Frame: the problem and its names

This file states the problem of showing a deep tree of text to a reader, then the names research and practice give it, and the sub-questions the review answers.

The problem in one sentence

A body of writing is stored as a tree of units nested to any depth. Decide how it is shown, so the reader keeps the top levels in view and can still reach any leaf.

Terms used in every file of this folder

Term Definition
node one unit of the tree: a heading or summary that stands for it, and a body that may hold child nodes
depth in storage the number of levels the stored tree has
depth in view the number of levels a reader sees open at one time
folded a node shown by its heading or summary only, its children hidden until the reader opens it
focus the node the reader is reading now
context what is shown around the focus: its ancestors, siblings, or an overview of the whole

What the field calls it

Field name What it names Entry point
stretchtext text that expands and contracts in place, so the surrounding text stays as context Wikipedia 2026c: defined in print in 1970, described in an unpublished note by Nelson in 1967
adaptive presentation, canned-text adaptation a system chooses which fragments of a page to show or stretch for this reader Brusilovsky 2001
focus plus context (focus+context, F+C) focus and context fused in one view, often by eliding or shrinking the far parts Cockburn 2009
fisheye view, degree of interest (DOI) show a node when its importance minus its distance from the focus passes a threshold Furnas 1986
overview plus detail (O+D) a separate overview pane beside the detail view Cockburn 2009
zooming, zoomable user interface (ZUI), semantic zoom one view at a time, moved in scale; semantic zoom changes content, not only size Bederson 1994; Hornbæk 2002
expand/contract table of contents a contents tree opened one level at a time Chimera 1994
hypertext structure: hierarchical, network, mixed how the links between nodes are arranged Mohageg 1992; McDonald 1998
disorientation, "lost in hyperspace", cognitive overhead the reader not knowing where they are or how to go on Conklin 1987
layered notice a short first layer with the full text one step away McDonald 2009
recursive summarization, tree-organized retrieval a model summarizes chunks, then summaries, into a tree it reads or searches Wu 2021; Sarthi 2024
progressive disclosure for agents an agent loads a short description first and the body on demand He 2026

What practitioners call it

Practitioners say The field says
accordion, collapsible section, expander, "show more" stretchtext; folded node
details and summary (HTML), disclosure widget stretchtext with one level (WHATWG 2026)
tree view, outline view, sidebar navigation expand/contract contents; fisheye of a tree
progressive disclosure, staged disclosure disclosure of function, later of content (Nielsen 2006b; Carroll 1984)
code folding, #region manual fisheye of program text (Cockburn 2009)
outliner, block editor, zoom into a bullet ("hoist") outline processor; zooming on a subtree
breadcrumbs the ancestral line of the focus, the zero-order fisheye of Furnas 1986
table of contents beside the text, minimap overview plus detail
skills, llms.txt, "load on demand" progressive disclosure of context (He 2026; linked review below)

Sub-questions

Id Question
Q1 Does showing a unit with its children folded change comprehension, search time or recall?
Q2 Hierarchical against networked hypertext, beyond what the linked reviews hold
Q3 Focus plus context, fisheye and zoomable views of documents and outlines
Q4 How many levels a reader handles in one view, and whether depth in storage matters once depth in view is capped
Q5 Outliners and block editors in practice, with adoption numbers
Q6 Trees of summaries for models as readers, measured on benchmarks
Q7 Hidden content left unread, disorientation, deep nesting

Search terms tried, three ways

  • Academic: stretchtext, expandable text, collapsible outline, progressive disclosure, focus plus context, fisheye view, overview plus detail, zoomable user interface, hypertext structure, disorientation, layered notice, recursive summarization, tree-organized retrieval.
  • Acronym: F+C, O+D, ZUI, DOI, RAG, RAPTOR.
  • Author: Nelson, Furnas, Bederson, Cockburn, Hornbæk, Shneiderman, Chimera, Egan, McDonald, Rada, Brusilovsky, Höök, Sarthi, Wu.
  • The word "outliner" could not be searched: both engines returned unrelated millions or tens of thousands, so outliners are covered through practice (protocol section 4).

Not reviewed here, linked instead

  • Recall of the top levels of a text, the levels effect: psycholinguistics.
  • Depth against breadth of menus and folders: file-structure.
  • General before detail, advance organizers, zoom-lens sequencing: reading-order.
  • Hypertext cognitive load, signaling, segmenting: learning-science.
  • Collapsed content missed, as one company article: typography-and-layout.
  • Progressive disclosure of context to agents, instruction files: llm-readers.
  • Hierarchical proofs, Lamport's steps and levels: mathematical-writing.
  • Slide builds as progressive disclosure: presentations.
  • Storing the tree, reuse and validation: topic document-model of this repository; unit types: topic information-typing.