Two editors, two jobs

The Reports page has two editors, and they are not the same tool wearing different hats. Fields is the fast path: fill in the title page, abstract, discussion and declarations, press Generate, and get a correct document in the house style. Doc Editor is for the other case — a thesis chapter, a journal submission, a report whose section order the template did not anticipate — where you need the page itself, not a form.

Fields vs Doc Editor

The one rule

A linked value whose experiment has since been deleted resolves to a visible [missing: …] rather than quietly keeping the number it used to hold. A document that silently preserves a deleted result is worse than one that admits the gap.

Templates

A template here is a real starting document, not a stylesheet: page size, typography, section order, and seed text that already contains your study’s numbers. A freshly created Methods section reads “Recordings were collected from 12 participants” with that 12 linked to the record.

What ships

Changing your mind later

Picking a different template on a document you have already written offers two things: apply only the page setup and typography (your text is restyled in place, nothing is deleted), or start again from that template’s own sections. The first is the default, because losing an afternoon’s writing to a dropdown is not an acceptable way for an editor to behave.

A page that is a page

Paper, margins, zoom

A4, Letter, Legal, A5, B5, A3, Tabloid, Executive or a custom size in millimetres; portrait or landscape; margin presets (Normal, Narrow, Moderate, Wide, Thesis with a bound edge, IEEE) or exact values; a binding gutter; header and footer distances from the paper edge. The document is laid out at the real text frame, so what you see paginates exactly as the PDF prints — the same object, laid out once.

┌──────────────────────────────┐  A4 · 210 × 297 mm
│           header             │
│   ┌────── margin ───────┐    │
│   │                     │    │  the text frame the
│   │   document content  │    │  document is laid out in
│   │                     │    │
│   └─────────────────────┘    │
│      1  ← footnote band      │
│           footer        1    │
└──────────────────────────────┘

Zoom (50–300%, fit width, whole page) is a view transform, so point sizes stay canonical: at 100% an A4 page is A4.

Columns that snake

One to four columns, and the text really does flow from the foot of one column into the head of the next — on screen and in the PDF. The IEEE template is two-column out of the box. Page numbers count sheets, not columns.

Headers, footers, numbering

Header and footer text, page numbers, and a “different first page” switch so a title page carries no running head. Heading numbering is decimal (1, 1.1, 1.1.1), IEEE roman (I., A., 1)) or none — and sections that should not be numbered (Abstract, Contents, References, Declarations, Appendix) take no number and consume no counter, along with everything under them.

Linked study values

Type two braces

Typing {{ opens your study: participant counts, sampling rate, dataset names, every experiment’s every metric, the best model by each metric, the results digest. The list shows the value beside the name, so you can see it is the right one before it lands in a sentence you will have to defend.

The proposed model reached {{acc
                           ├── Accuracy · exp.eegnet.accuracy
                           │   Experiment · EEGNet · 87.0%
                           ├── Best Accuracy · best.accuracy
                           │   Best result · 87.0%
                           └── Accuracy · exp.csp_lda.accuracy
                               Experiment · CSP+LDA · 82.0%

Values are typed, not strings: an accuracy renders as 87.0%, a loss as a number, a date as a date, a list as a comma-separated list. The inserted run keeps the key, so it re-resolves rather than going stale.

What a refresh does

Refresh from study, and every export, runs the same pass:

1. rebuild the value catalogue from the study record;
2. re-resolve every linked value;
3. rebuild generated tables from the record;
4. renumber headings and captions, and re-measure the footnote band;
5. move any heading stranded at the foot of a page or column;
6. refill the contents with the page numbers a reader will turn to.

Tables and figures from the record

Insert ▸ Study content offers a results table (one row per experiment, every metric the record holds), a study summary row, a configuration table, the recorded statistical comparison, saved Analysis figures, a pipeline snapshot, the integrity statement with its results digest, and a table of contents. Where the record holds nothing, the document says so in place — “No statistical comparison is recorded for this study” — rather than leaving a gap or inventing a result.

Tables, pictures, pipelines

Tables

Insert any size; add and delete rows and columns; merge and split cells; set column widths or distribute them evenly; shade cells; toggle a header row; pick a look (Grid, Academic rules-only, Striped, Plain); add a caption that numbers itself. Linked values work inside cells, so a comparison table can be wired to the record cell by cell.

Pictures and text wrapping

Insert from a file or paste a screenshot straight in. Resize by width in millimetres, fit to the measure, scale by a percentage, rotate, flip, crop by a percentage off each edge, and add a caption. Four wrap modes:

Square wrapping survives into HTML (as a CSS float) and LaTeX (as wrapfigure).

Pipeline snapshots

The report gets its own copy. In the arrange dialog you can drag blocks, nudge them with the arrow keys and tidy them into rows to make the figure read on a printed page — and nothing else: no adding, deleting, rewiring or reconfiguring, and no change of any kind to the pipeline in your study. Double-click a snapshot later to rearrange it again.

Writing tools

The ribbon is deliberately unoriginal — Home, Insert, Layout, References, Review, View, with bold where bold has always been. Fonts, sizes, colours, highlight, super/subscript, small caps, alignment, line and paragraph spacing, indents, bulleted, numbered and multilevel lists, styles (Title, Heading 1–4, Body, Quote, Code, Caption, Reference), format painter, undo/redo, find and replace, word count, and the usual keyboard shortcuts.

Outline, sections, numbering

The outline panel lists every heading. Click one to jump to it; drag one to move the whole section — its sub-headings, tables and figures travel with it, and the document renumbers itself. That is the answer to “this journal wants Methods after Results”, which is not a thing anyone should do by dragging text selections through forty pages.

Captions, references, footnotes

Captions number themselves in document order (Figure 1, Table 2 …). Cross-references insert the target heading’s text. Footnotes are real footnotes: a superscript marker where you put it, the note set in a strip at the foot of that page, renumbering automatically when you insert one in the middle.

Export

What it does not do

• Cross-references insert text, not a live number. Renumbering rewrites the heading, so a stale reference is visible rather than silently wrong.
• The footnote strip is uniform. Every sheet reserves the height the busiest sheet needs. Spread-out notes look exactly like Word; twenty notes on one page cost that strip everywhere.
• Word’s own footnote format is not written on DOCX export — notes go in under a Notes heading instead.
• Nothing spans the columns. In a two-column document the title and abstract sit in the first column rather than across the top. A venue that wants a spanning title wants the LaTeX export and its own class file, which is what that template is for.

Plan and storage

The Doc Editor is available on Researcher and up, the same plan gate as report generation. The document is stored inside the study itself, alongside the report template, and is saved when you press Save, when you leave the page, and before switching studies — so a report travels with the study rather than as a file someone has to remember to keep next to it.