Search
An exact-match search over what this record publishes about each tracked entry, subdivision, finding, organisation and entry of the exact-term index: names, the identifier this record gives each tracked entry, finding, organisation and term (a subdivision is found by its own name), the publisher it records, and the publisher’s own words where the record banked them with the address they were read at. The index is in this page; searching it makes no request of any kind.
A query matches a span exactly or it does not match. There is no stemming, no synonym list and no ranking; the one normalisation is case, and a case-folded match is labelled as one. Each row names the field it matched in, the window around the match, the address the span is cited to, and the evidence state of the entry it is or subdivides. The method is stated in full under the exact-match index.
- 198 rows carrying 1191 spans: 132 × a tracked entry, 26 × a subdivision of a tracked entry, 29 × a published finding about this record, 6 × an organisation that produces documents, 5 × an entry of the exact-term index.
- 140 of the 1191 spans are cited to a publisher — a span this record banked together with the publisher URL where it was read; the other 1051 are text this record publishes on a page of its own, cited to that page.
- Evidence states over the same 198 rows, in the order the classifier tests them: 30 not a document, 26 a register, 3 unstated by the publisher, 77 verified by a person, 7 source access limited, 1 publisher metadata, 13 observed, 1 not currently retrievable; and 40 rows whose type carries none.
The box is inactive until this page’s script enables it. With scripting off it stays inactive, and every row it searches is reachable without it: tracked entries from the board and the sources, subdivisions from the pages of the tracked entries they subdivide, findings from the findings, organisations from the bodies, and terms from the exact-term index.
The filters are inactive until this page’s script enables them.
Each one set is written into this page’s address after the
#, the part a browser never sends, so a copied address opens
on the same view.
What this index does not carry.
Across the tracked entries, the record reads 32 keys taken, when the
site is built, from the publisher addresses those entries record, and
serves them on /api/v1/instruments.json.
This index adds none of them as a span of its own. 28 of the 32 are contained in no span
it holds, byte for byte or with case folded, so a search for one of those
finds nothing here; the others are found inside a span it already holds,
such as a title or an alias. 3 publisher
sentences sit beside prose naming where they were read rather than an
address, and are not in it either. The passages the exact-term index quotes
are on its page; only each entry’s name is
here. No document’s body is indexed: the publisher’s words it
holds are the passages the record banked, each beside its address.