The algorithmic discovery of latent connections within a corpus, presenting them as suggestions rather than claims about semantic structure.
Hard and soft links
Where explicit hyperlinks and references are author-curated (hard data), dynamic hyperlinks are inferred (soft). A heatmap layer shows connection density across the corpus — varying intensity signals how many related mentions exist elsewhere. Selecting text surfaces those mentions ranked by similarity.
The implicit network is irreducibly dense — any sufficiently rich corpus contains countless valid relationships. Dynamic hyperlinks acknowledge this by treating discoveries as exploration aids, not as “true” structure. This humility distinguishes them from explicit links, where actors commit to meaningful connections.
Related components
- Reference — explicit, typed entity linking (the hard counterpart)
- Bubble menu — selection-anchored interaction surface
Resources & references
- Inspired by Linus Lee’s Hyperlink Maximalism and the Notation app.
- Rao (2009) The rhetoric of the hyperlink — the reader-side complement to the “irreducibly dense network” claim: each reader’s click-trail produces a unique text from the same hypertext.
Related patterns
Complements
- Link preview — inferred links benefit from preview since actors have less prior expectation of the destination
- Suggestion — suggested actions and connections
- Annotation — user-created markers (hard) and system-generated insights (soft)
- Deep linking — URL-based addressability
Related
- Information architecture
- Modality
- Prose — Dynamic hyperlinks are prose-as-affordance: the link text is the prose, and the move from authored phrase to actionable target is itself the prose move.
Preceded by
- Data & Information — hard and soft data layering