Every concept in a taxonomy lives in a single `xs:element` declaration in a schema document. Open the file, find the element by id or by name, and ten seconds of reading will tell you almost everything you need to know to use it correctly in a filing.
The attributes you care about, in order: `name` (the local part of the QName), `id` (the NCName id used to anchor links from labels and references), `type` (which item type it is), `substitutionGroup` (whether it is an item, a tuple, a hypercube, or a dimension), `xbrli:periodType` (instant or duration), `xbrli:balance` (debit or credit, and only on monetary concepts), `abstract` (whether the concept can carry a fact value at all), and `nillable` (whether facts of the concept may be reported with `xsi:nil="true"`).
`abstract="true"` is the attribute that catches the most filings off guard. An abstract concept exists to be a parent in presentation arcs and to anchor structure in definition linkbases. It cannot legally bear a fact value. ESEF specifically codes anchoring extensions to abstract base concepts as `ESEF.3.3.1.ExtensionConceptAnchoredToAbstractConcept`, and authors keep doing it because the label looks right and the namespace looks right.
`nillable="true"` permits explicit absence: a fact reported with `xsi:nil="true"` says "I am asserting that this concept has no value for this context", which is different from omitting the fact entirely. SBR rule FR-NL-5.07 governs nil usage in Dutch filings; ESEF and EFM treat nil more permissively but still flag certain numeric concepts where nil is clearly inappropriate.
The deprecation signal lives in the label linkbase, not the schema, but it is the next thing to check. A concept with a `deprecatedLabel` and a `deprecatedDateLabel` is on its way out of the taxonomy, and the regulator's validator will warn (EFM 6.5.42 and analogous codes) when it appears. Migrate to the replacement concept in the next filing.
Reading one `xs:element` answers "what is this concept". The question one level up is "where did all these concepts even come from", and the answer has a name worth knowing cold: the Discoverable Taxonomy Set, or DTS. Nobody hands a processor a list of taxonomy files. An instance points at a single entry point through `link:schemaRef` (inside `ix:references` for inline filings), and from that one starting document the processor discovers everything else by following references: `xs:import` and `xs:include` to other schemas, `link:linkbaseRef` to presentation, calculation, definition, and label linkbases, `link:roleRef` and `link:arcroleRef` to role declarations, and `xlink:href` from locators back to the concepts they point at. It follows those edges, then follows the edges of whatever it just found, and keeps going until nothing new turns up. XBRL 2.1 section 3.2 calls that the DTS discovery process, and section 5.1 makes the rule precise. The DTS is the result: the complete, closed set of documents reachable from the entry point.
If that sounds like a reachable set in the language of graphs, the vocabulary the next chapter makes explicit, that is exactly what it is. The taxonomy documents are vertices, the references are directed edges, and the DTS is the transitive closure starting at the entry point. Three consequences follow and all of them bite in practice. First, the DTS is discovered, not declared, so you cannot judge a filing by the files sitting next to it on disk; you have to follow the graph. Second, the entry point is load-bearing: point a Dutch KvK filing at the schema for the wrong entity-size class and the processor discovers a different, wrong DTS, and every concept and rule downstream is quietly the wrong one. Third, errors that live in this graph carry the `xbrldte:*` prefix (DTS errors, like `TooManyDefaultMembersError`), which is why Chapter 7 routes `xbrldte:*` to a different, earlier place in your triage than the instance-level `xbrldie:*` codes: a broken DTS is a broken graph, and it has to be healed before any fact inside it can be trusted. Processors cache a resolved DTS by content, which is also why an offline taxonomy package has to reproduce the same closure, byte for byte, that an online discovery would have reached.
The fastest practical question, when you meet a concept you have never used before, is "is this a stock or a flow?". `periodType` answers it, and the rest of the metadata then tells you which unit, which balance, which substitution group, and whether you can tag against it at all.
Open the schema, find the xs:element by id or by name. Read its type, periodType, balance (if monetary), substitutionGroup, abstract, and nillable attributes.
Type tells you whether the concept is monetary, share-typed, percent, decimal, string, text-block, date, boolean, or QName-typed.
abstract='true' means the concept cannot carry a fact value. It is a placeholder used in presentation and definition arcs.
nillable='true' allows the fact to be reported with xsi:nil='true' to indicate explicit absence. SBR rule FR-NL-5.07 governs nil usage.
deprecatedLabel and deprecatedDateLabel mark concepts no longer recommended; using them produces a warning in EFM 6.5.42 and similar regimes.
<xs:element name="GrossProfit"
id="ifrs-full_GrossProfit"
type="xbrli:monetaryItemType"
xbrli:periodType="duration"
xbrli:balance="credit"
substitutionGroup="xbrli:item"
abstract="false"
nillable="true"/> report-2025.xhtml
| ix:references -> link:schemaRef (the ONE entry point)
v
my-entrypoint.xsd ---- xs:import ----> ifrs-full.xsd ---- import ----> xbrl-instance.xsd
| |
| link:linkbaseRef | link:linkbaseRef
v v
my-ext_pre.xml ifrs_lab-en.xml
my-ext_cal.xml ifrs_cal.xml
my-ext_def.xml --xlink:href--> back into ifrs-full.xsd concepts
THE DTS = every document reachable by following schemaRef, import,
include, linkbaseRef, and xlink:href edges (and everything THEY reach),
until the set stops growing. That "until it stops growing" is the
transitive closure: the same reachable-set idea from Chapter 4,
applied to documents instead of concepts.3 questions. Druk op 1 tot 4 om te kiezen, Enter om te bevestigen.
Door deze quiz af te ronden, markeer je deze sectie als voltooid.
Tip: kies met 1 tot en met 4 en bevestig met Enter.