I think it was back in the 80's when I was in a networking standards group (was it ANSI?? It's been a long time) and we were trying to define programming layers on top of sockets for UNIX. The effort completely failed - but the write-up about what we were trying to accomplish seemed to live on - getting revised and revised until it became the OSI model - with all its layers that we sort-of tried to implement. Of course, the new drivers and software layers didn't actually work, and the performance was abysmal, so the effort was abandoned. But the press and everyone else really liked the understandable text write-up about networking, which almost nobody understood at the time, so the write-up seemed to get a life of its own.

Since then I've been scorned at job interviews and among peers for not having the model memorized, or not being able to explain the meaning of various layers. Just another example of our crazy methods of screening job candidates. Answering stupid questions from stupid people means more than actual experience.