Domain-Driven Design

Evidence (5)

• 5Hacker News·7/14/2026research
  DSLs Help LLMs Produce Reliable Code

  The article argues that large language models are most reliable when constrained by clear abstractions and domain-specific languages, rather than asked to implement everything from a broad natural-language spec. It uses Tickloom as an example of building and then using a DSL for distributed system behavior, with LLMs acting both as a design partner and a natural-language interface.

• 6Martin Fowler·6/16/2026research
  Fowler on LLMs, DDD, and AI’s impact on programming

  Martin Fowler highlights Prag Dave Thomas’s view that programming with LLMs can be more enjoyable by reducing drudgery, speeding feedback, reviving old projects, and helping exploration. He also reflects on DDD Europe talks about managing LLM context windows and conversation modes, and shares Charity Majors’s warning that AI adoption creates both real productivity gains and real risks to code quality and organizational knowledge.

• 4Martin Fowler·5/12/2026research
  What Code Means in the Age of LLMs

  The article argues that code has two roles: executable instructions for machines and a conceptual model that captures a domain's vocabulary and design. It says LLMs can commoditize code generation, but the more valuable part remains the shared concepts, boundaries, and meanings encoded in software. The piece also highlights the risk of cognitive debt when teams rely on generated code without preserving understanding.

• 6Martin Fowler·5/5/2026open_source
  Lattice Brings Structured Patterns to AI Coding

  Rahul Garg has released Lattice, an open-source framework designed to reduce friction in AI-assisted programming by organizing reusable engineering patterns into atoms, molecules, and refiners. It also maintains a living project context in a .lattice folder so AI tools can adapt to a team’s standards, decisions, and review history over time. The post also notes updates to an SPDD article and reflects on tools for conversation logs and the risks of AI-generated misinformation in Google Search.

• 4Martin Fowler·3/3/2026research
  Design-First Collaboration for AI Coding

  The article argues that AI coding assistants often skip the design phase and generate implementation immediately, which hides important architecture decisions inside the code. It proposes a structured, whiteboard-like conversation pattern to align on design before coding, reducing review overhead and misunderstandings.