Idea 💡

[rotund birch]

Hello, I love this game, and I have an idea, sorry for my english.
I love collect any type of things, and in this game, I think it would be cool to have like a "tree" of all the item, or a collection board, something like that with all the items that we've discovered and like a "?" on the one's that we didn't discovered yet.
I think that some people surely already asked for something like that but I wanted to ask this, it's just an idea, to share my point of view!! But such a great game !!

[bitter sedge]

EVERY item you haven't discovered?

that's probably more than 20 million and thats only ones which people have discovered

[naive vapor]

you're a bit short there. The number of unique combinations of the letters a-z, numbers 0-9, and a space, with a limit of 30 characters, is 1.11^47.

That doesn't count that lower and upper case are different. Then there are all the emoji, symbols, and non-english characters. And that you can create strings longer than 30 characters, even if you can't use them again.

So while its not technically infinite, it's large enough to effectively be infinite.

"which ones we haven't found" is not actually possible to list out.

[bitter sedge]

i was talking about ones already discovered

although the ones we havent found is probs closer to that

also the limit is actually 20 tokens, 30 chars is just when the element becomes dead

[cobalt hamlet]

based on a figure of 32k tokens in llama 2 dictionary

[prisma tundra]

technically you have a bunch of exceptions like trimmed white Spaces, trimmed elements that contain +

but yeah

[cobalt hamlet]

if anyone wants to account for every edge case to find a true upper bound be my guest

[tidal stump]

Good idea with ib/builtin collections tho (maybe mass spawning of ghost elms from sandwich utils?)

@tidal stump Userscript 'implementation'?

[ocean monolith]

while impossible that would be really cool

[tidal stump]

h

[foggy flame]

there is way more

all unicode characters, uppercasing, and way more

[cobalt hamlet]

what are you talking about

[foggy flame]

there is over 80000 unicode characters

[cobalt hamlet]

i see

i didn't remember the number so i didn't know if that might've already been included in the figure or not

wait no

i'm pretty sure many unicode characters will be split into multiple tokens

[cobalt hamlet]

32k was sourced from llama papers themselves and unicode characters will be regularly parsed into multiple tokens (ie overlap is possible) so i will stick with this figure

[tidal stump]

h

[tidal stump]

b @tidal stump

[red pawn]

Eye blinding