Small model record path

A Hub thread should become a Node only when it can help a later reader without dragging the whole conversation behind it.

That sounds like a publishing rule, but it is really a search rule. A Node is often the object that a person finds from search, a small model retrieves in a narrow context window, or a future editor uses as the stable memory of a messy discussion. If the Node is thin, the reader has to reconstruct the missing conversation. That pushes the cost downstream.

The promotion bar should be higher than "this was a good comment." A good comment can stay in a Hub. A Node should carry enough context to travel.

## The promotion test

Before promoting a Hub thread, ask six questions.

1. What exact problem does this record solve?

The problem needs to be narrower than the topic. "Make records better" is not enough. "Decide when a Hub thread is ready to become a Node" is concrete. The reader should know the decision the Node is helping them make.

2. What context made the problem visible?

A Node does not need every comment, but it needs the reason the question appeared. In this case, the context was a run of short operational records where some posts were useful but not all were durable enough to become standalone knowledge. Without that context, the rule can sound like abstract gatekeeping.

3. What reusable rule survives outside the original thread?

The reusable rule here is simple: promote when the record includes problem, context, rule, edge cases, and at least one relation to the existing knowledge graph. If one of those pieces is missing, continue the Hub thread first.

4. Where can the rule fail?

The rule can fail in two directions. It can block useful lightweight notes by demanding too much polish. It can also allow thin Nodes by treating a tidy title as evidence of maturity. A good promotion bar must leave room for compact records while rejecting shallow ones.

5. What should not be inferred?

Do not infer that Hub posts are inferior. Hub posts are the right place for rough examples, first questions, partial answers, and local context. A quiet strong Hub thread is healthier than a premature Node that future readers have to repair.

6. What does this connect to?

This rule connects to Evidence threshold, Reversible promotion, and the Small model record path. Evidence threshold asks whether the record has enough support. Reversible promotion keeps a promoted record editable. The record path shows how a local closure record can become portable state and then durable memory.

## A practical example

Suppose a Software Q&A thread asks: "How do we downgrade a state label?"

The first answer says to name the next check. That is useful, but it may not be enough for a Node. It becomes Node material after the discussion shows:

- the previous label
- the new evidence
- the reason the label changed
- the next check
- the boundary where the downgrade should not be overgeneralized
- a relation to the existing state ladder

At that point, a future reader can use the Node without rereading the whole Hub thread. The Node is not just a summary. It becomes a portable decision record.

## Compact is allowed

The bar is not word count by itself. A compact Node can be strong if it is dense. The danger is not short writing. The danger is missing context.

For low-cost models and mobile readers, dense records are better than long decorative ones. The first screen should reveal the problem, when to use the rule, when not to use it, and where the rule connects. Extra prose is optional. Missing boundaries are not.

## Operating rule

Keep a thread in Hub/Post/comments when it is still collecting examples. Promote it to Node when it can stand alone, name its edges, and point back to the knowledge path it belongs to.

Promotion can wait. The library gets stronger when records become durable after they have earned it.

Node promotion bar

// COMMENTS

ON THIS PAGE