Glossary

The words the site uses, from cited and stated to schema.org and sameAs, each in a sentence or two. Every term links to the page that explains it in full.

Four terms of the glossary, cited, namesake, OAI-PMH and sameAs, each joined to the About page that explains it.

The words the site uses for what it holds and where it came from, each with the page that explains it. A link's kind of evidence is told by a mark on every person's page: a filled dot for cited, a ring for stated, a dashed ring for claimed or inferred.

The record

Lineage: A person's advisors, their advisors' advisors and so on upward, and their students and their students' students downward, followed through doctoral supervision. The record: people, theses and works
Supervision: A thesis record that names a student and the advisor who supervised them. A doctoral supervision is a link of the lineage. The record: people, theses and works
Level: The degree a thesis was written for: a doctorate, a master's, a licentiate, a bachelor's, or unknown where its record does not say. Only doctoral supervision counts as lineage. The record: people, theses and works
Opponent: At a public defence in Sweden or Finland, the scholar from outside the university appointed to question the candidate on the thesis. The record: people, theses and works
Custos: At a Finnish defence, the professor who presides over it, often the candidate's supervisor. The record: people, theses and works
Coverage: How far the record reaches, counted each night by country, decade and level. Where a count stops, the record stops, which is not evidence that nothing more exists. What the record does not know

Where a link comes from

Provenance: Where a statement came from: the record that says it, the kind of evidence it is, and when it was read. Where every link comes from
Provenance knowledge graph: A graph in which every link carries its provenance. The record is built as one: no link stands without the source and the kind of evidence behind it. Where every link comes from
Cited: A link a source's own record gives, as a thesis record naming its supervisor; drawn with a filled dot. The doctoral supervisions counted are the cited ones. Where every link comes from
Stated: A link Wikidata states, as a doctoral advisor or student; drawn with a ring and shown apart from the repositories' records. Where every link comes from
Extracted: A supervisor found in the text of a thesis itself, its title page or its acknowledgements, by a language model, kept with the line it was found in. Where every link comes from
Inferred: A link a rule suggests rather than a record states, such as coauthorship drawn from shared papers; drawn with a dashed ring and never counted. Where every link comes from
Claimed: What a person says of themselves on their ORCID record, such as the works it lists; shown apart from the record's own and never counted with them. Where every link comes from

Sources

Source: Anywhere the record reads from: a university repository, a national register, an aggregator, Wikidata, OpenAlex or ORCID. Where the data comes from
Repository: A university's or institute's open archive of the theses and papers written there, describing each in a record a program can read. Where the data comes from
Register: A list an authority keeps: of a country's theses, as RENATES is for Portugal, or of people, as ORCID, IdRef and GND are, to which a person may be linked. Where the data comes from
Aggregator: A service that gathers the records of many repositories, as OpenAIRE, BASE and CORE do. A repository that cannot be reached may be read through one, its records then counted as the aggregator's. Where the data comes from
OAI-PMH: The protocol of the Open Archives Initiative, by which a repository hands the descriptions of its records to any program that asks, a page at a time, and those changed since a given day. Where the data comes from
Wikidata: The open knowledge base of the Wikimedia projects. Its statements of doctoral advisors and students are part of the lineage, marked as stated. Where the data comes from
OpenAlex: An open index of scholarly works and their authors: the source of a person's papers, coauthors and citations, none of them counted as lineage. Where the data comes from

Who is who

Namesake: Someone who shares a name with another person. Across sources namesakes stay apart until an identifier or a checked match joins them; within one repository a name is taken as one person. Who is who
Linked: A person is linked when a source gives them an id used beyond this record, such as an ORCID iD, or a checked match ties them to Wikidata or OpenAlex. Who is who
ORCID iD: The identifier a researcher registers for themselves at ORCID. Where a record states one, the person can be linked to Wikidata or OpenAlex, and through them to the same person elsewhere. Who is who
A checked match, a likely match: Two records joined as one person after a check: a checked match is one confirmed, a likely match one the person reading the pair was unsure of, shown as likely and not confirmed. Who is who

Measured

Check: A sample of what a reading or a join gives, compared by a person with what the source's own pages say, and scored against a bar set in advance. Measured, not asserted

Models and assistants

Large language model: A program trained on a great deal of text to read and write language. No link of the record is written by one; models are used only in a few bounded places. Where AI is used, and where it is not
Retrieval-augmented generation: Answering a question by first finding the documents that bear on it and handing them to a language model to answer from. The site's assistant does the same with exact retrieval: what it answers from is the sheet in front of the reader. Where AI is used, and where it is not
Prompt cache: Anthropic's keeping of the start of a request, here the sheet, so the later questions of a conversation about it are answered without sending it to be read anew. Where AI is used, and where it is not
Model Context Protocol: An open standard for giving an AI assistant tools it can call. The record's tools let an assistant such as Claude look up people, lineages and routes under the reader's own account. Your assistant and the record

Open data

Schema.org: A shared vocabulary for describing people, theses and other things as data that search engines and programs read. Each person and thesis page has its contents in it at a second address. Open data and citation
sameAs: The schema.org property that names other pages about the same person: on a person's data, the registers a source or a checked match links them to. Open data and citation