<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0">
  <channel>
    <title>智码方舟博客</title>
    <link>https://thesis.polars.cc/blog</link>
    <description>AI 项目生成、全栈项目交付、Directus 内容管理和 SEO 优化实践。</description>
    <language>zh-CN</language>
    <lastBuildDate>Sun, 31 May 2026 15:21:13 GMT</lastBuildDate>
    <item>
      <title>AI 毕设怎么从题目变成可运行系统</title>
      <link>https://thesis.polars.cc/blog/ai-bishe-from-topic-to-running-system</link>
      <guid>https://thesis.polars.cc/blog/ai-bishe-from-topic-to-running-system</guid>
      <description>计算机毕设不应该从写代码开始，而应该先把题目拆成角色、页面、数据模型和验收清单，再进入 AI 生成流程。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>用 AI 做毕设，需求应该怎么写</title>
      <link>https://thesis.polars.cc/blog/write-requirements-for-thesis-ai</link>
      <guid>https://thesis.polars.cc/blog/write-requirements-for-thesis-ai</guid>
      <description>一份适合 AI 生成项目的毕设需求，需要包含题目背景、用户角色、功能模块、页面清单、数据对象和运行要求。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>Spring Boot + Vue 毕设项目，范围怎么定才不容易翻车</title>
      <link>https://thesis.polars.cc/blog/springboot-vue-bishe-scope</link>
      <guid>https://thesis.polars.cc/blog/springboot-vue-bishe-scope</guid>
      <description>Spring Boot + Vue 是常见毕设组合，但功能范围要控制在可运行、可解释、可演示，而不是越大越好。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>计算机毕设交付前要准备哪些材料</title>
      <link>https://thesis.polars.cc/blog/thesis-defense-deliverables-checklist</link>
      <guid>https://thesis.polars.cc/blog/thesis-defense-deliverables-checklist</guid>
      <description>交付毕设不只是提交源码，还要准备数据库脚本、运行说明、测试账号、截图、论文材料和答辩讲解线索。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>已有代码项目，怎么继续生成论文或项目报告</title>
      <link>https://thesis.polars.cc/blog/code-to-thesis-report-workflow</link>
      <guid>https://thesis.polars.cc/blog/code-to-thesis-report-workflow</guid>
      <description>代码项目是论文的材料来源。先整理功能、数据模型、截图和运行说明，再让 AI 生成论文初稿会更稳定。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>接外包时，先交付一个能演示的版本</title>
      <link>https://thesis.polars.cc/blog/developer-outsourcing-first-demo</link>
      <guid>https://thesis.polars.cc/blog/developer-outsourcing-first-demo</guid>
      <description>外包项目最怕需求长期空转。先用 AI 把客户口头需求变成可运行演示版本，可以更快锁定范围和报价。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>外包需求边界怎么定，才能减少返工</title>
      <link>https://thesis.polars.cc/blog/outsourcing-requirement-boundary</link>
      <guid>https://thesis.polars.cc/blog/outsourcing-requirement-boundary</guid>
      <description>外包项目返工多，通常不是技术问题，而是没有把角色、页面、数据、权限、验收和变更边界写清楚。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>管理后台和 CRUD 项目为什么适合 AI 先生成</title>
      <link>https://thesis.polars.cc/blog/admin-crud-projects-ai-scaffold</link>
      <guid>https://thesis.polars.cc/blog/admin-crud-projects-ai-scaffold</guid>
      <description>管理后台、CRUD、内部工具和轻量 SaaS 往往结构相似，适合先用 AI 生成可运行骨架，再由开发者做业务细化。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>用积分生成 AI 项目，怎么控制成本和风险</title>
      <link>https://thesis.polars.cc/blog/points-cost-and-risk-control</link>
      <guid>https://thesis.polars.cc/blog/points-cost-and-risk-control</guid>
      <description>AI 项目生成前先试聊、拆范围、确认技术栈，再付费生成完整项目，可以减少无效消耗和返工。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>毕设和外包项目，技术栈应该怎么选</title>
      <link>https://thesis.polars.cc/blog/tech-stack-selection-bishe-outsourcing</link>
      <guid>https://thesis.polars.cc/blog/tech-stack-selection-bishe-outsourcing</guid>
      <description>技术栈选择不要只看流行度，要看老师或客户能否接受、团队是否能维护、项目是否能按时跑起来。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>AI 生成项目为什么需要检查点和继续修改</title>
      <link>https://thesis.polars.cc/blog/checkpoints-continue-modify-ai-project</link>
      <guid>https://thesis.polars.cc/blog/checkpoints-continue-modify-ai-project</guid>
      <description>AI 项目生成不是一次性赌结果。检查点、暂停恢复和继续修改，让生成过程更适合真实交付。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>边生成边预览，为什么能提高项目验收效率</title>
      <link>https://thesis.polars.cc/blog/live-preview-and-delivery-acceptance</link>
      <guid>https://thesis.polars.cc/blog/live-preview-and-delivery-acceptance</guid>
      <description>实时预览让用户在生成过程中看到页面、文件、日志和运行效果，能更早发现方向问题，降低最终验收风险。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>从一句需求到数据模型，AI 项目生成中间发生了什么</title>
      <link>https://thesis.polars.cc/blog/from-chat-to-data-model</link>
      <guid>https://thesis.polars.cc/blog/from-chat-to-data-model</guid>
      <description>AI 生成全栈项目时，真正关键的中间层是把聊天需求提取成结构化需求和数据模型。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
    <item>
      <title>AI 生成项目和模板代码有什么区别</title>
      <link>https://thesis.polars.cc/blog/ai-generated-project-not-template</link>
      <guid>https://thesis.polars.cc/blog/ai-generated-project-not-template</guid>
      <description>模板代码是固定骨架，AI 项目生成应该根据需求、角色、页面、数据模型和技术栈重新组织可运行项目。</description>
      <pubDate>Sun, 24 May 2026 00:00:00 GMT</pubDate>
    </item>
  </channel>
</rss>
