[Capability question] A chart axis cannot be type-checked at authoring time — a dataset MEASURE has no datatype to declare, and type is already spent as an aggregate alias #16959

Escalated from the objectui domain:ui execution seat (session session_01YBWFb5YgMU5dw8p2VKj16S), measured on objectui#7195. ⛔ Not graded, not assigned, and ⛔ no fix proposed — the first question here is a capability question, and answering it is above the seat that measured it.

The symptom that led here (objectui side, already recorded)

A chart axis bound to a column whose every row is boolean draws zero marks and says nothing. Measured through the real renderer, each zero behind a control that drew:

• x all-boolean: 0 of 2 marks. Full tile text was exactly ym010203040 — the legend label plus the y ticks. A confident y scale, no x scale, no marks, no sentence.
• One real number present among the booleans: 3 of 3 marks.

⭐ Recharts needs one real number to establish the scale, then coerces the booleans onto it — so "does a boolean coordinate draw" is a property of its neighbours, not of its own value. That is why objectui refused to guard it in the renderer: a refuse-booleans predicate would give the mixed tile a footnote reading "2 of 3 rows were not drawn" while three points are visibly on screen, and a false sentence about the picture is worse than silence (objectui#7146 / #7147).

⇒ objectui#7195 stays open as the symptom record. This card is the upstream half.

⭐ The actual finding: the type is UNDECLARABLE, not merely unread

Measured on this repo:

1. Every chart-axis binding key is a bare string. ChartAxisSchema.field is z.string() ("Data field key"); ReportChartSchema.xAxis / yAxis are bare strings naming a bound-dataset dimension or measure; ChartAggregateSchema.field names an object field. None carries or resolves a datatype.
2. The only datatype on the path is DatasetDimensionSchema.type — and it is optional, so it can simply be absent.
3. ⭐ DECISIVE: DatasetMeasureSchema has no datatype at all. Its type key is an alias for aggregate — this repo's own comment calls it "THE dangerous one", because a Cube-trained author writing type: 'sum' silently lost the aggregation.
4. Scatter binds two MEASURES. objectui states it in its own source: "scatter is excluded on purpose: both of its axes are numeric measures."

⇒ On the tier a scatter axis actually binds, there is no datatype to declare. Refusing a boolean there is not a tightening of an existing check; there is no check and no place to put one.

5. Nothing cross-references a chart axis to a dataset declaration. The closest existing refinement, checkReportOrder, checks name membership within the same document — never a type. A sweep for a type-aware refinement across the chart / report / dashboard schemas returned only boolean flags and the count-needs-a-field refinement.

The capability question — please rule on this first

Should a chart axis resolve to its dataset measure or dimension declaration at authoring time, and should a measure be able to declare a datatype at all?

⛔ Everything else depends on that answer, which is why nothing is proposed here. Any refusal is a narrowing of an accept surface and therefore above the manual floor whichever repo it lands in.

⚠️ A hazard for whoever takes it: the obvious key name is already spent

If the answer involves letting a measure declare a datatype, ⛔ it cannot be spelled type. On DatasetMeasureSchema that word is already the aggregate alias, so an author writing type: 'boolean' would be renamed onto aggregate and then fail the aggregation enum — a silent rename into a confusing error. A different spelling is required, and it should reuse the dimension tier's existing vocabulary so the two tiers do not drift into two datatype vocabularies.

⭐ Note this is additive if taken alone: adding an optional key to a strict object widens the accept surface rather than narrowing it, so it could be a cheap first step. ⛔ But it only pays off if the capability question is answered the way it presumes, so it should not be taken before the ruling.

⚠️ Scope: this governs at least four renderer sites and both axes

Measured on objectui, each behind a control that drew: the y axis has the identical hole (0 of 2, silent); both axes all-boolean collapses the tile text to just ym with no scale on either side; and line, bar and area all draw zero and say nothing on an all-boolean measure. Pie / donut / funnel / treemap already speak, because Number(false) is 0 and trips an existing magnitude test.

⇒ The silent family is the cartesian / positional one. A ruling should be made once, upstream, rather than per chart arm — splitting it invites different answers to one mechanism.

What is NOT claimed

• ⛔ No priority claimed; objectui graded its symptom card p3 (narrow trigger, not a regression, no dataset in a sweep produced it accidentally) and that grading is not transferred here.
• ⛔ Not measured by me: whether any authored dataset in this repo or in cloud actually binds a boolean column to a cartesian axis. The objectui sweep found none arising accidentally; that is a reachability reading on one repo, not a census.
• ⛔ The objectui-side declarations (ObjectChartSchema.xAxisField / yAxisFields) are bare strings too, but they cannot be fixed there alone — objectui pins that these schemas belong upstream and that a local fork is the failure mode.