How to Connect Notes and Sources on a Research Canvas

Limit the board to a few object types. Use source objects for files, URLs, or recordings; evidence notes for checked claims; question notes for open work; conflict markers for unresolved differences; and draft objects for output. Give each type a consistent shape, label, or lane. Color can help, but never make color the only carrier of meaning.

Write the relationship vocabulary in a visible legend. Useful links include supports, qualifies, contradicts, defines, exemplifies, and answers. Avoid unlabeled arrows: they force the next reader to guess whether the connection means agreement, sequence, ownership, or mere association.

A practical left-to-right layout is question, source intake, candidate notes, checked evidence, synthesis, and draft. Put a rejection or parking lane underneath so weak material remains visible without contaminating the accepted set. A PDF, video, and interview can all sit in intake; their verified claims meet later in the evidence lane.

State-based lanes show what needs attention. A source that has not been reviewed should not sit beside accepted evidence as if their status were equal. Movement across lanes becomes an editorial decision: imported, inspected, verified, accepted, used, or retired.

An evidence note should be readable at a glance but carry enough detail to reopen the source: source ID, page or timestamp, one claim, short support, conditions, and status. Link the note to the source object, then link it forward to the question or draft section it serves. The same note can support several sections without duplicating its text.

If the source is a scanned PDF, include the OCR condition. If it is a video transcript, record whether the timestamp was replayed and whether a visual check was needed. The locator belongs inside the note because canvas position can change.

Group evidence around subquestions rather than around whichever source produced it. A cluster may contain two supporting notes, one qualification, and an open question. This makes coverage visible: a lonely claim with no independent support looks different from a pile of duplicate summaries from the same source.

Create a conflict object when sources disagree. Connect both evidence notes to it and record the suspected reason—date, definition, population, method, or unresolved error. Do not draw a direct contradiction line unless both sources address the same proposition.

Use stable IDs and descriptive titles, not “note 14.” Keep raw captures near their source, archive rejected material, and summarize dense clusters with an index note that links to the parts rather than replacing them. When a board becomes wider than one screen, add a compact map of zones and the current research question at the entrance.

Review connections periodically. Remove lines that express no useful relationship, merge duplicate notes only after checking their locators, and move finished evidence into an archive zone without breaking draft links. Spatial organization should reduce retrieval time, not become another artifact that needs interpretation.

Pick each planned conclusion and trace backward: draft section, synthesis cluster, evidence note, source object, and original locator. Then trace forward from every accepted note to see whether it is used, intentionally reserved, or obsolete. Orphaned claims and dead-end evidence are signals to repair the structure.

The finished canvas is not proof by itself. It is an interface for inspecting the proof trail. Export or describe the claim-to-source path with the final output so the reasoning survives after the spatial arrangement is no longer visible.