Engineering Foundation
OpenLogos isn’t invented from scratch. It’s built on 40+ years of battle-tested software engineering theory — carefully recompiled for an era where AI writes the code.
Six pillars
Section titled “Six pillars”| Theory | Origin | Maps to in OpenLogos |
|---|---|---|
| BDD — Behavior-Driven Development | 2006, Dan North | Phase 1 requirements, GIVEN/WHEN/THEN acceptance criteria, scenario structure |
| TDD — Test-Driven Development | 2003, Kent Beck | Phase 3 Step 3 test-first design, openlogos verify |
| DDD — Domain-Driven Design | 2003, Eric Evans | Phase 3 Step 0 architecture, scenario modeling, ubiquitous language |
| Stage-Gate Process | 1986, Robert Cooper | Three-layer progression gates, Gate 3.5 acceptance report |
| Docs-as-Code | 2014, Write the Docs | Documents as Context, logos/ directory, version-controlled artifacts |
| Executable Methodology | AI-native pattern | AI Skills (SKILL.md), AGENTS.md auto-detection |
Why theory matters
Section titled “Why theory matters”Predictability
Section titled “Predictability”These practices have been validated across millions of projects over decades. When applied to AI-assisted development, they produce predictable results — not random walks.
Composability
Section titled “Composability”BDD + TDD + DDD + Stage-Gate compose naturally:
- BDD defines the scenarios
- TDD verifies them
- DDD models them
- Stage-Gate ensures quality between phases
OpenLogos wires them together into a single coherent workflow.
Teachability
Section titled “Teachability”AI agents learn from established patterns far more effectively than from ad-hoc instructions. When you write “use GIVEN/WHEN/THEN,” AI understands the pattern because it appears in millions of training samples.
Addressing common concerns
Section titled “Addressing common concerns”“Isn’t this just waterfall?” No. Waterfall sequences all requirements before any design before any code. OpenLogos sequences per scenario. S01 can be in Phase 3 while S04 is still in Phase 1.
“AI is supposed to be fast. This adds process.” A 5-minute prompt that produces garbage you debug for 3 hours isn’t fast. A 30-minute structured process that produces correct, tested, traceable code in one shot is fast. Speed is measured in shipped features, not generated tokens.
“This is overengineering for small projects.” OpenLogos scales down gracefully. A solo project might skip the HTML prototype phase. A weekend hack might compress Phase 1 + 2 into one doc. The methodology is a toolkit — use what you need.
“Why not just use Agile?” Agile assumes human developers who can hold context in their heads. AI doesn’t have a head — it needs context written down. OpenLogos is Agile-compatible but context-explicit, because that’s what AI requires.
See also: Interactive deep dive — Engineering Foundation →