digital object identifier
Enable an AI agent to recognise a digital object identifier, establish what it identifies, assess its registration and resolution evidence, and select justified lookup or maintenance actions.
Research draft, second pass
A second pass drafted this model: the structure a model of this thing needs, and what is known about it in the world. The line under this one says how the second half was obtained - researched against sources, or recalled without web access, in which case nothing here was read anywhere and every claim is a lead to verify. Unreviewed either way.
Researched by: Codex + Grok
Purpose and description
Enable an AI agent to recognise a digital object identifier, establish what it identifies, assess its registration and resolution evidence, and select justified lookup or maintenance actions.
A Digital Object Identifier (DOI) is a persistent, resolvable identifier assigned to a digital or physical object under the DOI system, consisting of a unique character string that is registered in a central directory and bound to current location and metadata so that the object can be located even when its URL changes.
It can be Extract and compare DOI names while retaining the original representation and transformation evidence.; Discover the registration agency and retrieve available registration or descriptive metadata.; Resolve a DOI and inspect its destinations or supported negotiated representations.; Match the declared referent and granularity to a citation, object or task requirement.; Investigate conflicting metadata, suspected duplicate assignments or broken destinations.; Prepare or perform authorised metadata and destination corrections through the responsible service..
Distinguishing features
A candidate DOI name has a prefix beginning with the directory indicator 10 and a slash-separated suffix; matching syntax alone does not establish registration. [DOI Handbook: Numbering](https://www-old.doi.org/doi_handbook/2_Numbering.html)
A DOI name and a URL presenting that name must be distinguished from the destination URL returned by resolution. [DOI Handbook: Resolution](https://www-old.doi.org/doi_handbook/3_Resolution.html)
A DOI names a declared referent rather than identifying content through a computed hash; referent meaning must be established from metadata, not decoded from the suffix. [DOI Handbook: Numbering](https://www-old.doi.org/doi_handbook/2_Numbering.html)
A DOI can identify a physical entity or an abstraction as well as digital material, so being a downloadable file is not a recognition requirement. [DOI Handbook: Numbering](https://www-old.doi.org/doi_handbook/2_Numbering.html)
Two different DOI names may identify the same referent; identifier equivalence and referent equivalence therefore require separate tests. [DOI Handbook: Numbering](https://www-old.doi.org/doi_handbook/2_Numbering.html)
Scope
+ DOI name recognition, representation and equivalence
+ Registration evidence and responsible registration agency
+ Referent identity, granularity and identifier relationships
+ Descriptive metadata provenance and agreement
+ Resolution behaviour and destination correspondence
+ Persistence obligations and authorised record maintenance
- The content, quality and lifecycle of the identified work, dataset, physical object or other referent
- Full bibliographic descriptions and citation-style rendering
- Hosting infrastructure and preservation repositories
- Copyright ownership, licensing decisions and access entitlements
- The internal operation of the Handle System or a registration agency
Characteristics
- Observed representation
- bare DOI name | labelled DOI | DOI proxy URL | shortDOI alias | ambiguous text Determines the extraction and expansion needed before identifier comparison.
- Recognition assessment
- unassessed | syntactically plausible | malformed | extraction ambiguous Prevents a recognisable string from being treated as a verified registration.
- Registration evidence
- unverified | confirmed by identified authority | absent from queried service | conflicting evidence Keeps service-specific negative results distinct from universal nonexistence.
- Responsible parties
- DOI to registration agency, registrant and current maintainer, with evidence and observation time Identifies where authoritative information and correction routes can be sought.
- Referent granularity
- work or concept | version or edition | manifestation | component | collection | other declared scope | unresolved Determines whether the DOI identifies the exact entity required by the task.
- Metadata agreement
- unchecked | consistent for task | incomplete | conflicting Controls whether metadata can support identification or needs reconciliation.
- Resolution observation
- unobserved | response obtained | identifier-not-found response | resolver failure | network failure | ambiguous result Separates observed lookup behaviour from registration and referent state.
