A term from the interface between linguistics and computer science. Specifically, it is about which proper names (company names, persons) a computer program can recognize in a text and extract from it (information extraction).
For example, the name Christoph can be spelled in many different ways, but it always sounds the same: Christof, Kristof, Kristoff, Krzysztof. And 'Apple' can be an apple, the computer company or Apple Records the label of the Beatles. With a first name, the computer hardly stands a chance; with a company name, the 'correct' term results from the context or domain.