Mount the framework as a submodule at framework/ (v0.4.0), add AGENTS.md and the installed skill copies for Codex CLI (.agents/skills) and Claude Code (.claude/skills).
32 lines
1.6 KiB
Markdown
32 lines
1.6 KiB
Markdown
# Domain Dictionary (DICT)
|
|
|
|
Cite in `CrossReference` only if the instance exists (`new-artifact.sh` checks this for you):
|
|
|
|
- **BC** — The business goals the vocabulary serves
|
|
- **SA** — The Product Owner (and other business stakeholders) whose terms are recorded
|
|
- **DM** — Concepts whose PO terms are recorded
|
|
|
|
## Required sections (after Metadata / Version History)
|
|
|
|
1. **Purpose and Scope** — the PO language and domain (from the registry's
|
|
`Languages` section, also in the `Language` and `Domain` Metadata rows) and
|
|
what the dictionary covers.
|
|
2. **Dictionary** — one row per term:
|
|
`PO term | Language | IT term | Definition | Used as PO term in | Used as IT term in`.
|
|
The definition is written in the PO language. "Used as PO term in" and
|
|
"Used as IT term in" list artifact types (for example `DM` and `OC, SD,
|
|
DCD, ERD`).
|
|
3. **Rules** — the register split (PO term in the Domain Model, use cases and
|
|
user stories; IT term in the Operation Contract, Sequence Diagram, Design
|
|
Class Diagram and ERD) and one IT term per PO term.
|
|
|
|
One dictionary has one domain: the PO terms are the domain's own words (for
|
|
example `medical`), the IT terms are professional IT. A project whose PO terms
|
|
come from two domains keeps one dictionary per domain, each with its own
|
|
`Domain` row; a term never appears in two.
|
|
|
|
Keep the dictionary in step with the Domain Model: a new concept gets a row
|
|
in the same change. When the PO language is English, the PO and IT columns can
|
|
still differ (a business word against a technical one); keep the file anyway
|
|
when the two registers use different words.
|