# PDR 000: Publish typed financial and business subjects in the connected graph

- HTML version: https://robbiepalmer.me/projects/personal-knowledge-graph/pdrs/000-typed-financial-and-business-subjects
- Project: Personal Knowledge Graph
- Status: Accepted
- Opened: 2026-10-31
- Decided: 2026-10-31

## Context

The graph can publish an abstract idea, a technology, a project, a blog post,
and a job role. A role already connects one period of employment to the
technologies used there, the projects completed there, and the blog posts
written there. Those links make the role part of the public account of the
work rather than an isolated resume entry.

There is no equally precise home for researched financial and business
subjects such as IR35, umbrella companies, dividends, Capital Gains Tax, and
the 60% tax trap. They are not all ideas or laws. A generic reference page
would classify the document rather than the subject and could also be mistaken
for a link to a third-party source.

These subjects have different meanings and constraints. IR35 is a regulatory
regime. An umbrella company is an organisation type. A dividend is a company
distribution. Capital Gains Tax is a tax. The 60% tax trap is an effect caused
by tax rules. Flattening those distinctions into tags or an open-ended `kind`
field would make the graph harder to query and easier to fill with ambiguous
records.

## Decision

Publish financial and business subjects as first-class, typed entities in the
existing connected knowledge graph. Do not add a `Reference` entity or treat
these subjects as a separate public graph.

The initial vocabulary will support the concrete classes needed by published
content:

| Subject                  | Class             |
| ------------------------ | ----------------- |
| IR35                     | Regulatory regime |
| Umbrella company         | Organisation type |
| Dividend                 | Distribution type |
| Capital Gains Tax        | Tax               |
| Personal Allowance taper | Tax rule          |
| 60% tax trap             | Tax effect        |

Law and work arrangement will also be distinct classes when content needs
them. A law represents an enacted legal instrument. It does not become a
catch-all for a regime, tax, company form, or observed tax effect.

Each subject will have a canonical page whose route and presentation identify
its class. The site may group these pages under a plain navigation label such
as "Finance and business", but that label will not replace their types in the
graph.

The new entities must join the relations that already explain work. A job role
will keep its existing technology, project, and blog connections and may also
record its work arrangement, the organisation form through which the work was
performed, and any regime that governed the engagement. Projects may record
which financial or business subjects they model. Blog posts may explain or
discuss them. The result should support a traversal such as:

```text
Role -> project -> models -> Capital Gains Tax
Role -> governed by -> IR35
Role -> engaged via -> Umbrella company
Role -> blog post -> explains -> Dividend
```

Authoritative third-party pages remain sources for claims. They do not define
the class of the local page.

Private financial facts remain outside the public content graph. A personal
finance application may later classify an employment, payment, or disposal
against the same public domain entity. That reuse must not publish the private
record or turn the public subject into an instance of one person's finances.

## Consequences

Readers can move between professional history, projects, writing, and the
financial or business rules that affected the work. The graph preserves the
existing role as a junction instead of placing finance content in an isolated
section.

Authors must choose a real class and relationship when adding a subject. That
cost is intentional. If a proposed subject does not fit the vocabulary, its
addition requires a small ontology decision rather than a generic fallback.

Some subjects vary by jurisdiction and effective date. Their pages must make
that scope visible and keep dated rule facts separate from the stable identity
of the subject. The site will need maintenance checks for stale effective
rules and their authority sources.

## Alternatives considered

### Add a generic reference page

This would make new pages cheap to add. It was rejected because `Reference`
describes an editorial use. It does not identify the entity or distinguish a
tax from an organisation type or regulatory regime.

### Extend ideas with a kind field

This would reuse the current idea routes and relations. It was rejected
because ideas represent abstract concepts. A discriminating string would
also leave one broad schema responsible for unrelated legal, financial, and
commercial entities.

### Publish a separate finance graph

This would create a clear storage boundary. It was rejected for public
knowledge because employment roles, projects, technologies, and writing
already form one connected account. A second graph would require joins to
recover relationships the site already knows. Private personal data still
keeps its own access and storage boundary.

## Related records

- Project: [Personal Knowledge Graph](https://robbiepalmer.me/projects/personal-knowledge-graph.md)
- Informed by ADR: [ADR 033: In-Memory Content Graph](https://robbiepalmer.me/projects/personal-knowledge-graph/adrs/033-content-graph-indexes.md)
- Implemented by ADR: [ADR 060: Model financial and business subjects as typed content graph nodes](https://robbiepalmer.me/projects/personal-knowledge-graph/adrs/060-typed-financial-and-business-subjects.md)

---

Markdown index of this site: https://robbiepalmer.me/llms.txt