Intention Space
Cognitive Execution Paths Without Hidden Logic
A research framework exploring how perception and intention can become first-class computational constructs — eliminating hidden control flow through declarative, data-driven design.
Vision
Software today is built around functions and control flow.
Intention Space explores a different model — one where perception, intention, and execution become explicit computational primitives.
The goal is not another framework, but a different way of thinking about computation itself.
How it works
The runtime follows a three-stage structure:
1. Perception
The system observes signals and constructs a perceptual state instead of reacting directly to events.
2. Intention
Intentions are generated from perception, user actions, or AI reasoning.
They describe what should happen, not how to execute it.
3. Execution
The CPUX engine executes intentions through transparent execution paths, eliminating hidden control flow.
The Core Invariant
Traditional computation manipulates values through functions. While values may carry names in source code, those semantic identities are typically lost during execution and compilation.
Intention Space proposes a different invariant:
Semantic Identity Survives Execution
Traditional computing operates on values and functions. Although programmers give meaningful names to variables, those semantic identities are largely lost once execution begins.
For example:
customerBalance = 100
patientTemperature = 100
The measure is identical:
100
Yet the computations that make sense for each are completely different. The distinction comes from the semantic identity attached to the measure, not from the number itself.
Modern AI systems can infer these meanings from names and context, revealing something that software has always depended upon but rarely treats as a first-class computational entity.
Intention Space makes semantic identity a runtime invariant.
A fixed collection of Pulses forms the semantic basis of an Intention Space. Intentions are constructed from these Pulses and executed through CPUX (Common Paths of Understanding and Execution), Design Nodes, Signals, Objects, Fields and Field Boards.
Every execution step retains a semantic coordinate:
(CPUX, IC, DN, Signal)
rather than being identified only by memory addresses, call stacks or instruction pointers.
This enables Stability Intelligence — computation in which intent, context and execution locality remain visible, traceable and bounded throughout the lifetime of the system.
In Intention Space, semantic identity survives execution.
Pulses
A Pulse represents a fundamental intent coordinate within an Intention Space instance.
Pulses are not transient variables. They form the stable semantic basis from which higher-level Intentions are constructed.
Intentions
An Intention is a composition of Pulses describing a meaningful unit of behaviour.
Rather than computing solely on anonymous measures, computation occurs on measures associated with stable intent coordinates.
measure + intent = computable meaning
Design Nodes
Design Nodes (DNs) provide uniquely identifiable design-time scopes whose identities remain intact during execution.
Each DN participates in CPUX execution through designated Signal pairs makig an Intention Container, making every execution step semantically traceable.
Fields and Field Boards
A Field represents the active semantic state of an Intention Space CPUX.
Field Boards coordinate interactions between CPUX paths, allowing execution flows to trigger, observe and compose with one another while preserving semantic locality.
CPUX
Computation proceeds through CPUX (Common Paths of Understanding and Execution).
Every execution has an explicit locus coordinate:
(CPUX, IC, DN, Signal)
rather than only an instruction pointer or memory address.
This makes execution transparent, traceable and bounded.
Stability Intelligence
Every situation contains constraints, relationships and possible resolutions. Intelligence is the process of discovering and maintaining those relationships while moving the situation toward a stable state. The result is a model of computation in which semantic coordinates are treated as first-class entities.
Instead of hidden control flow operating on anonymous values, Intention Space enables execution within a stable semantic geometry where every computational step can be located, understood and reasoned about through its intent.
For example:
Customer has not paid invoice.
This situation already contains structure:
customer invoice amount due date payment status
The intelligence is not created by the software.
The intelligence already exists in the situation.
The software merely helps navigate it.