FAQ on Pi-Calculus (2002) [pdf] | Hacker News Reader

FAQ on Pi-Calculus (2002) [pdf]

cs.cmu.edu

44 pointskushti5 comments

There has been tons of research on the pi-calculus. But very little has been lifted to the "real world". Maybe because the pi-calculus model isn't suited for programming (unlike the lambda calculus that led to Scheme, ML, Coq...).

I'd love to be proven wrong, but it seems to be a niche for a community of researchers to publish useless technical papers. At least, they should devote their energy to try to apply their results to solve real problems.

There's at least LuaPi: https://github.com/LuaDist/luapi/blob/master/doc/LuaPiTut.pd... - doesn't seem hard or "bad suited for programming".

I'd be wary of claiming that something isn't practical just because it hasn't escaped the academic world yet. Bringing research stuff to "real world" is a hard skill, and in my opinion of a very high value.

> ... they should devote their energy to try to apply their results to solve real problems.

You might instead have said this about any researcher in abstract mathematics, computer science, and even theoretical physics.

Some people enjoy working directly on "real problems", and that is fine. But plenty of other people enjoy working simply to push the frontier of our knowledge, and that's also fine.

pi-calculus and lambda calculus are both theoretical frameworks, engineered for clarity of notation and thought, in order to reason better about theoretical problems in computation. Neither of them were designed to be practical programming tools, and their merit does not rest on whether they happen to inspire such tools.

One thing that bothers me a little with the pi-calculus, is that unlike some abstract maths or theoretical CS, many researchers sell their results with potential real-life application arguments. After several decades of the same papers studying very similar type systems and equivalence with no application, one can legitimately starts to wonder if this is worth the effort.

There was also WS-CDL, a standard for choreography that was intended to improve upon many of the orchestration standards (and implicit centralised control within them). It was one of the few WS-* standards to come from W3C rather than OASIS. http://www.w3.org/TR/ws-cdl-10/

Reply on news.ycombinator.com