- Destination correspondence
- unchecked | matching referent | matching status notice | ambiguous | mismatching | unreachable A completed redirect is insufficient evidence that the destination serves the intended referent.
- Observation age
- seconds since the recorded registration, metadata or resolution check Supports task-specific decisions about whether evidence needs refreshing.
- Maintenance authority
- unknown | lookup only | authorised for specified record changes | authority disputed Distinguishes the ability to inspect a DOI from permission to change its records.
Also called
Where this came from
wikidata · CC0 1.0
Drafted structure
Bundle to layer to finding to question, as the second pass will find it: 7 bundles · 13 layers · 20 findings · 41 questions.
DOI name recognition Establish the identifier represented by an observed string.
Copying, URL encoding and citation punctuation can make identifier recognition uncertain before any lookup occurs.
Name extraction
Separate the candidate DOI from its presentation context.
Candidate boundaries
Record the original text, proposed prefix and suffix, and any unresolved boundary ambiguity.
- Which characters belong to the DOI name rather than a label, URL wrapper or surrounding citation? boundary
- Could apparent trailing punctuation or an encoded character be part of the suffix? definition
Name equivalence
Assess whether differing presentations identify the same DOI name.
Comparison transformations
Record reversible extraction steps and the applicable comparison rule without treating guessed repairs as confirmed identifiers.
- Which decoding and case-comparison rules apply to these exact representations? definition
- If the input is a shortDOI alias, what full DOI does the alias service establish? provenance
- Can the agent establish equivalence without dropping potentially meaningful characters? action
Registration and responsibility Establish registration evidence and the parties responsible for this DOI.
A plausible DOI string does not establish assignment, and maintenance actions need a specific responsible service.
Assignment evidence
Distinguish an observed candidate from an evidenced registration.
Registration confirmation
Record the queried authority, environment, result and observation time.
- Which authoritative response confirms assignment of this exact DOI? provenance
- Does a negative result concern this service or environment only, or establish something broader? boundary
Maintenance responsibility
Identify the registration agency, registrant and current record maintainer separately.
Responsible service
Record evidenced responsibility and the route for registration questions or corrections.
- Which evidence identifies the current agency and record maintainer? provenance
- What service or contact accepts corrections for this DOI? action
- Does the evidence distinguish an original registrant from the current maintainer? boundary
Referent binding Establish exactly what the DOI identifies and how that identity relates to other identifiers.
An agent can retrieve a valid DOI yet use it for the wrong version, component or conceptual level.
Declared identity
Assess the entity and granularity declared for the DOI.
Referent boundary
Record enough referent description to distinguish the intended entity from neighbouring versions and parts.
- What entity does the registration metadata declare this DOI to identify? definition
- Does it identify a concept, a specific version, a manifestation, a component or a collection? boundary
- What evidence connects the object in the agent's task to that declared entity? provenance
Related identifiers
Distinguish equivalent referents from related but separately identified entities.
Identifier relationship evidence
Record claimed equivalence, version, part and other relationships with their asserted direction and source.
- Are these DOI names asserted to identify the same entity, or entities related by version, part or another relationship? boundary
- Who asserted the relationship, and what evidence supports it? provenance
- Should the agent retain distinct identities, select a task-appropriate DOI or flag a suspected duplicate assignment? action
DOI metadata evidence Evaluate the provenance and sufficiency of descriptions obtained for the DOI.
Metadata retrieval supports identification only when its source, interpretation and conflicts are visible.
Metadata origin
Identify where each descriptive assertion came from.
Retrieval context
Record the service, response format, retrieval time and whether a record is authoritative, mirrored or transformed.
- Was this description obtained from the registration agency, a resolver representation, a landing page or an aggregator? provenance
- Which schema or transformation determines how its fields should be interpreted? definition
Metadata sufficiency
Assess whether available descriptions support the intended identification decision.
