Discussion for: Add Taiwan Traditional Chinese Collation Support for utf8mb4

[rkajiyama]

Abstract

MySQL currently provides a generic Chinese Unicode collation for utf8mb4, such as utf8mb4_zh_0900_as_cs, but it does not appear to expose an explicit Taiwan Traditional Chinese collation (臺灣繁體中文字排序).

For applications serving users in Taiwan, a generic Chinese collation may not always match expected Traditional Chinese ordering, especially where Taiwan-specific sorting, Zhuyin / Bopomofo order, or stroke-based order is expected.

This discussion proposes adding explicit Taiwan Traditional Chinese collation support for utf8mb4. Possible candidates include:

• **utf8mb4_zh_tw_0900_as_cs** -- Taiwan-specific Traditional Chinese, Stroke-based
• **utf8mb4_zh_zhuyin_0900_as_cs** -- Zhuyin / Bopomofo ordering
• **utf8mb4_zh_unihan_0900_as_cs** -- Radical-Stroke Sort Order, based on CJK Unified Ideographs (CJK統合漢字)

This proposal is related to the broader Unicode 17 support discussion #72

However, this discussion is scoped to the currently supported Unicode 9.0 / UCA 9.0 collation family. If Unicode 17 support is accepted and scheduled first, the same idea could instead be implemented for a future utf8mb4_zh_*_1700_* collation family.

Implementation details can be discussed separately in Issues. The goal of this discussion is to collect requirements and community feedback.

[rkajiyama]

Questions

1. Is an explicit Taiwan Traditional Chinese utf8mb4 collation needed in MySQL?

2. If yes, are the following candidates appropriate?

   • utf8mb4_zh_tw_0900_as_cs
   • utf8mb4_zh_zhuyin_0900_as_cs
   • utf8mb4_zh_unihan_0900_as_cs

3. Are any of these candidates unnecessary?

4. Are there other Taiwan Traditional Chinese sorting modes that should be considered?

5. Should the naming use zh_tw, zh_hant_tw, zhuyin, stroke, or another convention?

6. Should this be implemented for the current Unicode 9.0 / UCA 9.0 collation family, or should it wait for a future Unicode 17 collation family if Unicode 17 support is accepted?

[IvanTu2729]

Hi,

I concur what Kajiyama-san proposed, in Taiwan, most people use Zhuyin symbols for typing, and stroke number for sequencing/ordering. But Chinese don't have notion of capital, except numbering, However, we do have idea of as (Accent-Sensitive), I think "as" is required. So, it will be very helpful, if the following collations are there.
utf8mb4_zh_tw_0900_as
utf8mb4_zh_zhuyin_0900_as
utf8mb4_zh_stroke_0900_as