Remove benchmarks, ziglyph references

`ziglyph` is no longer maintained and basically abandoned, there's no
need to keep the comparison between them active going forward.

1 files changed, 0 insertions, 81 deletions

diff --git a/README.md b/README.md
index bfa8d5e..1d3899c 100644
--- a/README.md
+++ b/README.md
@@ -519,84 +519,3 @@ test "Scripts" {
     try expect(scripts.script('צ') == .Hebrew);
 }
 ```
-
-## Relation to Ziglyph
-
-zg is a total re-write of some of the components of Ziglyph. The idea was to
-reduce binary size and improve performance. These goals were achieved by using
-trie-like data structures (inspired by [Ghostty's implementation](https://mitchellh.com/writing/ghostty-devlog-006))
-instead of generated functions. Where Ziglyph uses a function call, zg uses an
-array lookup, which is quite faster. In addition, all these data structures in
-zg are loaded at runtime from compressed versions in the binary. This allows
-for smaller binary sizes at the expense of increased memory
-footprint at runtime.
-
-Benchmarks demonstrate the above stated goals have been met:
-
-```plain
-Binary sizes =======
-
-172K ziglyph_case
-109K zg_case
-
-299K ziglyph_caseless
-175K zg_caseless
-
-91K ziglyph_codepoint
-91K zg_codepoint
-
-108K ziglyph_grapheme
-109K zg_grapheme
-
-208K ziglyph_normalizer
-175K zg_normalize
-
-124K ziglyph_width
-109K zg_width
-
-Benchmarks ==========
-
-Ziglyph toUpperStr/toLowerStr: result: 7756580, took: 74
-Ziglyph isUpperStr/isLowerStr: result: 110959, took: 17
-zg toUpperStr/toLowerStr: result: 7756580, took: 58
-zg isUpperStr/isLowerStr: result: 110959, took: 11
-
-Ziglyph Normalizer.eqlCaseless: result: 626, took: 479
-zg CaseFolding.canonCaselessMatch: result: 626, took: 296
-zg CaseFolding.compatCaselessMatch: result: 626, took: 604
-
-Ziglyph CodePointIterator: result: 3691806, took: 2.5
-zg code_point.Iterator: result: 3691806, took: 3.3
-
-Ziglyph GraphemeIterator: result: 3691806, took: 78
-zg Graphemes.Iterator: result: 3691806, took: 31
-
-Ziglyph Normalizer.nfkc: result: 3856654, took: 411
-zg Normalize.nfkc: result: 3856654, took: 208
-
-Ziglyph Normalizer.nfc: result: 3878290, took: 56
-zg Normalize.nfc: result: 3878290, took: 31
-
-Ziglyph Normalizer.nfkd: result: 3928890, took: 163
-zg Normalize.nfkd: result: 3928890, took: 101
-
-Ziglyph Normalizer.nfd: result: 3950526, took: 160
-zg Normalize.nfd: result: 3950526, took: 101
-
-Ziglyph Normalizer.eql: result: 626, took: 321
-Zg Normalize.eql: result: 626, took: 60
-
-Ziglyph display_width.strWidth: result: 3700914, took: 89
-zg DisplayWidth.strWidth: result: 3700914, took: 46
-```
-
-These results were obtained on a MacBook Pro (2021) with M1 Pro and 16 GiB of RAM.
-
-In contrast to Ziglyph, zg does not have:
-
-- Word segmentation
-- Sentence segmentation
-- Collation
-
-It's possible that any missing functionality will be added in future versions,
-but only if enough demand is present in the community.