Orientation only. This names the subdomains and how they relate — it is deliberately shallow and broad. The one context modelled deep is Tenancy & Arrears (the payments seam) in
domain-model.md. Everything else here is named to show where it sits, not modelled.Items marked (confirm) are my proposals awaiting your call.
DDD sorts subdomains by where modelling effort is worth spending:
- Core — the differentiator; go deep.
- Supporting — necessary, business-specific, not a differentiator; model lightly.
- Generic — every business needs it; buy / off-the-shelf; don't model.
Core ≠ critical. The axis is differentiation, not importance. Trust accounting is mission-critical (mishandle trust money and you lose your licence), yet it's Supporting, not Core — it's legislated, done identically by every agency, and buyable off the shelf (PropertyMe, Console). No landlord picks an agency for its trust ledger. The test: could I buy it, and would a customer switch for a better one? Core is where you can be better; trust accounting is where you must merely be compliant. The differentiator here is tenancy/arrears management (the tribunal-ready timeline), which is why PM is Core, Accounts Supporting — even though we may still build Accounts deeply for our own reasons.
PM is the core domain. It contains several bounded contexts (each its own model + language, even though they're all "PM"):
| Bounded context | Language / concerns | Depth here |
|---|---|---|
| Tenancy & Arrears | tenancy, lease terms, rent due, arrears, days-behind / periods-behind, notice, overstay; arrears collections (notices, NCAT/tribunal, termination) live here | Deep — domain-model.md |
| Maintenance / Work Orders | tenant requests, work orders, tradies, owner approval | Named only |
| Inspections | routine inspections, entry/exit condition reports | Named only |
| Compliance | smoke alarms, safety checks, bond lodgement, certificates & due dates | Named only |
A context in its own right (deep in reality, stubbed here to the payment-facts
edge). Owns: receipting, trust ledger, suspense (UNKNOWN), reversals,
invoices (incl. maintenance/repair bills paid on behalf), disbursement, owner
statements, fees.
Owner-facing origination: landlord onboarding, management / agency agreements (BDAs), fee structures.
Advertising vacancies (incl. sign boards and property signage), applications, screening, approving a tenant, executing the lease — the pipeline that produces a tenancy.
Capability, not org chart. Who physically handles advertising / signboards varies by agency — in-house agents, the back-office, or an outsourced vendor tee'd up by the office manager — but it's the same capability, so it maps to the same subdomain. In-house vs outsourced is a sourcing detail, below the modelling altitude. The map models what the system does, not who does it.
Identity & users · Documents & e-signing · Payment rails / bank feeds · Notifications · Reporting / BI · Office / internal admin (procuring computers, stationery, running the office) — real agency work, but not part of this system's domain. A reminder that the map models the product, not everything the business does.
The interesting boundaries — a concept in one context becomes a different concept in another:
- Accounts → Tenancy & Arrears — a payment becomes an arrears reduction.
← modelled (
domain-model.md). - Maintenance → Accounts — a work-order invoice becomes a bill paid on behalf → a disbursement deduction on the owner statement.
- Leasing → Tenancy — an approved application + executed lease becomes
TenancyCommenced. - Business Development → Accounts + Tenancy — a signed management agreement sets up the owner + fee structure (→ Accounts) and authorises management (→ PM).
- Within PM — Inspections → Maintenance (a defect found becomes a work order); Compliance → Maintenance (an overdue/failed check triggers remediation).
Only seam #1 (Accounts ↔ Tenancy & Arrears) is modelled deeply. This map exists so we know where that seam sits — and which seams (#2–#5) would be the next interesting ones to model if the project grows.