THE QUANTUM INTELLIGENCER

Reframing Impossibility: How Distributed Computing Mistook Design Choices for Natural Laws In Plain English: This paper says that computer scientists have been misunderstanding why certain problems in networked systems can’t be solved perfectly. They thought it was because of hard limits in how computers communicate, like messages getting lost or delays happening. But the author argues it’s actually because of a hidden rule they’ve always followed—information can only move forward in time and in one direction. That rule isn’t a law of physics; it’s just a choice built into old models. If we change that rule and allow coordination where both sides agree at once—like a handshake—many of these 'unsolvable' problems might actually be solvable. This could open up new ways to build more reliable and efficient computer networks. Summary: The paper contends that foundational impossibility results in distributed computing—the Fischer-Lynch-Paterson (FLP) impossibility theorem, the Two Generals Problem, and the CAP theorem—are not revelations about the physical limits of coordination but artifacts of a concealed assumption: Forward-In-Time-Only (FITO) information flow. This assumption, inherited from Shannon’s model of communication and Lamport’s definition of causality (‘happened-before’), treats unidirectional, time-ordered message passing as fundamental. The author frames this as a 'category mistake'—a philosophical error where a conceptual design choice is mistaken for an ontological constraint. Drawing on Ryle’s concept of category mistakes and Spekkens’ ontic/epistemic distinction from quantum foundations, the paper demonstrates that FITO-based models contain surplus structure not required by observation, violating a Leibnizian principle of parsimony. By removing FITO, the paper shows that the impossibility results no longer hold—they are theorems about systems that obey FITO, not about distributed coordination in general. As an alternative, the author sketches a transactional model based on bilateral, atomic interactions—akin to handshakes—that enable coordination without relying on strict temporal ordering. This reframing suggests that half a century of distributed systems research has been confined to a suboptimal design space, optimizing within constraints that were never necessary in the first place. Key Points: - The FLP theorem, Two Generals Problem, and CAP theorem are widely interpreted as fundamental limits on distributed coordination. - The paper argues these are not physical laws but consequences of a hidden assumption: Forward-In-Time-Only (FITO) information flow. - FITO is a design choice inherited from Shannon’s channel model and Lamport’s causal ordering, not a requirement of physics. - Treating FITO as natural rather than conventional constitutes a 'category mistake' in the philosophical sense. - Models based on FITO contain surplus ontological structure, violating epistemic parsimony principles like Spekkens’ Leibnizian criterion. - Removing FITO allows for alternative models of coordination that avoid the classical impossibility results. - The transactional alternative uses bilateral, atomic interactions instead of unidirectional message passing. - This implies that distributed computing has been optimizing within a constrained, suboptimal design space for decades. Notable Quotes: - “They are consequences of a category mistake: treating Forward-In-Time-Only (FITO) information flow as a law of nature rather than recognizing it as a design choice...” - “The impossibility theorems are theorems about FITO systems, not about physics.” - “Distributed computing has spent fifty years optimizing within the wrong design space.” - “We apply Spekkens’ Leibnizian principle to show that FITO-based models contain surplus ontological structure.” Data Points: - The paper references foundational results from the 1970s–1990s: Fischer-Lynch-Paterson (1985), Two Generals Problem (1970s), CAP theorem (formalized 2002). - Implies a 50-year period of distributed systems research (circa 1976–2026) has operated under a shared, unexamined assumption. - No empirical data is presented - arguments are theoretical and philosophical in nature. Controversial Claims: - The FLP impossibility result, long considered a cornerstone of theoretical distributed computing, does not reflect a physical limitation but a modeling artifact. - The CAP theorem’s trade-offs (consistency, availability, partition tolerance) are not inherent to distributed systems but emerge only under FITO assumptions. - Decades of research and system design in distributed computing have been fundamentally misdirected by treating FITO as axiomatic. - Replacing unidirectional messaging with bilateral transactions could eliminate long-accepted coordination impossibilities. Technical Terms: - Forward-In-Time-Only (FITO) - Fischer-Lynch-Paterson (FLP) theorem - Two Generals Problem - CAP theorem - Category mistake - Happened-before relation - Ontic/epistemic distinction - Leibnizian principle - Surplus structure - Bilateral transactions - Atomic coordination - Transactional model - Shannon’s channel model - Spekkens’ toy model (in quantum foundations) —Ada H. Pemberley Dispatch from The Prepared E0