« first day (1107 days earlier)      last day (368 days later) » 

12:50 AM
@Ginger also, there's a very good chance that if you're using the fractional byte encoder/decoder offline, you're already someone who would have java - the only people who use the offline interpreter are those who need it for either a) get requests (which crosshatch can be used for) or b) vyxal development testing. If you're in category b, there's also a chance you've installed scala for version 3 development, which requires java anyway
to me, it's an acceptable compromise
my estimations are that no java would be a problem for 0.01% of people. only about 1% of people use the offline interpreter, and then, only 1% of people would use fractional bytes offline (because you might as well just run sbcs, the fracbytes are for scoring purposes), and then, once you've gotten past that, you probably have 10% of people without java
1% * 1% * 10% = 1e-5
and you only need java if using the fracbyte stuff
the process is never called if you don't use the ! or = flags
 
Whatever we're doing with fracbyte stuff, we should probably add it to the h flag menu
 
oh dang yeah that's right
i forgor
turns out I didn't
 
 
2 hours later…
2:41 AM
@lyxal when you have java but not scala
 
@Seggan that's perfectly fine. I more meant people who also have done v3 work.
 
it be like that
"version" here means the algorithm + corpus version
not the actual encode/decode functions
CMQ: Where should a guide on how to use vyncode be stored? In the Vyxal repo or the Vyncode repo?
by guide, I mean explanation of what vyncode even is
 
3:53 AM
Vyxal repo could have a summary and link to the full explanation in the Vyncode repo
 
 
4 hours later…
8:07 AM
I'll be adding a encode/decode utility section before merging the pr
 
9:04 AM
The interesting thing about vyncode is that some combinations are shorter than others, even though they achieve the same thing
So that might open up a whole new facet of golfing
 
 
3 hours later…
12:01 PM
Lyxal deleted branch gh-readonly-queue/main/pr-1720-da42ac2850c4adad5a5c4a27d44afb7c7d7537c2 in Vyxal/Vyxal
Lyxal deleted branch v2-expand-into-fracbytes in Vyxal/Vyxal
 
Thus it begins
!!/prod
!!/production
 
@Ginger An error occurred while handling message 63474173!
 
!!/prod
 
@Ginger An error occurred while handling message 63474176!
 
wonderful
for once that's not my fault
watch it inadvertently be as a consequence of too much snowy conversations
hey wait it did open a PR
Vyncode is now live on vyxal.pythonanywhere.com
well
almost
it'll probably take a bit
huh, so turns out it didn't copy vyncode over
okay so it did copy it over
must be caching
yeah it was
23
A: 1, 2, Fizz, 4, Buzz

lyxalVyxal, j, 65 bits1, 18 17 14 12 11 8.125 bytes ₁ƛ₍₃₅kF½*∑∴ Try it Online! (link is to bitstring) I refuse to be beaten by Arn, Ash and Fig. I absolutely will not be beaten by any of those languages. Vyxal forever lads. And yes, that really is 8.125 bytes. We have a fractional byte encoding syst...

you know I had to :p
So I guess as a general rule of thumb: don't go changing every single vyxal answer to use vyncoder. If you find a post that you think "yeah that'd be cool with vyncoder" use it, but don't go doing all the posts. And don't go commenting "can be shorter now" on old scored posts either.
2
obviously you can use it for new answers as much as you want
but if someone isn't using it, vyncode is not a golfing suggestion
also, (this isn't a rule nor a suggestion, just a little trolling): it'd be funny if we didn't tell anyone that vyncode is a thing outside of answers and then let people in TNB be like "hey wait since when is vyxal fracbyte"
no pressure to adhere to that, and it's completely fine to mention it in TNB
I just think it'd be funny
personal opinion
I cannot stress enough how much that's just something that'd be funny if we did, not something we necessarily do
 
12:26 PM
Using vyncode for things that are already in the corpus falls into the category of making a language for a specific challenge
Which kind of feels like cheating
 
@AndrovT So then what about when more programs get added to the corpus? Are those answers cheating if they have a rescore?
 
Imo answers should only use vyncode older than the answer
 
I guess when it's used will be like flags
some view them as cheating, others don't
 
Otherwise you could make a language where whenever a challenge gets posted you update it such that the solution is only one or two bytes
 
12:33 PM
@TheThonnu already outgolfed :p
 
1:08 PM
@lyxal what is vyncode?
 
