docs: keep usage guidance version-agnostic

Also add Traditional Chinese README

Author: frankslin

README.md

@@ -1,55 +1,22 @@
 # opencc-data [![](https://badge.fury.io/js/opencc-data.svg)](https://www.npmjs.com/package/opencc-data) [![JSDelivr badge](https://data.jsdelivr.com/v1/package/npm/opencc-data/badge)](https://www.jsdelivr.com/package/npm/opencc-data)
 
-A collection of word lists for Simplified and Traditional Chinese conversions from the [OpenCC](https://github.com/BYVoid/OpenCC) project.
+[繁體中文](README.zh-TW.md)
+
+A collection of dictionary data, configs, and test data for Simplified and Traditional Chinese conversions from the [OpenCC](https://github.com/BYVoid/OpenCC) project.
 
 ## Compatibility
 
-This project is primarily maintained for use with [opencc-js](https://github.com/nk2028/opencc-js).
+This package is intended for [opencc-js](https://github.com/nk2028/opencc-js) and other OpenCC-compatible implementations that consume OpenCC dictionary and config data.
 
 - **Strict Version Matching**: Compatibility is only guaranteed when the version of `opencc-data` strictly matches the version of the consumer package.
 - **Breaking Changes**: We do not guarantee compatibility between different versions of `opencc-data`. Structure or file names may change to align with upstream OpenCC updates.
 
 ## Data Sync Policy
 
-This package syncs upstream OpenCC dictionary `.txt` files, config `.json` files, and `test/testcases/testcases.json`. It also generates and checks in derived dictionaries during data sync: `TSCharactersExt.txt` from `TSCharacters.txt`, plus the reverse variant dictionaries `HKVariantsRev.txt`, `TWVariantsRev.txt`, and `JPVariantsRev.txt` from their corresponding variant dictionaries.
+This package syncs dictionary `.txt` files and config `.json` files from OpenCC's generated resource zip, plus `test/testcases/testcases.json` from the upstream repository. Each sync keeps the upstream resource manifest as an internal baseline so future updates can detect substantive data changes by resource hash.
 
 ## Usage
 
-The data files prioritize canonical data from the OpenCC project. The following configurations are strictly consistent with the logic in OpenCC's `.json` configuration files:
-
-- **Outer Array (Stages)**: Represents the conversion-chain stages.
-- **Inner Array (Groups)**: Represents a dictionary group used within one stage; dictionaries earlier in the group have higher priority when their entries overlap (merged relationship).
-
-**From Chinese variants to OpenCC standard:**
-
-```json
-{
-  "cn": [["STPhrases", "STCharacters"]],
-  "hk": [["HKVariantsRevPhrases", "HKVariantsRev"]],
-  "tw": [["TWVariantsRevPhrases", "TWVariantsRev"]],
-  "twp": [["TWPhrasesRev", "TWVariantsRevPhrases", "TWVariantsRev"]],
-  "jp": [["JPShinjitaiPhrases", "JPShinjitaiCharacters", "JPVariantsRev"]]
-}
-```
-
-**From OpenCC standard to Chinese variants:**
-
-```json
-{
-  "cn": [["TSPhrases", "TSCharactersExt", "TSCharacters"]],
-  "hk": [["HKVariantsPhrases", "HKVariants"]],
-  "tw": [["TWVariantsPhrases", "TWVariants"]],
-  "twp": [["TWPhrases"], ["TWVariantsPhrases", "TWVariants"]],
-  "jp": [["JPVariants"]]
-}
-```
-
-`TSCharactersExt` contains rare inferred simplified forms that may render as tofu (missing glyph boxes) in common fonts. It should generally be omitted unless those extended mappings are explicitly required.
-
-**Explanation of the Chinese variants above:**
+Use the config files shipped in `data/config/` as the source of truth for dictionary order and conversion-chain behavior. Config contents can change between OpenCC versions, so consumers should load the matching config file for the package version they depend on instead of hard-coding dictionary lists from this README.
 
-- `cn`: Simplified Chinese (Mainland China)
-- `tw`: Traditional Chinese (Taiwan)
-- `twp`: Traditional Chinese (Taiwan, with phrase conversion)
-- `hk`: Traditional Chinese (Hong Kong)
-- `jp`: Japanese Shinjitai
+Dictionary text files are shipped in `data/`. Config files reference those dictionaries by file name and preserve OpenCC's stage/group ordering semantics.

README.zh-TW.md

@@ -0,0 +1,22 @@
+# opencc-data [![](https://badge.fury.io/js/opencc-data.svg)](https://www.npmjs.com/package/opencc-data) [![JSDelivr badge](https://data.jsdelivr.com/v1/package/npm/opencc-data/badge)](https://www.jsdelivr.com/package/npm/opencc-data)
+
+[English](README.md)
+
+本套件收錄來自 [OpenCC](https://github.com/BYVoid/OpenCC) 專案的簡繁轉換詞典資料、配置與測試集。
+
+## 相容性
+
+本套件適用於 [opencc-js](https://github.com/nk2028/opencc-js)，也可供其他相容 OpenCC 詞典與設定檔格式的實作使用。
+
+- **嚴格版本匹配**：只有在 `opencc-data` 與使用端套件版本完全一致時，才保證相容性。
+- **破壞性變更**：不同版本的 `opencc-data` 之間不保證相容。資料結構或檔名可能會隨上游 OpenCC 更新而調整。
+
+## 資料同步政策
+
+本套件會從 OpenCC 產生的 resource zip 同步詞典 `.txt` 檔與設定 `.json` 檔，並從上游 repository 同步 `test/testcases/testcases.json`。每次同步都會保留上游 resource manifest 作為內部基準，之後可透過 resource hash 判斷是否有實質資料變更。
+
+## 使用方式
+
+請以套件內 `data/config/` 提供的設定檔作為詞典順序與轉換鏈行為的依據。設定內容可能隨 OpenCC 版本改變，因此使用端應載入其依賴版本對應的設定檔，而不是從 README 固定寫死詞典列表。
+
+詞典文字檔位於 `data/`。設定檔會以檔名引用這些詞典，並保留 OpenCC 的 stage/group 排序語意。