Lambda the Ultimate

Back in 2003, Martin Berger conducted a fairly lengthy interview with Robin Milner. The transcript includes some interesting tidbits on the development of ML, CCS, and the pi-calculus. Among other things, you'll find a recounting of how Milner and David Park came up with the idea of bisimulation, a discussion of the rationale behind some of the design decisions Milner and his colleagues made in creating the pi-calculus, and Milner's thoughts on how theory should influence programming languages:

I do care that languages should be informed by theories...I actually think the best way forward for us now is to look at concurrent calculi as modelling theories for modelling interactions, whether they occur in programs or in outside programs...Languages should emerge from that. They should be treated as a part of a modelling theory. Up to now I don't think we had sufficient incentive to make sure that our languages are close to scientific models. It's only with the onset of computation as a global phenomenon that modelling those interactions becomes so scientifically important that it is bound to have its effect on programming languages.