A laboratory notebook has rules that look, at first, like fussiness. It must be bound, not a folder of loose sheets. The pages must be numbered before anything is written on them. Every entry carries the date. Nothing may be erased — a mistake is struck through with one line, so that what was written underneath can still be read.
A typical bound notebook holds two hundred numbered pages, and a researcher who fills three a day works through one in a little over two months.
The reason for the rules is not tidiness. A notebook kept that way can be produced years later as evidence: in patent disputes over who had an idea first, in investigations when a result cannot be repeated, in any argument about what was actually done and when. Loose pages prove nothing. A page that has been rubbed out proves less than nothing.
The notebook is not there to help the scientist remember. It is there to make the work checkable — by a colleague, by a court, and above all by the writer on the morning after, when the reasoning that seemed obvious at midnight has to be followed step by step.
Working paper in an exam is a small version of the same idea. It is not a place to scratch numbers while you think. It is where the question you were actually answering stays visible, so that ten minutes later, when you come back to it, you do not have to reconstruct your own thinking from nothing.