>concurrency models We started out with OS threads (I guess processes came first but whatever) and now we're trying to figure out what the next paradigm should

>concurrency models We started out with OS threads (I guess processes came first but whatever) and now we're trying to figure out what the next paradigm should be. It looks to me like it's Hoare (channels, etc) for systems programming and actors for distributed systems, both really really old ideas. To be fair there are other ideas (STM, futures, etc) that fill their own niches, but they either specialize on a smaller problem (futures) or they're still not quite ready for popular adoption (STM). If this is cyclical then I think we're pretty early in the first cycle. Sure, the spotlight moves from one model to the other and back, but that's because the hype train cannot focus on many things at the same time, not because the ideas go out of style.