#vocab/meta # PKM Vocabulary — Open Questions Version 0.1.3 is a first draft and I expect to be wrong about things. These are the places I'm least confident, kept here rather than buried in the README so they can be argued with and revised independently of the announcement. Nothing gets deleted when I change my mind — a retired term stays resolvable and points at whatever replaced it via `dcterms:isReplacedBy` — so it is safe to link to any of this now. ## 1. Should the time periods nest, or sit side by side? Right now [[Day (Term)|Day]], [[Week (Term)|Week]], [[Month (Term)|Month]], [[Quarter (Term)|Quarter]], [[Year (Term)|Year]] and [[Decade (Term)|Decade]] are siblings under Calendar. The only nesting is partitive: a Day is *part of* a Week. The case for nesting: it is simply true. A Day is part of a Week which is part of a Month. The case against: the tree gets deep, and the notes at each level don't inherit anything from the level above — a Week Review is not a kind of Month Review, and nothing about a Month Plan is derivable from a Quarter Plan. Depth that carries no inheritance is depth that costs navigation and buys nothing. **Where it stands:** siblings, with `isothes:broaderPartitive` carrying the part-of relation where it genuinely holds. ISO 25964's partitive relation exists precisely for this — a whole/part link that doesn't claim kind-of. That feels closer to right than either extreme, but I keep going back and forth. ## 2. Are the names right? "Cluster" for the dozen notes belonging to a single day — [[Day Cluster (Term)|Day Cluster]], and the same shape for weeks, months, quarters, years. I went back and forth for a long time. It is not a *collection* in the SKOS sense (those are already used here for something else), not a *group* (overloaded everywhere), not a *set* (implies no structure). "Cluster" suggests things that belong together without claiming they form a container, which is what I meant. I am still not sure a reader arrives at that meaning unaided. ## 3. What's obviously missing? A whole layer of LYT practice is not modeled yet: MOCs, home notes, maps, bases, workflows. The reason it's missing is that I couldn't decide whether they are **terms or techniques**. A MOC is a note type — that's a term. "Make a MOC when a topic reaches a dozen notes" is a technique, and a vocabulary is the wrong place for it. But the boundary isn't clean: is a *home note* a note type, or a role a note plays? Roles may want modeling as something other than concepts. Related open question: several terms in the vocabulary are Obsidian *plugins* (`Advanced URI`, `Add Folder`) sitting alongside genuine domain concepts. That may be a category error of the same family — tools and concepts in one hierarchy. ## Known warts Not conceptual questions, just things I know are wrong and haven't fixed: - **`owl:versionIRI` is minted in the term namespace.** It currently resolves to `https://w3id.org/pkm/vocab/0.1.3`, which collides with the term URI pattern `https://w3id.org/pkm/vocab/{Term}` — a version is not a concept and shouldn't live in the concept space. - **Per-term filenames diverge from labels in a few places.** "Template" serializes to `TemplateCopy.ttl` and "Month Health" to `Month-Health.ttl`. Harmless for resolution, but it means the filename is not reliably derivable from the label. ## Feedback The fastest way in is the [hierarchy view](https://w3id.org/pkm/vocab). Pick at it — tell me what's wrong with it. That's what a 0.1 is for.