The Periodic Stack of Data Systems
A data system is its coordinate.
Storage and retrieval is a structured design space. Ten axes classify any system that keeps records and answers questions about them. Position predicts behavior — what it answers quickly, what it guarantees, what survives, and what it costs in joules.
c = ⟨ W, S, T, C, D, Q, V, R, N, τ ⟩
The cell coordinate · ten axes, one tuple
The thesis
Position predicts behavior.
A periodic table works because where an element sits tells you how it behaves. The same structure holds for data systems. Each axis is a finite, independent dimension of design. A system occupies one position on each; the tuple of those positions is what the system is.
Read across the coordinate and the system explains itself — the questions it answers for free, the ones it answers at the Landauer floor, the guarantees that travel with each answer, what survives a failure, and the joules each operation costs against its physical minimum.
The stack is a way of seeing, not a verdict. It classifies what exists, makes the structure of the design space legible, and lets a system be reasoned about before it is built — the same move the Periodic Stacks of Computation and Information make for their domains.
The design space
Ten orthogonal axes.
Five inherit from textbook database theory. Three — T, V, R — are shared with the sibling OpenIE stacks. Two close the distribution and temporal dimensions.
Workload zone
01 · WHAT IS ASKEDThe kind of question put to the system.
Storage primitive
02 · HOW RECORDS LIEThe physical organization of records.
Thermodynamic class
03 · LANDAUER FLOORThe cost floor of the operation, inherited from the Periodic Stack of Computation.
Consistency model
04 · WHAT A READER SEESWhat a reader is guaranteed to observe relative to concurrent writers.
Durability class
05 · WHAT SURVIVESWhich failure the data is promised to outlive.
Query interface
06 · HOW IT IS ASKEDThe shape in which the caller expresses the question.
Verifiability
07 · WHAT TRAVELS WITH ITWhat kind of trust accompanies the answer.
Encoding regime
08 · WHAT RECORDS MEANWhat the records are for.
Distribution topology
09 · HOW IT SPANS MACHINESHow the system is laid out across machines.
Temporal model
10 · HOW HISTORY IS KEPTHow time is represented in the records themselves.
The coordinate space
The addressable map.
The product of the axis cardinalities is the discrete coordinate space. Most of it is structurally constrained; what remains is the map on which every data system has a place.
5 · 12 · 4 · 7 · 5 · 7 · 5 · 4 · 6 · 5 = 35,280,000
Discrete cell count · the addressable map of data-system space
discrete cells
orthogonal axes
compute primitives below
Companion stacks
One method, applied across fields.
The same axes recur. T inherits from the compute stack. V and R are shared with the information and synthesis stacks. A primitive has consistent coordinates across all of them.
Compute
258 primitives, 33 families.
Synthesis
The seven-axis synthesizer stack.
State
The nine-axis state stack.
Communication
The eleven-axis channel stack.
Memory
The eleven-axis retention stack.
Knowledge
Seven axes · the object the others handle.
Agent
Four axes · the actor and selection.
Stack
The overview · three names, one substrate.
Verify
Seven axes · joules per verified output.
OpenIE
The parent thesis.