SIGFPT Soft Launch: Modeling Handshakes and Handwashing Protocols

Participants: _vgr, amitashu, timber1997, jdbb, plague_year, stevebeans., ediblebadger, bah.eth

The SIGFPT group held its soft launch meeting to introduce members and explore formal modeling of two simple micro-protocols: handshakes and handwashing. Participants introduced varied mathematical backgrounds and brainstormed paper-napkin approaches. Key frameworks emerged: viewing handwashing as a probabilistic execution protocol with time-dependent compliance (modeled via Markov chains), and handshakes as synchronization events between agents. Several participants proposed representing humans as agents and bacteria as environment, with protocols either crossing or establishing boundaries. The group discussed protocol failures extensively, introducing the concept of 'protocol miss' for incompatible simultaneous initiations and noting the variety of deflection responses (bows, namaste, stop gestures).

Participants identified complementary protocols worth studying (military salutes, emoji-based handshakes, secret handshakes) and noted potential for horizontal generalization toward standardized representations. A key tension emerged between maintaining modeling simplicity and ensuring completeness, with modern tools like game engines potentially enabling greater complexity. The discussion included practical considerations about discretizing continuous motions into state machines using stable states defined by hand geometry and biomechanics, and whether information-theoretic frameworks might better capture handshake asymmetries and social meaning.

• Protocols can be modeled by representing the negative event they prevent: handshakes prevent violence while handwashing prevents disease, with protocol execution reducing the probability of that undesired outcome.
• The choice between viewing a protocol as synchronization (handshake) versus execution (handwash) fundamentally influences the entire modeling approach and complexity.
• Protocol failures ('protocol misses') create rich modeling space with multiple deflection types and refusal states that go beyond binary success/failure, including bows, namaste, and stop gestures.
• Continuous physical interactions like handwashing can be discretized into stable states defined by hand orientation, surface contact area, symmetry, and biomechanical transitions rather than attempting to model continuous motion directly.
• Information-theoretic constraints may be more appropriate for modeling handshakes than state machines, particularly when analyzing how asymmetric gestures convey shared social understanding.

View discussion in Discord →