KDS is a proprietary, deterministic modernization capability built on an analysis engine that has no peer, no competitor, and no equivalent in industry. KDS exists for one purpose: to restore truth, structure, and engineering certainty to legacy systems that were never supposed to fail.
For decades, organizations have been forced to choose between two bad options: manual reverse engineering or probabilistic AI inference. Both introduce ambiguity, risk, and drift. KDS eliminates all three. It reconstructs system knowledge directly from the binary with forensic precision, producing certification-grade engineering documentation that is complete, auditable, and traceable.
This repository is not a software distribution. It is the public-facing doctrine of a capability built and proven under rigorous, disciplined engineering standards.
KDS is a deterministic reconstruction service for aerospace, medical, defense, industrial, and safety-critical domains. It transforms raw assembly into:
-
System Requirements Specifications (SRS)
-
Software Design Documents (SDD)
-
Data Flow Diagrams (DFDs)
-
Memory maps and interface definitions
-
Traceability matrices
-
Verification and validation artifacts
Every output is derived directly from Assembly. No inference. No guessing. No hallucination. No tribal knowledge required.
Every programming language, no matter how it's written -- 1960s COBOL, 1980s Mil-Std-1750A, modern C++ -- converges into one objective reality: machine code. KDS targets that layer directly, reading the signal of actual execution. The result is a reconstruction that stands entirely on its own: independent of the original source language, the original developers, or any existing documentation.
KDS performs millions of logical checks per second, turning 5,000 lines of assembly into a complete, navigable model in under a minute. Because the engine resolves the entire model simultaneously rather than building it line by line, a lesson learned this morning can be reflected in a corrected model that same afternoon -- real iteration, at a pace and scale this kind of reconstruction has never run at before.
KDS traces the life of a variable: how data enters through a port, moves through an accumulator, lands in a register bank, and influences a control branch. Every call the binary makes into its own compiler or runtime library is identified and drawn as its own named entity, standing alongside the application's own logic. Every entity and every relationship in the model traces back to a specific instruction sequence in the source binary.
The resulting Data Flow Diagram is 100% compliant with the Yourdon/DeMarco methodology, produced at a scale few have ever attempted. The CASE environment adds a further layer automatically: hover-text abstracts on every library call, giving readers full context on what a routine does without ever leaving the diagram.
Every reconstruction is held to one rigorous standard: every connection placed correctly, every route resolved cleanly. Each determination is backed by direct evidence from the engine's own execution logs -- exact candidate counts, exact reservation histories -- so the result is provably correct. The same discipline that produces the model also audits it, page by page, confirming its accuracy before it's called complete.
One representative reconstruction, publicly referenceable in this repository, illustrates the scale KDS operates at:
-
Source: 4,969,515 lines of ARM64 assembly (~5 million LOC)
-
Entities extracted: 112,480
-
Hierarchy depth: 3,553 levels
-
System complexity: 4,852 total (average node complexity 3.12 across 1,006 top-level entities; max node complexity 30)
-
Output: a 3,883-page Data Flow Diagram, fully cross-referenced page to page, published in
3880 Page DFD ARM64 ASM/
Every entity and every relationship in this diagram traces back to a specific instruction sequence in the source binary. No inference. No sampling. No reconstruction shortcuts.
Note: GitHub's own folder browser truncates directory listings above roughly 1,000 files, so not all 3,883 pages will appear when browsing that folder directly on GitHub. Every file is present in the repository regardless -- clone the repository, or link to a specific page directly (e.g. .../3880%20Page%20DFD%20ARM64%20ASM/Page_1500.pdf), to reach any page not shown in the truncated listing.
This repository contains the conceptual, philosophical, and methodological foundation of KDS:
Founder_Manifesto.md – The doctrine and worldview behind deterministic engineering.
Capabilities_Statement.md – A formal articulation of what KDS delivers.
Methodology.md – The high-level process that governs deterministic reconstruction.
CFP_Abstract.md – The conference-ready abstract introducing the KDS paradigm.
examples/ – Fictional but representative artifacts demonstrating output structure.
LICENSE.md – A proprietary license protecting all intellectual property.
These documents exist to inform, not to enable replication.
KDS is offered as a confidential, high-assurance modernization service. Engagements typically include:
1. Binary intake and scoping
2. Deterministic reconstruction of system behavior
3. Generation of engineering documentation
4. Traceability and verification mapping
5. Delivery of a complete, auditable engineering package
KDS serves organizations responsible for systems where failure is not an option:
-
Aerospace and avionics
-
Medical devices
-
Defense and national security systems
-
Industrial control systems
-
Critical infrastructure
If your system is undocumented, partially documented, or too risky to modify, KDS restores the engineering truth required to move forward.
All content in this repository is proprietary and protected. No rights are granted to copy, modify, distribute, or derive from any material herein. See LICENSE.md for full terms.
KDS is a sovereign, deterministic modernization capability. This repository is its public doctrine—not its implementation.
If a capability like this is relevant to what your organization is building, securing, or trying to understand, I'd welcome hearing from you. Reach out to talk about the technology, ask questions, or explore where a conversation might lead. KDS operates globally and supports classified and unclassified environments.
Steve Whiseant Knowledge Discovery Systems