The process record

Evidence you cannot make later

The strongest defence against any allegation about your work is a record of it being made, and it has one property that decides everything: it cannot be created afterwards. Version history is timestamped by the platform. Commits carry dates. Notes accumulate. None of it can be produced the week after an accusation, which is exactly why it is convincing.

So the record has to exist before you need it, which means it has to be a habit rather than a project.

The cheapest possible version

One rule: write in a tool that keeps version history, and never paste a finished draft into an empty document.

Google Docs keeps detailed version history automatically and free. Microsoft Word with autosave to OneDrive or SharePoint does the same. Both let you show a document growing over hours and days, with pauses, deletions, and reordering: the shape of somebody writing.

That single habit is most of the protection, and it costs nothing.

The common way people destroy their own evidence is drafting somewhere else and pasting the result in. A document whose history shows 2,000 words arriving in one action at 23:40 tells a story you cannot easily contradict, even when you wrote every word in another window.

What a good record contains

For a substantial piece of work:

• Version history, on, from the first sentence.
• Your reading notes, in your own words, dated, with page numbers.
• The outline, as it changed. The old versions are the valuable ones because they show thinking.
• Your search history for sources: a Zotero library with items added over weeks is a beautiful record.
• Any conversations with an assistant that you would disclose. Keep the links or exports.
• The disclosure statement, written as you went.
• Your unaided-check log from Module 3, which shows capability developing independently of any tool.

For code: git, with commits at the end of every session, including the failing ones. A history with forty commits showing an idea developing, with dead ends in it, is far more persuasive than a repository containing one commit called final.

The failure that looks bad and is innocent

Worth naming so you can avoid it. Students who draft on paper, or in a notes app, or in an assistant’s window, and then type or paste the result into the submission document, produce a history with no development in it. They did the work honestly and the record shows nothing.

If you prefer drafting on paper, many people write better that way, photograph the pages, dated. Two minutes, and your record is intact.

Do not fake it

There are tools that simulate typing to fabricate a document history. Using one is a separate and more serious offence than whatever it was meant to conceal: it is deliberate fabrication of evidence, and institutions treat it accordingly. It is also detectable, because simulated typing has a suspiciously even rhythm and no deletions, revisions or long pauses.

A real writing history is messy. That messiness is the signal.

The record helps you before it ever defends you

The framing that makes this bearable: none of this is extra work done out of fear. Every item on that list is something that makes the work better.

Notes in your own words help you write. An outline that changed is an argument that developed. Commits at the end of each session let you go back when you break something. A disclosure written as you go is accurate. The unaided-check log tells you what you can actually do.

You are keeping a record because it is how good work is done. That it also happens to be unanswerable evidence is a side effect, and it is the reason to start in week one rather than the week you need it.

Retention

Keep everything until your degree is awarded and the appeal window has closed. Cases sometimes arrive months later, and a cloud document deleted in a tidy-up is gone. Export a copy of anything that matters at the end of each year: institutional accounts are closed after you graduate, sometimes with little notice.