Glossary

Twenty-two terms that show up in OpenXiv policies, the API surface, and answer-engine queries. The list is intentionally small; if a term is missing, the AT Protocol spec and the OpenXiv policies pages have longer treatments.

Set @id: https://openxiv.net/glossary#terms

AT Protocol at-protocol: The protocol behind Bluesky and OpenXiv. AT Protocol gives each user a personal data server (PDS) holding signed records under reverse-DNS namespaces, supports account portability across providers, and publishes a public jetstream that indexers consume. OpenXiv uses AT Protocol for its source-of-truth records.
did:plc did-plc: The default decentralized identifier method on AT Protocol. A did:plc string is a stable identifier for an account that can be rotated to a new key or moved between PDS providers without changing the underlying ID. OpenXiv accepts did:plc as a sign-in identity.
did:web did-web: An alternative DID method that anchors an identifier to a domain name. A user with did:web:example.org publishes their DID document at https://example.org/.well-known/did.json. OpenXiv supports did:web identities for users who host their own PDS or want a domain-bound identity.
PDS (Personal Data Server) pds: An AT Protocol server that holds a single user (or a small group) and their signed records. Bluesky operates a default PDS; users may also self-host. OpenXiv writes app.openxiv.* records to the author's PDS at submission time so the records remain portable.
App View app-view: An AT Protocol indexer that aggregates records from many PDS instances and exposes them through a public API. OpenXiv runs an App View that reads from the jetstream, indexes OpenXiv lexicon records, joins them with binary artifacts (PDFs, HTML), and serves the public site.
Trust Passport trust-passport: A signed evidence vector attached to each OpenXiv paper. The passport tracks seven lanes (transparency, identity, provenance, citations, math, integrity, social review) with per-lane evidence items, open disputes, author responses, and any external attestations.
Refusal packet refusal-packet: The public document OpenXiv publishes when refusing a submission. The packet names the specific failure mode, points at the evidence, and explains what would have to change for re-submission. Stored as an app.openxiv.refusal record.
Typed endorsement typed-endorsement: An endorsement that carries a specific verb (verified derivation, reproduced result, checked references, useful background, important but flawed, needs correction) rather than a generic thumbs-up. The full vocabulary is at /vocabulary.
Explainer tier explainer-tier: One of three audience-tiered summaries that accompany every OpenXiv paper: school (curious teenager), undergrad (you have seen the field), expert (you are in this subfield). Generated by an LLM, labelled as AI-generated, editable by the author.
AI disclosure level ai-disclosure-level: One of four tags on every OpenXiv submission: none (no AI), assistant (drafting / grammar / search help), coauthor (AI generated substantive content), primary (AI is the principal generator). Stored on an app.openxiv.disclosure record and exposed as a badge on the paper page.
Provenance timeline provenance-timeline: An eight-stage public timeline showing how a paper got from upload to indexed: source upload, source validation, PDF compile, HTML conversion, identifier assigned, PDS write, Bluesky bridge, and feed publication. Each stage has a timestamp where one exists.
OAI-PMH oai-pmh: The Open Archives Initiative Protocol for Metadata Harvesting. OpenXiv exposes /oai-pmh with the standard verbs (Identify, ListSets, ListRecords, GetRecord) and oai_dc metadata so external aggregators (OpenAIRE, BASE, CORE) can harvest the corpus.
ScholarlyArticle scholarly-article: The schema.org type emitted in JSON-LD for every OpenXiv paper. Includes headline, identifier, abstract, author list with ORCID sameAs, license, datePublished, and a Periodical isPartOf block with the OpenXiv ISSN.
Crossref crossref: A DOI registration agency for scholarly content. OpenXiv plans to mint DOIs via Crossref as part of the preservation roadmap; until then, the OpenXiv identifier is the persistent citable ID.
ORCID orcid: A persistent digital identifier for individual researchers. OpenXiv lets authors sign in with ORCID and attaches the orcid.org URL to each author entry in the paper's JSON-LD sameAs and in the HighWire citation_author_orcid tag.
Federation federation: On OpenXiv, federation means writing a record to the author's PDS and producing a Bluesky bridge post so the paper shows up in normal social feeds. It is a direct side effect of AT Protocol, not a separate cross-poster.
JSON-LD json-ld: A linked-data serialization format. OpenXiv emits JSON-LD payloads (Organization, WebSite, ScholarlyArticle, FAQPage, DefinedTermSet, BreadcrumbList, ItemList, HowTo) so search engines and answer engines can read the page as structured data rather than free text.
ISSN issn: International Standard Serial Number, an eight-character identifier for serial publications. OpenXiv is registered under ISSN 3120-9556. The number is emitted as citation_issn, in dc.identifier as urn:issn:3120-9556, and on the Periodical block of every paper.
DOI doi: Digital Object Identifier. A persistent, resolvable identifier for digital objects. OpenXiv's DOI minting roadmap goes through Crossref; for now the OpenXiv identifier (assigned at acceptance) is the persistent citable string.
Jetstream jetstream: AT Protocol's firehose of newly created and modified records across the network, served as a websocket. OpenXiv's App View consumes the jetstream filtered to app.openxiv.* and app.bsky.feed.post collections to keep the index in sync with what authors publish on their PDSes.
Lexicon lexicon: AT Protocol's schema language. A lexicon describes a record type (its fields, validation rules, and parent collection). OpenXiv defines lexicons under the app.openxiv.* namespace: paper, summary, disclosure, endorsement, refusal, version.
BullMQ bullmq: A Redis-backed job queue used by OpenXiv's API service to coordinate the submission pipeline: source validation, PDF compile via Tectonic, HTML conversion via LaTeXML, figure extraction, summary generation, PDS write, and Bluesky bridge. Each stage is a separate BullMQ job.