The applications in my "core" homelab (i plan to develop)

• serving monitoring application
• budget tracker application
  • record expenses everyday
  • also record gains (from allowance, salary, revenue, assets etc.)
• reminder tool
  • daily quests as todos (workout, programming)
  • anime/manga/manhwa/manhua updates
  • tech/dev daily news
  • llmarena/latest AI models updates
  • backlog clearing/old reads to revisit
  • "Core reminds me to..." (philosophical, grounded mental)
    • self learn, have strong fundamentals (language-agnostic)
    • job hop for growth potential
    • the work you do needs to be broadly communicated for visibility in order to be considered for high ratings and compensation increases.
    • programming is just moving data around and managing state.
  • academic/organization work
  • job application/openings (opt, since i have linkedin, etc.)
• wardrobe management system
  • start picturing clothes with name (like in uniqlo), description
• learning management system (quiz or LMS)

  • catered to myself.

  • Course -> Lesson -> Topic -> Question -> AnswerOption

  • see: High-level approach: two buckets - content vs user state

    [COMMENT] High-level approach:

    Think in two buckets: content vs user state.

    Content is a tree and never knows about users Course → Lesson → Topic → Question → AnswerOption Everything is ordered explicitly (order fields). No progress, no users, just structure.

    User progress lives in junction tables Instead of flags on content, you store:

    Enrollment (user ↔ course)
    
    LessonProgress (user ↔ lesson)
    
    TopicProgress (user ↔ topic)
    

    That’s where completion, last activity, and resume pointers go.

    Resume beats percentages Store “last lesson / topic / question” so you can drop users exactly where they left off. Percent complete can always be computed later.

    Quizzes are sessions, not states A QuizAttempt = one user, one topic, one run. Each attempt has many QuizResponses. Attempts are immutable history; score is saved on the attempt.

    History vs snapshot Attempts/responses = truth. Progress tables = cached “current state” for fast loading.

    Prereqs = graph, not booleans Model prerequisites as TopicPrerequisite(topic, prereq_topic) so you can enforce or extend later without schema pain.

    If you separate content, progress, and attempts cleanly, everything else (analytics, retries, certificates, adaptive paths) falls out naturally.