VERSOFOLD 0.4.1 · HANDBOOK

Learn with the practice project

A hands-on paper with synthetic data, an editable figure, a table, and connected results.

← All guidesRead as Markdown ↓

The practice project is a small, editable research paper that connects several VersoFold tools. You can change every part of it. All six measurements, method names, and reviewer comments are invented for learning. They are not evidence for a real scientific claim. No AI account, remote connection, or downloaded dataset is needed for the exercises.

Create your own copy

On Home, choose Try a practice project. Choose a destination and a new folder name, then choose Create workspace. You can also select Practice project under New workspace → Start with. Creation never replaces an existing folder: choose a different name if that folder already exists.

The walkthrough opens over your new workspace. Close it whenever you want to try a step; reopen it through Documentation, search for “practice”, or open START-HERE.md in the project's file list. Help → Explore workflows connects common goals to the relevant tools and guides. You can use just one of the exercises at a time.

Your copy contains:

ItemWhat to try
main.texEdit the author, prose, equation, and captions.
data/results.csvSix synthetic runs for two fictional methods.
Figures → Figure StudioThe saved “Scores by run (synthetic)” chart and its linked dataset.
Write → Table editor“Practice scores”, an editable sheet with mean formulas.
Research → Linked results“Guided mean score (synthetic)”, connected to a number and a sentence.
Write → Manuscript boardCards linked to the existing sections; Discussion has a practice checklist.
references.bibA BibTeX entry describing the local toy dataset, not a published paper.
notes/reviewer-feedback.mdFictional reviewer requests to use in Rebuttal.

First success: edit and compile

  1. Open main.tex under Files and find \author{Your name}. Replace the name. You can edit supported prose in Visual too.

  2. Choose Recompile. The example uses pdfLaTeX and standard packages included with the desktop app. If using a development installation, a working TeX installation is needed.

  3. Read the PDF. Find the equation, the two-series plot, the table, and the bibliography.

  4. Use Review → Versions to save a named version such as “Before practice edits”.

Expected result: your name appears in the compiled PDF and the initial Guided mean is 82.0%. Saving source and compiling a PDF are separate actions. See Compilation & PDF if the build reports a problem.

Trace and update a number

  1. Open Research → Linked results and select Guided mean score (synthetic). Inspect its definition: file data/results.csv, column score, filter method = Guided, operation mean, one decimal place, percentage display. Three rows match.

  2. Close the tool. Open data/results.csv in Files. In the last row, change the score in Guided,3,0.84,9.5 from 0.84 to 0.90, and save.

  3. Reopen Linked results and Refresh. The mean of 0.80, 0.82, and 0.90 is 0.84, so the new displayed value is 84.0%.

  4. Review the proposed numeric replacement and the connected comparison sentence. Decide whether the sentence is still justified within this synthetic example. Select the intended numeric update and acknowledge the statement only after reading it.

  5. Apply the reviewed update, then recompile. The app preserves a checkpoint before changing the manuscript.

Expected result: the Results section displays 84.0% after you accept the update. The sentence is reviewed by you; it is not automatically rewritten. The CSV-linked figure follows the changed input when regenerated during the build. The independent spreadsheet table still has its original values until you edit and publish it.

To reverse this exercise, change the CSV score back to 0.84, refresh Linked results, review the return to 82.0%, and recompile. See Linked results for ambiguous anchors, stale data, and recovery.

Refine the figure

Open Figures → Figure Studio and select Scores by run (synthetic). Inspect the source rows, then change a line color, marker, or axis label in the design controls. Preview the result and publish the figure. It is already included in the manuscript, so you do not need to insert another copy. Recompile to inspect the PDF at its actual printed size.

The chart uses the CSV directly and shows individual runs. If you edit data inside Figure Studio, it may create an editable dataset copy; use the CSV file for the linked-result exercise above. Read Figure Studio for dataset preparation, uncertainty, themes, and publication.

Edit and republish the table

  1. Open Write → Table editor and select Practice scores.

  2. Inspect cell E3: its formula is =AVERAGE(B3:D3). The data cells use percentage formatting, while their underlying values are fractions.

  3. If you changed the CSV in the number exercise, change D3 from 0.84 to 0.90 here too. This sheet is an independent copy, so CSV changes do not rewrite its cells. E3 should display 84.0%.

  4. Choose Save table, inspect the publication preview, then Publish to paper. The existing manuscript already includes the generated table file; do not insert it a second time.

  5. Recompile and compare the table with the figure and Results paragraph.

Expected result: publishing changes the table used by the paper; merely saving the sheet does not replace its previous published output. See the table guide.

Plan and review a revision

Open Write → Manuscript board. The cards are connected to the actual document sections. Select Discussion, inspect its word target and checklist, then add a sentence to that section explaining a limitation of the toy data. Return to the board, review the changed source, check the relevant task, and move the card to In review. A card's stage is your planning decision, not a guarantee of scientific quality.

For response practice, open notes/reviewer-feedback.md. Copy its fictional review, choose Import reviews in the Reviews tab (or move the paper to the Rebuttal phase), and paste it. In the Response view, decide how to address one comment, make the change in the paper, and write a short answer under the comment explaining what changed. See Write the response and Copy for OpenReview.

Choose your next workflow

  • Reading: open one of your own PDFs from Home, then try notes and comparison in the reading desk. The practice bibliography has no external paper attached, so there are no source passages to check.

  • Sources: add a real paper with its PDF to your library, cite it, then check the sources of the sentence that cites it. Synthetic practice results cannot support an external claim.

  • Page layout: after a successful build, open Page layout and inspect where the existing figure and table land.

  • Submission: use Submission rehearsal with requirements you have checked. No real venue or deadline has been preselected for this example.

  • Sharing: create a web edition draft or try a local export. Creating this practice project publishes nothing online.

  • Agent help: if you connect a model, try “Read the practice-project documentation and explain how this paper's number is connected to its data.” Reading the handbook remains on demand; the local exercises do not start an agent conversation.

Start fresh or keep the project

The example becomes an ordinary folder in your chosen location. Keep it, rename its library entry, or make an independent copy from Home. To repeat the initial exercise, create another Practice project in a different new folder. Removing a workspace from the library leaves its files on disk. You can always start a separate blank or article workspace for your real research.

← Browse the handbookGet VersoFold →