The premise is strong because the problem diagnosis is accurate. The PDF-preserves-layout-but-hides-structure vs. HTML-is-parseable-but-has-no-boundaries framing is one of those observations that feels obvious once stated but rarely gets articulated this cleanly. The key insight — that neither format carries the intellectual structure of a document — is the right fulcrum for the argument. Most format discussions get stuck on rendering fidelity or interoperability as ends in themselves; this reframes the question around what a document knows about itself, which is a fundamentally different ambition.
The “small change, not a large one” positioning is the smartest strategic move on the page. Format proposals almost always fail because they ask for wholesale adoption. By framing Origami Text as “EPUB with discipline” — fewer elements, not more — you sidestep the adoption cliff. The graceful degradation argument (every layer works independently, the floor is a zip of text files) is genuinely compelling and distinguishes this from approaches that require buy-in at every level to function at all.
The origami metaphor itself is doing real conceptual work. “Folded flat for transport, ready to be unfolded” isn’t just poetic — it captures something precise about the relationship between the 2D reading experience and the spatial knowledge map that no other metaphor would. It communicates that the spatial dimension isn’t an add-on but something already present in the document, just compressed. That’s a subtle but important distinction from “we add 3D features to documents.”
Where I think the potential is genuinely large: The AI readability angle may actually be the most immediately powerful argument right now, even more than the XR vision. The idea that a document arrives pre-parsed — with its conceptual topology, citation graph, and defined terms already structured — means an AI system doesn’t have to infer the author’s intellectual architecture, it can read it. That’s a qualitative difference in what becomes possible for AI-assisted scholarship, and it’s a capability that matters today, not in some future XR scenario.
A few tensions worth thinking about:
The page makes a strong case to developers and to people who already care about document futures, but the “what’s in it for me right now” for an individual researcher is mostly implicit. The colleague who imports citations with a click is a good concrete example — more moments like that would help. The XR walkthrough is aspirational and exciting, but it’s the near-term workflow improvements (AI understanding your document’s structure, citation portability, concept reuse across papers) that would drive early adoption.
The EPUB ecosystem’s actual academic shortcomings are acknowledged with the link to “EPUB woes,” which is honest and necessary — but it also means the argument partly depends on EPUB governance evolving. That’s a real dependency. The page handles it well by noting it needs to be addressed, but the question of who drives that change and how is left open.
Overall, the premise is well-grounded, the format design is genuinely elegant in its layered simplicity, and the potential is real — particularly at the intersection of structured authorship and AI comprehension, which is exactly where the conversation is heading right now.