Stories from the atlas
A story presents an authored route through philosophical research, with a companion graph that follows the discussion. It connects continuous prose with inspectable treatments, people, ideas, works and sources.
Reviewed 14 October 2026 · Prototype documentationStories use research already saved in the atlas. They can be written by a person, developed with AI assistance, or edited collaboratively. The current prototypes test the reading experience and production method; a fully automated authoring interface is planned.
Reading and exploring
On a wide screen, the article appears beside a companion graph. Scrolling activates a section and changes the graph’s focus. Selecting a node opens its full title, summary, qualified dates, source links and a link to the full Explorer. Selecting a connection explains the relationship it represents.
Dragging, zooming or inspecting details pauses automatic following. Resume reading returns to the current article section; Fit scene restores that section’s framing. The article remains readable with JavaScript disabled.
Spatial companion
The selected-anchor prototype uses a stable spatial arrangement. The camera moves between the chosen accounts, while surrounding records remain recessed. Spatial distance is a reading aid rather than a measure of philosophical similarity.
Temporal companion
Treatments appear above a fixed context area containing People, Ideas and Works lanes. Treatment positions follow qualified dates; context lanes group records by type and do not represent chronology. Shared context records retain their slots between adjacent sections.
Hover highlights a node and its connections without a popup. Full titles remain available by selecting the node. The episode selector offers direct navigation; How to explore contains the legend and an option to show nearby treatments. Dashed date spans indicate approximate author context rather than an asserted date for a treatment.
Writing modes
| Mode | Organization and voice | Current status |
|---|---|---|
| Selected anchors | Compare a chosen set of thinkers or treatments around a shared problem. A conversational comparative essay can move between their distinctive accounts. | Prototype · Augustine, Aquinas and Hegel; spatial companion. |
| Development through time | Follow changing questions and accounts through dated episodes, retaining branches and gaps. Chronological transitions provide a more linear narrative. | Prototype · The Trinity through time; temporal companion. |
| Problem pathways | Follow subproblems, alternatives and conceptual connections across periods without requiring chronological order. | Planned · no published prototype yet. |
Writing mode determines the route through the material. Rhetorical style and graph presentation are separate choices, although the current prototypes pair each mode with a particular presentation. Future modes may require different companion designs.
From research to a story
- Define the question and scope. Select a collection, ideas or explicit records, and choose a writing mode. Set an intended length and level of detail.
- Inventory the saved material. Local tools gather active canonical records, qualified dates, source links and relationships. The temporal inventory also distinguishes treatment dates from author-context placements.
- Plan the route and scenes. Human or Codex judgment selects focal accounts, supporting context, section questions and transitions. Identify important gaps before drafting.
- Write from the existing research. Compose the article from saved accounts and their support. Essential missing evidence returns to the research workflow; story production does not initiate an unrecorded source crawl.
- Review grounding and prose. Check substantive claims, chronology and historical verbs against the selected material. Edit for a coherent reading experience, including readers who do not open the graph.
- Export and render. Local tools resolve identities, export the bounded graph bundle, render the Quarto article and refresh the story catalog. Review the article and transitions in the browser.
Interpretive selection and writing currently take place through human–Codex collaboration. Export, structural checks and catalog generation are local functions. The existing workflow produces reusable artifacts, but it is not a one-click story generator.
Authoring options
These options are currently expressed in the saved plan and editable article. They are not all exposed as controls in the application.
| Choice | What it controls | Boundary |
|---|---|---|
| Scope | A collection and/or selected canonical records. | A story is a selection of material, not a complete export of a collection. |
| Mode and route | Comparative anchors, chronological episodes or a future thematic pathway. | A narrative transition does not create a historical relationship in the graph. |
| Voice and length | Rhetorical style, intended audience, section structure and prose detail. | Currently editorial choices; automatic enforcement of general word and section budgets remains planned. |
| Scene focus and support | Which accounts receive emphasis and which records supply context. | Export uses bounded scenes and active, resolved identities. |
| Graph presentation | Spatial or temporal companion, with mode-specific navigation. | Appearance does not change node identity or relationship meaning. |
| Temporal range | Episode windows and the initial date-axis range. | Publication, composition, activity and lifespan dates retain their distinct qualifications. |
| Associated context | Include connected people, ideas and works around selected treatments. | Only existing structural connections are exported; display proximity supplies no new evidence. |
Editorial standards
The narrative discusses philosophical ideas directly. It should not refer to its own collection, graph records, software or production process. Interpretive qualifications belong in ordinary prose; software, versioning and AI assistance belong in About this reading.
Chronological succession permits a transition such as “later.” Claims of influence, response or transmission require appropriate support. Conceptual comparisons can be useful without establishing a historical connection, and a later thinker should not automatically be presented as completing an earlier one.
A story preserves links to the underlying accounts and sources. Its footer identifies the writing mode, article edition and date, software release and saved data bundle. Editorial prose does not become a treatment or graph edge unless separately researched and recorded.
Publication and current limits
The story package contains an editable Quarto article, a scene plan, a bounded JSON graph and shared reader assets. The temporal pilot also has a saved inventory and grounding receipt. The Stories index and each collection’s Stories tab use a rebuildable catalog of rendered articles.
Reading uses static HTML, CSS, JavaScript and JSON; it requires no database or live model call. The local application provides the shared navigation and links to the Explorer. Packaging as a reusable Quarto extension, checking the external site’s theme and paths, and validating the full reading experience on phones remain future work.
Each story uses a saved snapshot. Later graph corrections do not silently rewrite that article or bundle: updating a story requires re-export and editorial review. This supports a stable reading while preserving the possibility of revision.