Axiom One primitives

SkipSpace

Large constrained search spaces are often implicit: materialising the tree or retaining every explored prefix exceeds embedded memory budgets.

A heap-free traversal primitive represents each prefix with compact sufficient state, counts valid continuations using bounded saturating arithmetic, and lets the caller visit or skip an entire subtree without materialising the search tree.

DevelopmentMITC99
GitHubQuick Start
Last verified
2026-08-28
Platform scope
Portable C99; caller-owned state and no dynamic allocation

What it solves

Large constrained search spaces are often implicit: materialising the tree or retaining every explored prefix exceeds embedded memory budgets.

How it works

1Prefix→2sufficient state→3bounded subtree count→4visit or skip subtree

Use it when

Traverse a modeled constrained space with fixed caller-owned workspace
Count a subtree before deciding whether to visit or skip it
Keep navigation state compact when equivalent prefixes have equivalent futures
Use sparse skip checks in a large deterministic exploration

Quick Start

#include "skipspace.h"

/* Configure the modeled transition table and caller-owned workspace. */
/* push() updates sufficient state; count()/can_skip() decide a subtree. */
/* No search tree is allocated or materialised. */

Engineering evidence

  • The portable C core has no Arduino or ESP-IDF dependency and no dynamic allocation
  • V1 defines navigation, bounded/saturating count, can_skip and workspace sizing, with depth bounded at 1000
  • An ESP32-S3 uint32_t, 31-state, K=32 profile is documented with approximately 227,834 visited nodes/s using sparse skip checks

Known limits

  • It is not a general solver, AI framework, vector database or universal search engine
  • The modeled transition table must preserve future-relevant behaviour in its sufficient state
  • count/can_skip cost substantially more than push/pop; V1 does not put rank/unrank on the hot path