05AB1E 05AB1E Corpus

05AB1E Corpus

Open ghost opened this issue 3 years ago • 1 comments

Similar to this issue opened in the Jelly repo and this issue opened in the Husk repo, I decided to run Lynn's method on 05AB1E answers (with a few modifications).

SEDE query used to get all 05AB1E answers on CGCC. Code used to analyze the data. Final corpus.

 Digraphs: Trigraphs: Quadgraphs:
   261 vy     57 34ç     38 qqqq 
   113 })     40 qqq     32 "D34 
   100 Dg     32 "D3     32 D34ç 
    96 2ô     32 D34     30 34çý 
    91 .•     31  =      25 "34ç 
    85 .V     30 4çý     23 AAAA 
    84 Qi     29 AAA     19 ) =  
    81 ðý     25 "34     18 34çì 
    79 sè     22 γ€g     16 4çìD 
    76 2ä     20 ∞.Δ     12 0"D3 
    74 €g     19 vyN     10 çìD« 
    73 Q}     19 ˆ}¯      9 {γ€g 
    71 .Δ     19 ) =      9 ‚•„í 
    71 .ø     18 4çì      9 •„í† 
    67 12     18 øvy      9 „í†ì 
    66 34     17 vy`      9 í†ìˆ 
    61 '-     17 è.V      9 †ìˆÈ 
    59 '.     16 vDy      9 ìˆÈŒ 
    57 4ç     16 çìD      9 ˆÈŒš 
    56 Nè     15 }})      9 ÈŒšï 
    55 }}     14 nOt      9 Œšï¿ 
    55  ÿ     14 XXX      9 šï¿Ÿ 
    54 }J     14 ƶ0K      9 ï¿Ÿ¯ 
    52 .c     13 Sðý      9 ¿Ÿ¯¥ 
    51 0K     13 '-×      9 Ÿ¯¥Š 
    50 SO     13 ÂQ}      8 DÙQ} 
    48 D3     13 2ä`      8 ºS₂+ 
    47 ,      13 #vy      8 ƒËŠˆ 
    47 εD     12 yåi      8 ʒÂQ} 
    47 3ô     12 DÙQ      8 €µ‚•

Remarks:

Digraphs:

I think pretty much everything here should be added as a one-byte builtin. If these were all 1 byte builtins, there will be a dramatic improvement in 05AB1E's golfiness.

vy is the most common digraph here. So we need an enumerated map command that auto-pushes the current iterated variable.
For }), a one-byte built-in should be made that 1) closes the current if statement/loop and 2) Wraps entire stack into a list.
.• and .V are very common in this list. 1-byte alternatives should definitely be made for those.
Apparently, Dg (Duplicate + length) is pretty common, so a one-byte built-in should be made for this.
05AB1E doesn't have a built-in for 2ô (split into chunks of length 2). So we also need a builtin for this.
Qi: We probably need to make an if that compares equality in advance.
ðý: Command that joins by spaces.
sè: Indexing command with swapped arguments.
2ä: Apparently, splitting something into two halves is quite common as well.
€g: Map every item by length.
.Δ: Make it a one-byte built-in, the other option is explained below.
.ø: We need a one-byte command for surround.
'. and '-: Apparently, these get used pretty often.
4ç: Guess it's caused by the lack of a quote (34ç) builtin
Nè: Index by iteration variable - sounds good for a one byte builtin
ÿ: Implicit prepending space can be useful in string interpolation.
}J: Looks like people want to join after looping

Trigraphs:

34ç is the most common trigraph. So this should just be a one or two byte built-in.
∞.Δ: A command that does .Δ over [1 .. ∞]. I think making .Δ a one-byte built-in is better than making this a two-byte built-in, since .Δ appears much more frequently.

Quadgraphs:

I think all of the quadgraphs here are either 1) overly specific; 2) covered by Digraphs/Trigraphs; or 3) inside a string.

Apr 05 '21 09:04 ghost

A list of most common 1-byte builtins:

Apr 13 '23 08:04 jfioasd

05AB1E 05AB1E copied to clipboard

05AB1E Corpus

Remarks:

Digraphs:

Trigraphs:

Quadgraphs:

05AB1E
05AB1E copied to clipboard