This generator makes printable note paper; it does not sit in the lecture for you or summarise anything automatically. The sheet is divided into a cue column, a main note area and a summary strip at the bottom, and the paper size, column width, header, ruled lines, dots or grid can all be changed. PNG and SVG show the first page, PDF produces as many pages as you configure, and JSON saves every template parameter.

The page lays out in physical units rather than inferring paper size from the screen width. It suits using one recording habit consistently across a course, a reading block or an interview, and then turning the paper into questions, recall practice and follow-up actions.

1. 01Choose A4 or Letter and set a margin of 10–25 mm. Test-print one page first to find the printer's unprintable edge.
2. 02Set the cue column to 20%–45% of the writable area. Narrow it when you need short keywords, widen it when you want full questions; there is no ratio that is simply “more effective”.
3. 03Choose ruled lines, dots, grid or blank, set a spacing of 5–12 mm and a summary strip of 25–60 mm. Leave more room when your handwriting is large or you want to draw.
4. 04Fill in the course, project, an optional date and custom column headings. Avoid pushing long body text into the header; full content belongs in the note area.
5. 05Export a PDF of 1–30 pages, choose “actual size / 100%” when printing, and switch off the printer software's extra headers and footers. Use SVG when you need vector lines or further layout work.
6. 06After the session, add cue questions, cover the main notes and try to answer, then check what you missed and write the summary. Save the JSON so the same template can be restored next time.

The Cornell Learning Strategies Center attributes this method to the education scholar Walter Pauk and points to his How to Study in College. The key to the method is not the three rectangles on the page but the questioning, recall, reflection and review that follow the record. Cornell's instructions use the cue column as a recall prompt rather than a second copy of the lecture text. [1][2]

Compared with ordinary ruled paper, the divided layout reminds you to separate the record from the retrieval cue; compared with a mind map, it suits sequential recording and answering against questions rather than expressing topic relationships radially. The template itself measures nothing about your understanding. Retrieval-practice research supports the general direction that actually attempting recall is worth building into study, but it cannot be used to promise that this tool, a particular column width or handwriting works better for everyone. [3]

When the same layout is borrowed for an interview, the cue column can hold follow-up questions, the main area the observations and consented quotations, and the summary strip the findings and counter-examples. That is a layout borrowing; it should not be claimed as a validation of the original study method.

The template “Lecture notes · A4 ruled, two pages” uses A4, a 15 mm margin, a 30% cue column, 8 mm line spacing and a 40 mm summary strip. A4 is about 210 mm wide, so the writable width is about 210 − 2×15 = 180 mm and the cue column about 180×30% = 54 mm. The PDF export has two pages, while PNG and SVG preview the first page only.

The second template is “Interview record · Letter dotted”: a 35% cue column, 6 mm dot spacing, and columns renamed to follow-up questions, observations and quotations, key findings, and next steps. The example keeps a blank participant code rather than prefilling any real interviewee details.

One workable study loop is: record “the difference between a sample and a population” in the main area; write “when can a sample result not be generalised directly?” in the cue column; cover the main area and answer; then write the conditions of applicability and what is still uncertain into the summary strip.

• The PDF reuses the repository's existing text-to-PDF canvas path and then generates the file with pdf-lib, so no font is copied and no new font dependency is added.

• To stay compatible with the CJK and Latin system fonts a browser can reach, the PDF pages are raster images: the text cannot be selected and this is not a tagged accessible PDF.

• SVG keeps the lines as vectors, but how the text appears still depends on the fonts of the receiving device.

• A printer may rescale or crop the edges, so the on-screen preview is no substitute for a test print.

• An over-long header eats into the writable area; check the warning shown in the preview.

• There is no handwriting recognition, no automatic review scheduling and no multi-device sync.

Filling the right-hand column and never using the cue column for recall means the main learning action of the method is unused. Writing a summary as an abbreviation of the original text proves nothing about understanding either. Another frequent problem is treating “fit to page” in the print dialogue as 100%: after printing, the line spacing and column width may no longer match the settings.

• Lectures, close reading of an article, technical training and any learning activity that needs recall afterwards can all try it.
• Interviews, observation and project records can borrow the divisions, as long as facts, interpretations and open questions are kept apart.
• When a mathematical derivation or a complex diagram needs more free space, choose grid or blank paper, or switch to another recording method.

Paper generation, previewing and exporting all happen in the local browser. Autosave is subject to browser quota and cleanup, and JSON is the better backup. A share link contains readable template and header content and is not an encrypted share; real interviewee identities, personal classroom details and confidential project names should not be written into a public link.

[1] Cornell Learning Strategies Center: method background and modules, which lists the works and chapters of Pauk and Owens. （访问日期：2026-09-22）

[2] Cornell: note-taking instructions adapted from Walter Pauk, covering record, question, recall, reflect and review. （访问日期：2026-09-22）

[3] Karpicke and Blunt, 2011, retrieval practice research. This is related learning research, not an empirical test of this generator. （访问日期：2026-09-22）