Perspectives
What Computation Really Is
Computation formally transforms representations according to rules, but a physical system instantiates a computation only under a mapping between physical states and computational states. Different computational models…
The easiest way to misunderstand computation is to begin with a metaphor and never return to the mechanism. Computation formally transforms representations according to rules, but a physical system instantiates a computation only under a mapping between physical states and computational states. Different computational models choose different abstractions.
For me, the useful sequence is the opposite: observe the phenomenon, identify what changes state, locate the constraints, and only then borrow language from engineering. A metaphor should reduce cognitive load; it should not silently replace the thing being explained.
Start with the mechanism
Computation formally transforms representations according to rules, but a physical system instantiates a computation only under a mapping between physical states and computational states. Different computational models choose different abstractions.
A mechanism-first explanation asks what physically carries the effect, what can vary, what is conserved, which feedbacks exist and how an intervention would change the outcome. This is the same discipline that keeps a production incident from turning into random log-reading. The difference is that nature has no obligation to expose a convenient API.
What an architect notices
Software architecture teaches that the same computation can run on different hardware and the same hardware can realize many abstractions. That flexibility is why computational metaphors travel so easily.
The comparison is valuable because it generates questions: where is state, how is it propagated, which processes are local, where are delays, what resources are scarce, and what conditions make the system leave a viable region? Those questions are portable even when the implementation is radically different from software.
Where the shortcut breaks
If every physical evolution can be called computation under some mapping, the claim risks becoming too broad to explain anything. A weather simulation computes a model of a storm; the storm obeys atmospheric physics.
The failure of the analogy is part of the explanation. It tells us which assumptions came from our engineering culture rather than from the phenomenon itself. In natural systems, history, material embodiment and environment are often not external concerns; they are part of the mechanism.
Scale changes the answer
At one scale we can talk about components. At another, interactions become the useful objects. Move farther out and population, tissue, institution or planet-level patterns appear. Good explanations do not insist that one scale is the only real one; they connect the scales without pretending the connection is trivial.
Why this matters
The crucial philosophical question is not whether the universe can be described computationally, but what that description lets us predict or explain.
What evidence would distinguish 'can be modeled as computation' from 'is computing'?
That is where the subject becomes more than a scientific fact. It becomes a way to think about systems whose organization was not designed for our convenience.
Reading trail
These links are starting points for the scientific and historical ideas. The systems interpretation, analogies and conclusions here are my own.