ABOUT

The Story Behind
Zero Point Logic

Built by one person, verified by 2.64 billion computations

THE CREATOR
C
Ciciu Alexandru-Costinel
RESEARCHER & DEVELOPER
Romania 🇷🇴
Mathematics Boolean Logic API Design Open Science Python Game Dev

How This Started

Zero Point Logic began as a personal research question: can a deterministic boolean system — one with no randomness and no training — exhibit genuine equilibrium properties? Not approximated balance, not statistical averaging, but a mathematically provable convergence to exactly 0.5 output frequency regardless of what you feed it.

The hypothesis seemed almost contradictory. Deterministic systems are fixed. Their outputs follow rules. How could a fixed rule produce equilibrium across every possible input distribution, from 1% ones to 99% ones? This contradiction was the seed of the entire project.

What followed was months of Python experiments, failed architectures, and eventually the discovery of the AIN (Absolute Input Neutralization) property — a structural characteristic of certain boolean matrix configurations that forces output equilibrium. The 8N+3 theorem emerged from trying to understand why. The 2.64 billion verified computations emerged from trying to prove it rigorously.

This was published as open science: paper on ArXiv, dataset on Zenodo, clients on PyPI/npm — because results that can be independently verified should be independently available.

2.64B
COMPUTATIONS
86,016
CONFIGS TESTED
100%
AIN VERIFIED
12
LANGUAGES
HERITAGE
Romanian Scientific Tradition
ZPL stands on the shoulders of extraordinary Romanian mathematicians who redefined the boundaries of logic and computation. Their work lives in ZPL's foundations.
🧮
Grigore Moisil
1906 – 1973
Founder of Romanian computer science and pioneer of many-valued logic — the idea that truth is not just binary (true/false) but can exist in degrees. Moisil algebras directly shaped algebraic logic. His work on switching circuits was decades ahead of practice. ZPL inherits his conviction that boolean structures can express richer truths than 0 or 1.
MANY-VALUED LOGIC
🧬
Gheorghe Paun
1950 – present
Creator of membrane computing (P systems) — a model of distributed computation inspired by the structure of biological cells. Paun showed that non-standard computational architectures can achieve remarkable properties, including universality and efficiency beyond conventional Turing models. ZPL's layered matrix architecture shares this spirit of biologically-inspired structural design.
MEMBRANE COMPUTING
🌀
Florentin Smarandache
1954 – present
Founder of neutrosophic logic, which extends classical and fuzzy logic by introducing a third component: indeterminacy. Truth, falsehood, and the space between. ZPL's AIN property achieves something related through entirely different means: the system does not compute indeterminacy — it eliminates bias entirely, arriving at a neutral equilibrium through deterministic structure.
NEUTROSOPHIC LOGIC
JOURNEY
The Road to 2.64 Billion
Every significant result has a backstory. Here is how ZPL went from a single question to a published, deployed, commercially available system.
PHASE 1
The Idea
The central hypothesis: can a boolean matrix system reliably converge to 0.5 output frequency regardless of input bias? The starting intuition was that if a system is structured correctly — not just randomly — certain symmetry properties might force equilibrium. No existing literature described this for deterministic systems without statistical mechanisms.
PHASE 2
First Tests
Initial Python experiments using small N values (N=3, N=5). The first time a configuration produced 0.4998 output ratio from a 10% input bias was striking enough to keep going. Further tests revealed the AIN (Absolute Input Neutralization) signature: output hovering at exactly 50% across wildly different input distributions. The observation was surprising enough to demand rigorous verification.
PHASE 3
Scaling Up
Systematic expansion: 86,016 unique matrix configurations tested at 11 bias levels each, 1,000 samples per configuration per bias. This produced the 2.64 billion computation dataset. The Railway API was deployed to make the system accessible and to validate it under real-world HTTP conditions. Results were consistent: AIN configurations maintained equilibrium without exception.
PHASE 4
Publication
The findings were written up and submitted to ArXiv as an open-access preprint. The full 2.64 billion computation dataset was deposited on Zenodo under CC BY 4.0, making it permanently citable and independently reproducible. Any researcher can download the CSV files, re-run the analysis, and verify the results from scratch. Open science was a deliberate choice, not an afterthought.
PHASE 5
Commercialization
Client libraries published on PyPI (zpl-engine), npm (zpl-engine), and as a Unity package (com.zplengine.client). RapidAPI listing for marketplace discovery. This website launched as the primary hub with documentation, live demo, member portal, and subscription tiers. The algorithm engine remains proprietary; the clients are MIT-licensed open source.
TRANSPARENCY
Open Science Commitment
Science should be verifiable. Every empirical claim in ZPL's publications can be independently checked using publicly available data and tools.
📊
Full Dataset
The 2.64 billion computation dataset is publicly archived on Zenodo. 86,016 configurations, 11 bias levels, raw CSV format — fully downloadable.
CC BY 4.0
📝
Open Access Paper
The ArXiv preprint is freely readable, citable, and permanent. No paywall, no publisher embargo. The full methodology is documented.
OPEN ACCESS
📦
Open Source Clients
PyPI, npm, and Unity client libraries are all MIT-licensed. You can read every line, modify them, and contribute to them freely.
MIT LICENSE
🔒
Protected Algorithm
The core ZPL engine algorithm is proprietary intellectual property. The what and why are published; the how is the foundation of the commercial service.
PROPRIETARY

Get in Touch

Research questions, commercial inquiries, collaboration proposals, or just curiosity — all welcome. Direct email, always read personally.

✉ cicic.alexandru@gmail.com