| In that vein, earliest mention of UUID I could find (thanks Apollo Archives): > 2.4 Unique Identifiers > An important aspect of NCA is its use of universal unique identifiers (UUIDs) as the most primi- > tive means of identifying NCA entities (e.g. objects, interfaces, operations). UUIDs are an exten- > sion of the unique identifiers (UIDs) already used throughout Apollo’s system [Leach 82]. Both > UIDs and UUIDs are fixed length identifiers that are guaranteed to refer to just one thing for all > time. The principal advantages of using any kind of unique identifiers over using string names at > the lowest level of the system include: small size, ease of embedding in data structures, location > transparency, and the ability to layer various naming strategies on top of the primitive naming > mechanism. Also, identifiers can be generated anywhere, without first having to contact some > other agent (e.g. a special server on the network, or a human representative of a company that > hands out identifiers). https://jim.rees.org/apollo-archive/papers/ncs.pdf The paper is not dated but probably was around 1986 or later? |