Sorting Isn't Simple–You're Just Seeing ASCII Order, Not Real Collation
7 days ago
- #Readyset
- #collation
- #ICU4X
- Alphabetization is more complex than it appears, especially in software engineering.
- Different languages and regions have unique sorting rules, making collation challenging.
- Readyset previously had limited collation support, relying on simple, sometimes incorrect methods.
- ICU4X was chosen over MySQL's collation due to its correctness, despite some MySQL incompatibilities.
- ICU4X initially lacked collation key support, which was contributed by Readyset's team.
- The new collation system in Readyset is more accurate and aligns with MySQL for most use cases.
- Readyset's contributions to ICU4X continued with bug fixes and additional features post-implementation.