Page:Sm all cc.pdf/118

 * Remember in note-taking that some ‘facts’ assumed during the experiment may later be rejected. Associated data therefore may be inaccurate or imprecise and need unanticipated corrections. Whether these corrections are possible probably will depend on completeness of the notes.
 * Flag any unexpected observations. Immediately consider whether they may indicate a problem with experimental procedure. Later, consider whether or not they can offer a new insight (Chapter 8).
 * In deciding whether or not to record something, remember how cheap paper is compared to the cost of redoing an experiment.
 * Similarly, paper is too cheap to warrant tiny, crowded marginal notations that might later be overlooked or found to be unreadable.


 * Keep and regularly update a backup of your most vital notes and data. Be prepared for the eventuality of losing notes, data listings, or data files. Take steps to postpone that eventuality: do not take prime data on airplanes as checked baggage; be cautious about carrying prime data around with you routinely; both lock and back up your personal computer; keep backups in a different room from originals.

14) protect your experimental setup, experiment, and samples from accidental damage by yourself or others.
 * Make a sign such as “Experiment in progress, do not disturb” and display it whenever the experiment is untended. I have seen incidents where using a seemingly overcautious sign could have prevented heartbreaking setbacks.
 * When leaving a shared computer while it is number-crunching or otherwise in use, put a sheet of paper saying “In use” over the screen or keyboard.

I know of incidents of janitors innocently discarding:
 * data or samples stored in a container that resembled a trash can;
 * delicate samples wrapped in Kleenex and left on a desk;
 * boxes that were almost empty.

15) avoid minor changes during the experiment. They separate data obtained before and after the change with a gulf of ambiguous comparison.