docs: replace the all-done changelog with a forward roadmap (EN+CN)

The Roadmap section was a list of shipped checkboxes. Split it into a
short 'Shipped' summary and a real 'Planned' list (PR reading mode,
whole-project chat, shareable export, wider annotation languages,
persistent project library, jump-to-definition).

Author: he-yufeng

README.md

@@ -206,21 +206,22 @@ CodeABC/
 
 ## Roadmap
 
-- [x] Project overview generation
-- [x] Hover annotations (Python priority)
-- [x] Terminology dictionary (hover keywords for definitions)
-- [x] Natural language editing ("change the stock from Maotai to BYD")
-- [x] Q&A mode (select code and ask questions)
-- [x] Multi-language UI (English interface)
-- [x] Test-coverage map (which files have tests; untested core files ranked by risk)
-- [x] Git-history insight (change hotspots, co-change coupling, and code ownership / knowledge silos)
-- [x] Tech-debt map (collect the codebase's own TODO/FIXME/HACK/XXX markers, ranked by file)
-- [x] Env-var surface (the environment variables a project reads, split into required vs optional)
-- [x] Entry-point detection (where execution starts: `__main__` scripts, declared console commands, conventional entry files)
-- [x] Logic-complexity ranking (which Python files carry the most tangled branching, from the AST — a different axis than centrality or churn)
-- [x] One-command launcher (`run.py` / `start.bat` — builds, serves, and opens the app)
-- [x] Single-process serving (the backend serves the built UI; one URL, no separate dev server)
-- [x] Native desktop app (Tauri — `npm run tauri:build` wraps the same UI in a native window)
+### Shipped
+
+The reading experience is already complete end to end: a plain-language project manual, hover annotations, a terminology dictionary, natural-language editing, and snippet Q&A, all in a bilingual UI you launch with one command (or as a native Tauri desktop app). On top of that sit a dozen deterministic, no-API-key maps — reading map, core-module ranking, test coverage, git-history hotspots and ownership, tech-debt markers, env-var surface, entry points, external integrations, silent-failure spots, under-documented files, and logic complexity.
+
+### Planned
+
+These are the directions I want to take next, roughly in priority order:
+
+- **Pull-request reading mode** — paste a PR link and get the diff explained in plain language: what changed, why it might matter, and which files to look at first. This is the most-requested extension and a natural fit for the existing annotation engine.
+- **Whole-project chat** — Q&A today is grounded in one file or snippet; the next step is a conversation that can reach across the project's maps and source at once, while staying honest about what it actually read.
+- **Shareable report export** — turn the generated manual and maps into a single self-contained HTML/PDF you can hand to a non-technical stakeholder, no server required.
+- **Annotation coverage beyond Python** — hover annotations lead with Python today; bringing JavaScript/TypeScript and Go up to the same depth is mostly prompt and tokenizer work.
+- **Persistent project library** — the project store is in-memory and single-process today; a small on-disk library would let you reopen past analyses without re-scanning.
+- **Jump-to-definition reading** — click a function or import and jump to where it's defined or used, so reading a call doesn't mean losing your place.
+
+Have an idea or a codebase that confuses CodeABC? Open an issue — the roadmap is shaped by what real non-coders get stuck on.
 
 ## Contributing
 

README_CN.md

@@ -188,21 +188,22 @@ npm run tauri:dev
 
 ## 路线图
 
-- [x] 项目说明书生成
-- [x] 悬停批注（优先支持 Python）
-- [x] 术语词典（鼠标悬停关键词弹出解释）
-- [x] 自然语言编辑（"把分析的股票从茅台换成比亚迪"）
-- [x] 提问模式（选中代码随时问）
-- [x] 英文界面
-- [x] 测试覆盖地图（哪些文件有测试；未覆盖的核心文件按风险排序）
-- [x] Git 历史洞察（变更热点、协同变更耦合、代码归属 / 知识孤岛）
-- [x] 技术债地图（汇总代码里作者自留的 TODO/FIXME/HACK/XXX 标记，按文件排序）
-- [x] 环境变量清单（项目读取的环境变量，区分必填与可选）
-- [x] 入口点检测（程序从哪开始运行：`__main__` 脚本、声明的命令行命令、按惯例的入口文件）
-- [x] 逻辑复杂度排序（从 AST 看哪些 Python 文件判断分支最多、逻辑最绕——和“中心度”“变更频率”是不同的角度）
-- [x] 一键启动（`run.py` / `start.bat`：自动构建、起服务、打开应用）
-- [x] 单进程服务（后端直接托管构建好的界面，一个地址搞定，无需另起开发服务器）
-- [x] 桌面端（Tauri —— `npm run tauri:build` 把同一套界面包成原生窗口）
+### 已交付
+
+阅读体验已经完整跑通：大白话项目说明书、悬停批注、术语词典、自然语言编辑、片段问答，全部在双语界面里，一条命令就能启动（也可以打包成 Tauri 原生桌面应用）。在这之上是十几张无需 API Key 的确定性图谱：阅读路线、核心模块排序、测试覆盖、Git 历史热点与代码归属、技术债标记、环境变量清单、程序入口、外部依赖、错误被吞点、最缺文档的文件，还有逻辑复杂度。
+
+### 后续规划
+
+按优先级大致排了一下接下来想做的方向：
+
+- **PR 阅读模式**：粘贴一个 PR 链接，用大白话讲清楚这次改动改了什么、为什么可能重要、先看哪几个文件。这是被问得最多的扩展，也最贴合现有的批注引擎。
+- **整项目对话**：现在的问答只盯着单个文件或片段，下一步是让对话能同时调动项目的各张图谱和源码，同时如实交代它到底读了哪些内容。
+- **可分享报告导出**：把生成的说明书和图谱打包成一个自包含的 HTML/PDF，直接发给不懂技术的同事或老板，对方不用起任何服务。
+- **批注覆盖更多语言**：悬停批注目前优先 Python，把 JavaScript/TypeScript 和 Go 做到同样的深度，主要是 prompt 和分词的工作量。
+- **项目库持久化**：现在的项目数据存在内存里、单进程一次性，做一个小的本地项目库后，重开历史分析就不用再扫一遍。
+- **跳转到定义**：点一个函数或 import 就跳到它定义、被用的地方，读到一处调用时不会丢失上下文。
+
+有想法，或者有让码上懂犯迷糊的代码库？欢迎开 issue，这份路线图就是跟着真实非程序员卡住的地方走的。
 
 ## 贡献