Hacker News Reader

Comment by sandreas | Hacker News Reader

What I'm missing in modern ebooks (like epub format) is more metadata. Who's talking (character data)? What emotional aspects does the scene have (angry, happy, sad, in a hurry)? Where does the conversation take place (geodata)?

I'd love to see at least:

  - character: ID, Name, Gender, Age
  - mood: ID, Name (Happy, Sad, Angry, ...)
  - place: ID, Name, Acoustic (Outside, Inside, Cave, ...)

This could be prepared by the author, work as a glossary, enrich the whole ebook experience and also would be a great preparation to teach AI voices how to convert a book into an audiobook.

View on HN

What's the point of reading a book, then? The joy of reading fiction is to try to understand the humanity in the scene. I don't need the author to force feed me all of these details. I want to wrestle with the answers, to try to grasp what it might mean.

TEI is something like that, but the amount of effort required to mark a book up like that would be astronomical.

Starts to sound like the kind of task an AI could do reasonably well though

If the goal of these tags are metadata for AI consumption, and the solution to generate them is “use an AI”… what is the point?

Specialization I presume, so one produces the metadata that can be consumed by another.

Also, the thing from the above post that stood out to me would be to act as a reminder for the reader. Not so much the location and emotion, but the character data. I've often found myself wondering who the character is that's appeared in a scene, forgetting that they previously appeared earlier.

The challenge would be balancing that metadata richness without turning the book into a spreadsheet, but if done well (maybe opt-in layers or a toggle), it could really deepen the experience

That sounds like you are asking for a play.

If it can be derived from the book text, then LLMs or reader can already derive it.

If it can’t be derived from the book text, then it’s extra content that probably shouldn’t be there because it came from elsewhere.

Reply on news.ycombinator.com