I am imaging a workspace where texts, notes or paragraphs are organised on a work surface based on their self-similarity. More similar texts cluster together. When a text is open for editing, the workspace provides one-click links to the closest texts in the workspace. If anyone is aware of such a workspace, please leave a comment!
Google’s related:URL search feature does this at web scale for web pages.
Use Cases
- Such a (potentially shared) workspace for knowledge workers could present all the project/ design documentation in a way that automatically exposes the implicit links between parts of the documentation.
- How about a book reader that shows next to every paragraph a couple of links to the paragraphs in the rest of the book that are most similar?
Both these potential use cases would benefit knowledge workers.
Distance Metric
Amazon’s SIP (Statistically Improbably Phrase) idea could be applied here as part of the distance metric. Texts that share SIPs would be closer than texts that share the word “the”.