A fractional byte compression system
 
think of it like Huffman coding, but better
 
Basically, sbcs in, bitstring out
And that bit string is how you score the submission
@AndrovT and then what if I change a character in the answer I'm rescoring?
just as I said
new meta golfing game :p
 
1:26 PM
nice
also how do i input a variable, use it, then use it later?
 
so you want to place the top of the stack into a variable?
→name # Set
←name # Get
 
oh yes
 
Just note that the register is available (£ = set, ¥ = get), which saves a byte over named variables
 
^
 
ok
thx
 
1:36 PM
hey welcome back btw, been a while since I've seen you around these parts :p
 
:+1:
 
both of you :p
 
@lyxal I'm not sure. The proper way to do it is to count each version of vyncode as a separate language. But that's probably not necessary.
 
@AndrovT it kind of already is
 
@lyxal I was saying :+1: to Alan
 
1:40 PM
@AndrovT technically speaking, it's a flag variant
@TheThonnu and I was saying it's nice to see you here too :p
 
@lyxal :+1:
@lyxal I think the header should at least mention Vyncode in bitstring answers
 
it does
 
Like instead of "Vyxal, x bits, y bytes" it should say "Vyxal <sup>Vyncode</sup>, ..."
 
the bytes hyperlink is to vyncode
and the superscript 1 is v1 of vyncode
 
I think it should be next to Vyxal
But it doesn't really matter
 
1:43 PM
I think the fact that there's even a bit count when it's widely known that vyxal is SBCS by default should suggest there's an alternate scoring happening :p
 
Sure
@lyxal in the last release you said that 2.20.1 would be when #1683 is merged
But it's been merged already
3 weeks ago in fact
 
Well jokes on you next release is 2.21 now I decided
 
2.20.0 -> 2.21.0 ?
 
@TheThonnu Also dang 3 weeks ago
Times goes way too quick
 
ikr
It doesn't feel like 3 weeks
 
1:48 PM
@TheThonnu vyncode is enough to push it to 2.21 imo
If it wasn't for that I
Would have made it 20.1
And I probably wouldn't have hit the enter key too early
 
!!/amilyxal
 
@TheThonnu You are not lyxal.
 
Good
 
!!/amilyxal
 
@AlanBagel You are not lyxal.
 
1:57 PM
good
 
!!/cookie
 
@TheThonnu No.
 
What?
It says no?
!!/hug
 
@TheThonnu (つ ͡° ͜ʖ ͡°)つ
 
!!/amilyxal
 
2:00 PM
@lyxal You are lyxal.
 
!!/cookie
 
@lyxal Here you go: 🍪
 
skill issue
 
There's a 25% chance if you're not admin
!!/cookie
 
@TheThonnu Here you go: 🍪
 
2:02 PM
There we go
 
 
2 hours later…
4:03 PM
Writing code at vyxal.pythonanywhere.com with the live bitcount enabled is really slow to the point of being nearly useless. Is it not possible to run the encode asynchronously somehow?
 
!!/info
 
Also why does it take the web version the same amount of time to compress a 50 byte program as it takes the python version to compress every Vyxal program ever written???
 
4:19 PM
@AndrovT It is JavaScript
@AndrovT It's possible the web version's really inefficient?
Can we incrementally encode stuff in blocks?
 
@VyxalBot bruh
 
 
5 hours later…
9:15 PM
!!/info
 
 
2 hours later…
11:06 PM
I was doing some thinking about intra-language competition and I've come up with an idea that could be cool. Currently, the only metric for scoring vyxal answers is byte count, and once an optimal byte count is found, there's no reason for anyone else to answer. What if there was a metric that wasn't minimisable? My idea is something like a style points metric (I'm referring to it as the Gaming Score for now) where highest Gaming Score wins per question
Things like lesser used element combos would score better and things like variables, comments and 1 byte answers would give penalties
Think of it like a secondary, unofficial leaderboard
Answers would still have to comply with site rules that they have a serious golfing attempt applied
And programs longer than a certain length would start penalising the Gaming Score too
Because while you might get more style combos in a longer answer, a horribly bloated program isn't stylish
It'd be an opt-in thing
Marked somehow in the answer body
No clue about any finer rules, just an idea
And Gaming Score wouldn't be an encoding system, rather a secondary "in-the-know" competition
It could encourage more people to use vyxal on answers where there's already a 6 byte answer
But that's just an idea at this point
 

« first day (1107 days earlier)      last day (368 days later) »