OBSERVABILITY Signal 486
Co-author of Lamport's anti-types specification paper reflects on its contentious publication and how typed verification prevailed
↗
Illustration only Photo by Igor Saikin on Unsplash
A retrospective blog post recounts how the author came to co-author Leslie Lamport's controversial paper arguing against types in specification languages, a paper born from a rejected TOPLAS submission and eventually published with a disclaimer.
The post offers a rare behind-the-scenes look at academic peer review politics and the evolution of type systems in formal verification. It also provides a candid assessment that Lamport's anti-type thesis has not aged well, given the industrial-scale successes of typed verification in CompCert, seL4, and Amazon's Nitro Isolation Engine.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
Lamport submitted a note arguing specification languages should use untyped set-theoretic formalisms rather than typed ones, and both TOPLAS referees rejected it.
Editor Andrew Appel persuaded one rejecting referee to co-author a revised version with Lamport, but a subsequent editor sent the revised paper to fresh referees who also rejected it.
The author concludes that 27 years later, typed systems have proven their worth in industrial-scale verification while untyped formalisms still struggle with practical issues like notation overloading and the freedom to write anything.
THE CLUSTER