« first day (767 days earlier)      last day (1009 days later) » 

00:58
@VyxalBot you good bro?
@lyxal I am doing poorly.
Yes
I can tell
You haven't reported any PRs not PR reviews
Also everyone needs to sign up for the ping list again lol
Because the js migration cleared the list lol
!!/ping me
@user I have put you on the ping list.
01:26
!!/ping me
@emanresuA I have put you on the ping list.
01:57
@lyxal according to the github org settings, it's because hyper's server's certificates need updated
02:09
which is indeed the case, you can see that the certirficate is expired if you go to his site
@hyper-neutrino your certificates are expired lol
02:27
!!/ping me
@Seggan I have put you on the ping list.
@lyxal sounds like myxal
02:48
Can't believe we only have 5 stars
 
2 hours later…
04:21
!!/ping me
@PyGamer0 I have put you on the ping list.
 
3 hours later…
07:45
If I could get some more opinions on the core functions list (github.com/Vyxal/Vyxal/blob/v3-planning-documents/…) that'd be great
!!/hyperping
@lyxal @lyxal @Steffan @emanresuA @user @hyper-neutrino @Seggan @PyGamer0 ^
(cc @UnrelatedString)
!!/ping me
@UnrelatedString I have put you on the ping list.
@lyxal steal quicks from flax :P
07:47
That doesn't fill the full list though :p
maybe add ngn/k's monadic &?
what does that do
anyhow i'm still reading through and everything else so far looks good but dedicated hex conversion builtins seem questionable
so replicate indices pretty much?
07:51
i can see it being useful but not that useful
@UnrelatedString any suggestions you'd add that aren't currently part of vyxal?
good question
i just finished reading through and nothing's obviously missing
wait actually indexing is obviously missing lmao
but in terms of non-silly-mistake omissions i'll have to think on that longer
maybe multiple function -> infinite list elements? like one to iterate on the single previous value and one to do something more like a recurrence relation
maybe the latter one could take a list of starting inputs but then the function it takes sees those as dumped into the stack
though that could make it harder to hypothetically pass additional parameters in
08:38
I have been pinged
@lyxal TBH I think having both structures and complicated modifiers takes up codepage space and causes redundancy
But the moment you start removing structures, you start to lose the easiness of vyxal
Besides, it won't take up as much codepage as it does now
There'll only be one closing character
Instead of several
And besides, using just modifiers for things like if statements only makes golf sense if you have things like chain rules or precedence
I guess... but it's hard to combine the two
@lyxal Jelly's modifiers take a single atom :P
How so?
@emanresuA but that atom can have multiple atoms because chain rules
+3 could be considered an atom
08:51
*link
except that can't for any purposes in jelly
@lyxal idea: break and continue statements for loops, the while and for loops can also have an else branch...
it will be ab;e to in perhaps :)
@UnrelatedString but my point was more that things are automatically grouped as a single unit based on arity
@UnrelatedString perhaps
they aren't but yeah the point is they could be
husk is probably a better example
08:54
What I'm saying is that structures are better for when you have more things than a modifier can group at once
also structures are just important cultural heritage
Say you had an if statement with 7 elements inside. A structure just has those 7 elements plus 2 brackets. Modifiers would need 1 modifier for the if statement, and a weird combination of at least 2 more modifiers to group the 7 because there's only options for a maximum of a group of 4 at once
Structure = 9 bytes, Modifiers = 10+ bytes depending on how you do it
and in cases where if or while comes up especially you're very likely to need multiple things in there
08:57
Exactly
Plus you can make it so that multiple structures can be closed at once
I guess ¯\_(ツ)_/¯
At the same time, I think modifiers are good because structures predominantly operate on a single value.
8 mins ago, by PyGamer0
@lyxal idea: break and continue statements for loops, the while and for loops can also have an else branch...
[Discord Relay]
Razetime#5187: hm whats the discussion
[Discord Relay]
Razetime#5187: husk is a good example an also a bad example of automatic grouping
[Discord Relay]
Razetime#5187: sometimes you have a great program and no type inference agreement which kills everything
Razetime#5187: often adding many unnecessary bytes
yeah husk's inference can be remarkably unhelpful sometimes lmao
@lyxal oh yeah wait what's the plan for the unifying closing character
in terms of actual codepage allocation
09:43
@UnrelatedString }
It was originally going to be ; but Seggan showed me why that was a bad idea
Curly closing bracket stands out better
@emanresuA well why can't there be both?
09:58
yeah a semicolon doesn't have the same kind of visual footprint as a closing bracket
the problem with choosing a single closing bracket to close brackets is you then have the other two doing something else but it shouldn't fuck with intuition too hard
10:14
The other two are modifiers now
Or will be
Because counting to three is hard lol
@lyxal why?
[code;code]
It looks like the ; could be separating two branches when in reality that's an if statement followed by a modifer
[code}code]
That makes it clearer that a block has been closed
Because the } disrupts the perception of [] being a full structure
It makes you think for a bit
And then see what's really happening
hmm true
well
how about using ; to close structures
and use letters like W,F,? to start structures?
05ab1e does that and it obscures that it's a structure
That's one of the things I've never liked about 05ab1e
Looks too letterish
Brackets are better aesthetically
hmmmm
10:27
yeah 05ab1e structures are super hard to parse visually if you don't know the language already
That's exactly right
And that's the reason I made two esolangs that both use brackets for structural stuff
Hell, that's why letters push themselves in Keg
Because letters are, at best, commands, not structures
[Discord Relay]
Razetime#5187: this same problem is in stax which goes `{..letter`
[Discord Relay]
Razetime#5187: but i don't mind it much since the debugger highlights blocks
and don't those letters also implicitly take everything afterwards if they don't have a block to close
what about <letter>{...}
well that's just wasteful :P
10:40
Btw, for v3, I think some functions/overloads make a lot of sense and should stay the way they are.
e.g. T, truthy indices / triple
@lyxal is sort lambda even that useful?
Or A, All / is-vowel
@Seggan Occasionally
whats #85 for
maybe also a spaceship element?
i.e. -1 if less, 0 if equal, 1 if greater
actually not a bad idea
[Discord Relay]
Razetime#5187: yeah unrelated string, they do
Razetime#5187: but those are also highlighted correctly
10:50
hard to say how useful it would be but worth ocnsidering
[Discord Relay]
UnrelatedString#9744: i would certainly hope so 😛
Razetime#5187: oh no they're on discord
[Discord Relay]
Razetime#5187: help
Razetime#5187: phew
[Discord Relay]
Seggan#8111: *evil laugh*
[Discord Relay]
Razetime#5187: *dies*
11:08
can't believe you've done this seggan
done what
look what you caused by evil laughing!
oh no!
anyhow,
thank you copilot
that's definitely what I want to do
Well I mean I suppose I did screw up the function name
because ż is one-length, not
aw beep boop doesn't autolabel PRs now because webhooks no

« first day (767 days earlier)      last day (1009 days later) »