use binary search.

[jerome-laforge]

Hello,
Is it possible to use binary search for func (a *alphabet) Index(t rune) (int64, error)?

Thanks

[lithammer]

Hmm, benchmarks seems to indicate that this is slower. I guess maybe because of the small search space...?

           │   old.txt   │               new.txt               │
           │   sec/op    │   sec/op     vs base                │
UUID-8       2.965µ ± 3%   2.942µ ± 0%        ~ (p=0.085 n=10)
Encoding-8   2.428µ ± 2%   2.413µ ± 1%   -0.62% (p=0.002 n=10)
Decoding-8   368.4n ± 0%   448.5n ± 1%  +21.74% (p=0.000 n=10)
geomean      1.384µ        1.471µ        +6.28%

           │   old.txt    │                new.txt                │
           │     B/op     │     B/op      vs base                 │
UUID-8       2.588Ki ± 0%   2.588Ki ± 0%       ~ (p=1.000 n=10) ¹
Encoding-8   2.578Ki ± 0%   2.578Ki ± 0%       ~ (p=1.000 n=10)
Decoding-8     152.0 ± 0%     152.0 ± 0%       ~ (p=1.000 n=10) ¹
geomean       1020.7         1020.7       +0.00%
¹ all samples are equal

           │  old.txt   │               new.txt               │
           │ allocs/op  │ allocs/op   vs base                 │
UUID-8       98.00 ± 0%   98.00 ± 0%       ~ (p=1.000 n=10) ¹
Encoding-8   98.00 ± 1%   98.00 ± 0%       ~ (p=1.000 n=10)
Decoding-8   5.000 ± 0%   5.000 ± 0%       ~ (p=1.000 n=10) ¹
geomean      36.35        36.35       +0.00%
¹ all samples are equal

[jerome-laforge]

You are right.
Even with biggest random shortuuid list, I have slower decoding.
So we can reject this PR.