PM-FLT-0040
A Filter without a quantifier is not a proposition
Withdrawn. Superseded by PM-FLT-0060 — A Filter's proposition is existentially quantified on both positions. The reasoning below — what the sources actually say, and why this reading of them was wrong — is kept exactly as written at the time, not deleted.
A Filter MUST declare how many Subject-side instances must satisfy the Subject State for the proposition to hold. No source states this, and the specification MUST NOT assume a default.
Non-normative and important. Take the blog's own example, "Created Artefact isAuthoredBy Active Artist". Four readings, four different systems:
| Reading | Holds when | |---|---| | SOME | at least one authoring Artist is Active | | EVERY | the relation is non-empty and every authoring Artist is Active | | NONE | no authoring Artist is Active | | COUNT ⋈ n | the number of Active authoring Artists satisfies a comparison |
A modeller writing that sentence in a workshop means one of these and will not notice that the notation did not record which. Any implementation must therefore either demand a quantifier at authoring time or silently pick one, and silently picking one is how a model comes to mean something its author did not intend.
EVERYcarries a second question: is it vacuously true on an empty relation? Almost never what a business analyst means, but it is the standard logical reading.
Withdrawn 2026-09-08. The reasoning above was sound and the conclusion was wrong. The claims do not state a quantifier and the blog does not either, but the patent description does, twice, in the section describing the Filter construct. Kept as the record of what we believed, and as a standing reminder that "no source states this" is a claim about our reading, not about the sources. See PM-FLT-0060.
CONFLICT MUST · withdrawn · area FLT · since 0.1.0
Source
No source citation — trust token CONFLICT; see the reasoning above.
Relationships
Blocked by Q-G-0003