ABSTRACT
The inception of one of the very first distributed algorithms, published in 1977 - IFIP Congress 19771, is one of the most exciting moments in my scientific carrier. The discovery of synchronization problems in networks where two distant processes shall interact correctly through time varying and lossy channels prompted me to look at the general problems arising in systems where n processes need to coordinate their behaviours, despite differing individual views of “current system state” and no central locus of control. While the protocol itself died a painful and probably well-deserved death, the influence of many of its design ideas have been underreported, partially because they are considered “Internet” and we might be better off had we picked up one or two more of them. Unfortunately, the Internet history is also tainted by mistaken articles and biased interviews produced by improvised “experts”.
