Removing names and addresses from a dataset may protect privacy when the remaining information is ordinary and broadly shared. In a detailed location or medical dataset, however, combinations of dates, rare conditions, and travel patterns can reidentify people when linked with outside records. 'Anonymous' is therefore not a permanent property of data; it is a judgment about likely attackers, available auxiliary information, and future use. Sensitive projects should minimize collection, restrict access, and reassess risk instead of relying on one act of de-identification.