Research#
EB is a discrete event simulation language. It exists to demonstrate the usefulness of “event segment” theory – a concept that we hope is useful to creators of discrete event simulators. The concept unifies periodic, iterative, event-driven and agent-driven models by providing the minimal computation unit required and a means to map each into the scheme.
Side benefit of the research#
This document (Building a DSL with Claude) is one early result.
This work is useful to engineers and scientists who would benefit from having a domain specific language (DSL) but have no experience building programming languages. They would benefit from a DSL tailored to their domain—if only creating one weren’t so difficult.
These users need more than one-off LLM prompting. Ad-hoc requests produce inconsistent results — the LLM reinvents the wheel each time. A DSL provides structure:
A defined language that the LLM can learn deeply
Documentation the LLM can reference for accurate assistance
Consistent patterns for debugging and generating first drafts
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ [ DSL Development & Use ] ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ ┃
┃ ┃
┃ ┃
┃ ┃
┃ ┃
┃ ┌─────────────────┐ ┌─────────────────┐ ┌───────────────────┐ ┃
┃ Person │ │ │ Compiler and │ │ Model creation │ ┃
┃ ──────▶│ Specification │───────▶│ documentation │───────▶│ and use │ ┃
┃ │ │ │ │ │ │ ┃
┃ └─────────────────┘ └─────────────────┘ └───────────────────┘ ┃
┃ ┃
┃ ┃
┃ iteration ┃
┃ ◀ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ (Person / LLM) ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ─ ▶ ┃
┃ ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