Facts and Opinions
The other concern is over where facts end and opinion begins. There are hundreds of programming languages, dozens of development paradigms and more than one philosophy on project management. The academic world is riddled with its own challenges stemming from its distance from business concerns and practicality.
For Cardano, we first attempted to capture obvious deficiencies that can be universally agreed to be useful from an engineering perspective. For example, cryptography and distributed systems are both extraordinarily involved topics with far too many examples of how naive hands can make horrific mistakes. Therefore, any protocol requiring insight from these domains needs to be designed by an acknowledged expert and be submitted for review by other experts.
Ouroboros is our first case study of this area. It was designed by a team of cryptographers with a large, diverse and publicly verifiable publication history. It was built according to the standard cryptography process, with security assumptions, an adversarial model and proofs. These proofs were checked by submission to conferences20 and also independently by computer proofs written in Isabelle by a team at the University of Cambridge21.
Yet this work alone provides no guarantees of usefulness — just a rigorous check of a security model given some assumptions. For usefulness, one needs to implement and test the protocol. Our developers have done so in both Haskell and also Rust. This work revealed that more effort needed to be focused on the synchronization model, which led to the creation of Ouroboros Praos.
This art of iteration is what produces great protocols, with each step leading to new lessons and a requirement to re-verify the correctness of prior step22. It is costly, time consuming, and at times truly tedious, yet it is required to ensure a protocol is correctly designed.
Protocols — especially ones to be used by billions of people — are not short lived and rapidly evolving. Rather they are intended to be followed for years to decades. It seems entirely reasonable that, prior to burdening the world with a new financial system we all have to live with for the next 100 years, we want to demand some tedium and rigor from its designers.
20: Accepted Paper Number 71 of the IACR's Annual Crypto Conference in California.
22: Following a tangent for a stake of levity, one should watch Professor Halmo's discussion about how to write a math textbook.