Path Integrals and Datom Streams: History Is Primary
Path integral intuition clarifies why event streams are such a strong base abstraction: histories come first, derived state comes later.
The Path Integral Perspective
In Feynman's formulation, an outcome amplitude is computed as a contribution from many possible histories, not a single privileged trajectory.
At a conceptual level:
Outcome = sum over possible historiesThe key shift is from object snapshots to temporal structure.
The Datom Perspective
A datom stream records realized history as immutable events:
d1 -> d2 -> d3 -> d4 -> ...State is not stored as a mutable singleton. It is interpreted from bounded history.
Terminology note: history prefix means the initial contiguous segment of the event log. This is a standard distributed-systems concept used to compare node histories.
state = interpreter(history-prefix)Where the Analogy Is Tight
- History is the primary object
- Observers/interpreters extract observables from history
- Causality constrains valid orderings
Where the Analogy Is Not Literal
Path integrals operate over possible histories with complex amplitudes. Datom systems store one realized history with explicit causal metadata.
So this is a structural analogy, not an equation of physical and computational mechanisms.
Implication for System Design
When history is primary, architecture becomes simpler:
- Append-only facts instead of mutable hidden state
- Deterministic replay and time travel
- Multiple interpreters over one substrate
This is exactly the Datom.world split between stream substrate and interpreter semantics.
Conclusion
A practical summary is: computation is interpretation of causal history.
That statement is compatible with event sourcing practice and with the path-integral style insistence that histories, not snapshots, are fundamental.
Related Reading: