fix(docs): correct multiple documentation bugs

- Add missing RLE magic number in algorithm docs and C++ API
- Remove duplicate Streaming API link in architecture page
- Fix table formatting error in zh/algorithms guide
- Add missing State Machine link in EN sidebar
- Update BibTeX year from 2024 to 2026
- Unify file size units to KiB (IEC standard)

Author: LessUp

docs/.vitepress/config.ts

@@ -28,6 +28,7 @@ const sharedSidebar = {
       items: [
         { text: 'Algorithm Academy', link: '/en/academy/' },
         { text: 'Huffman Coding', link: '/en/academy/huffman' },
+        { text: 'State Machine Design', link: '/en/academy/state-machine' },
       ],
     },
     {

docs/en/algorithms/range.md

@@ -161,9 +161,9 @@ let decoded = rangecoder::decode(&encoded, &cum_freq, data.len())?;
 
 ::: warning Performance Issue with Large Files
 
-The current Range Coder implementation has a **known decode performance issue** for files larger than **500 KB**. The decode operation may become significantly slower or appear to hang.
+The current Range Coder implementation has a **known decode performance issue** for files larger than **500 KiB**. The decode operation may become significantly slower or appear to hang.
 
-**Workaround**: For testing purposes, use files smaller than 100 KB. This is reflected in the CI pipeline which uses 100 KB test files for Range Coder verification.
+**Workaround**: For testing purposes, use files smaller than 100 KiB. This is reflected in the CI pipeline which uses 100 KiB test files for Range Coder verification.
 
 **Status**: This is a known issue that is documented for future improvement. The encode operation works correctly for all file sizes.
 

docs/en/algorithms/rle.md

@@ -92,6 +92,7 @@ pub fn encode(data: &[u8]) -> Vec<u8> {
 
 | Field | Size | Description |
 |-------|------|-------------|
+| Magic | 4 bytes | `RLE\x00` (0x52 0x4C 0x45 0x00) |
 | Count | 4 bytes | Little-endian unsigned int (run length) |
 | Value | 1 byte | The repeated byte value |
 

docs/en/api/cpp.md

@@ -93,7 +93,10 @@ g++ -std=c++17 -O2 -Wall -Wextra -o <binary> main.cpp
 
 ### File Format
 
-Repeated `(count: uint32 LE, value: byte)` pairs.
+| Offset | Size | Field |
+|--------|------|-------|
+| 0 | 4B | Magic: `RLE\x00` |
+| 4+ | Variable | Repeated `(count: uint32 LE, value: byte)` pairs |
 
 ---
 

docs/en/architecture/index.md

@@ -180,6 +180,5 @@ Hidden complexity:
 
 ## Further Reading
 
-- [Streaming API](/en/api/streaming) - 5-state FSM details
-- [Streaming API](/en/api/streaming) - Complete API documentation
+- [Streaming API](/en/api/streaming) - 5-state FSM details and complete API documentation
 - [Cross-Language Testing](/en/testing/cross-language) - Conformance verification

docs/en/guide/algorithms.md

@@ -158,7 +158,7 @@ cargo run --bin rangecoder -- decode output.rcnc restored.bin
 :::
 
 ::: warning Performance Note
-The Range Coder decoder has a known performance issue for files >500KB. Use smaller test files for cross-language verification.
+The Range Coder decoder has a known performance issue for files >500 KiB. Use smaller test files for cross-language verification.
 :::
 
 ---

docs/en/reference/bibliography.md

@@ -107,10 +107,10 @@ This document collects classic papers, authoritative textbooks, and related open
 If you reference CompressKit in academic work, we suggest using the following format:
 
 ```bibtex
-@misc{compresskit2024,
+@misc{compresskit2026,
   author = {CompressKit Team},
   title = {CompressKit: Cross-Language Lossless Compression Algorithms},
-  year = {2024},
+  year = {2026},
   publisher = {GitHub},
   url = {https://github.com/LessUp/compress-kit}
 }

docs/en/testing/cross-language.md

@@ -89,7 +89,7 @@ Detailed format specifications are available in the specs directory:
 
 | Algorithm | Limitation | Workaround |
 |-----------|-----------|------------|
-| Range Coder | Decode performance degrades >500KB | Use smaller chunks |
+| Range Coder | Decode performance degrades >500 KiB | Use smaller chunks |
 
 ## Reporting Issues
 

docs/zh/algorithms/range.md

@@ -162,9 +162,9 @@ let decoded = rangecoder::decode(&encoded, &cum_freq, data.len())?;
 
 ::: warning 大文件性能问题
 
-当前 Range Coder 实现存在一个**已知的解码性能问题**：当文件大于 **500 KB** 时，解码操作可能会变得非常缓慢或出现卡顿。
+当前 Range Coder 实现存在一个**已知的解码性能问题**：当文件大于 **500 KiB** 时，解码操作可能会变得非常缓慢或出现卡顿。
 
-**临时解决方案**：测试时请使用小于 100 KB 的文件。CI 管道中已使用 100 KB 测试文件进行 Range Coder 验证。
+**临时解决方案**：测试时请使用小于 100 KiB 的文件。CI 管道中已使用 100 KiB 测试文件进行 Range Coder 验证。
 
 **状态**：这是一个已知问题，已记录以便未来改进。编码操作对所有文件大小均正常工作。
 

docs/zh/algorithms/rle.md

@@ -92,6 +92,7 @@ pub fn encode(data: &[u8]) -> Vec<u8> {
 
 | 字段 | 大小 | 描述 |
 |------|------|------|
+| Magic | 4 字节 | `RLE\x00` (0x52 0x4C 0x45 0x00) |
 | 计数 | 4 字节 | 小端序无符号整数（行程长度） |
 | 值 | 1 字节 | 重复的字节值 |