyxal, 3 bytes" and "vyxal 3, x bytes"
11:00 AM
Does wrapping in quotes work?
Nope still matches the comma, oh well
"Vyxal 3" -"Vyxal, 3 bytes"
Exlucde the 3 bytes option manually
Ah nice
After going through the codepage, I don't really have much to say. There aren't that many digraphs but I'm not sure how necessary most of them are. A few things do feel a bit odd e.g. ¤Ḍ₳⁻ᵈᵉᶠᴴЧ but I don't know how useful they all actually are
11:20 AM
The dyadic modifier lambdas exist because modifiers do different things based on arity
One thing I will say is that y'all went overboard with the modifiers. They only save bytes over having a function overload when you have exactly one builtin being used, and that's extremely rare for some of them e.g. fixpoint, key, invariant
E.g reduce doesn't automatically convert to dyadic anymore
@emanresuA key probably isn't as useful (I stole it from jelly so you'd have to ask Dennis why it's there), but the other two have practical uses
Especially invariant
There's been times with v2 I've wished it was one byte
I have never used either of those with a single-byte function
Which is the only case where it saves bytes over just ...)↔ or the appropriate lambda
A quick search shows I've used it like that at least once
Okay, aside from that digital root thing
11:29 AM
I feel like I've also seen it a fair few times in jelly answers
Modifiers taking dyads are rare enough that I think having them force dyadicity is fine - fold, outer product, pairwise if it was useful, and occasionally scan and nopop
I get where you're coming from trying to have everything cleanly defined but IMO it just doesn't work well with golfing
@emanresuA it's for saving modifier space
Overloading based on arity
/ given a monad is filter, given a dyad is reduce
Best way to save modifier space is to turn most of those into function overloads
It's the idea of how ~ in v2 does multiple things expanded to everything
@emanresuA I'll probably consider that when I get around to actually one day fully resuming development

« first day (1538 days earlier)      last day (11 days later) »