Identity agreement
Record missing or conflicting identity-bearing assertions without silently combining incompatible records.
- Which assertions distinguish the referent sufficiently for this task? boundary
- Which titles, creators, dates, types or version assertions disagree across the retrieved evidence? measurement
- Can the agent proceed with qualified evidence, or must it obtain clarification from the maintainer? action
Resolution and destination Assess DOI resolution separately from the usefulness and identity of its destination.
Resolution can return associated information rather than the object itself. [DOI Handbook: Resolution](https://www-old.doi.org/doi_handbook/3_Resolution.html)
Resolver observation
Capture the context and result of a specific lookup.
Resolution result
Record the resolver, request time, requested representation, response and redirect chain where applicable.
- Which resolver and request parameters produced this result? provenance
- Did the request produce a response, an identifier-not-found result, a service failure or a network failure? measurement
- Does the task require ordinary navigation or a metadata representation supported by this service? action
Destination assessment
Inspect what the resolution result actually provides.
Destination fit
Assess referent correspondence, access barriers and status notices independently of redirect success.
- Does the destination identify the declared referent, provide a status notice, or lead somewhere unrelated? boundary
- Is the result a landing page, content, metadata, an access challenge or an error? measurement
- Should the agent follow an evidenced alternate route, retry later or report a destination problem? action
Persistence and maintenance Support continuing identification as locations, descriptions and responsible parties change.
The DOI name persists through changes in ownership or management, while maintenance responsibility may move. [DOI Handbook: Numbering](https://www-old.doi.org/doi_handbook/2_Numbering.html)
Change classification
Determine what an observed or proposed change means for this DOI's binding.
Binding continuity
Separate record correction and relocation from changes that require an agency-specific identity decision.
- Does the proposed change preserve the declared referent or identify a distinguishable entity? boundary
- Which agency or community policy governs whether this change requires another DOI? provenance
- Is an apparent withdrawal or replacement a statement about the referent, the destination or the registration record? definition
Authorised intervention
Determine permitted maintenance and verify its effect.
Record change authority
Record the actor's authority, intended metadata or destination change, and evidence needed to confirm completion.
- What authorisation permits this actor to change this DOI's specific records? provenance
- Which supported operation corrects the problem while preserving the intended referent binding? action
- Which subsequent agency response and resolution check would establish that the requested change took effect? measurement
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
A model that cannot be lined up against existing standards, identifiers and practice cannot be adopted by anyone who already uses them.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Kinds and varieties
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- journal-article DOI
- dataset DOI
- software DOI
- book/chapter DOI
- conference-proceedings DOI
- preprint DOI
- grant/funder DOI
- physical-object / sample DOI
- Which of these kinds and varieties hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Identifiers and schemes
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Wikidata - Q25670 - Wikidata item for Digital Object Identifier
- ISO - ISO 26324 - International standard for the DOI system
- DOI syntax - 10.<registrant>/<suffix> - Handle-based DOI string; prefix 10.NNNN (or longer) assigned to a registrant, suffix assigned by that registrant
- URI form - https://doi.org/10.<registrant>/<suffix> - Preferred display and resolution URL via the DOI proxy
- Which of these identifiers and schemes hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Standards and regulation
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- ISO 26324:2022 Information and documentation - Digital object identifier system (ISO)
- ANSI/NISO Z39.84 Syntax for the Digital Object Identifier (NISO, historically; syntax now aligned with ISO 26324)
- Handle System / RFC 3650-3652 as the resolution substrate (CNRI / IETF informational RFCs)
- DOI Foundation policies and Registration Agency agreements (International DOI Foundation)
- Crossref and DataCite membership terms governing assignment, metadata deposit, and persistence commitments
- Which of these standards and regulation hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Real-world use
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Cited in scholarly articles, datasets, and software as a stable link that resolves to the current landing page
- Deposited with Crossref by journal publishers so references can be matched and cited
- Minted by DataCite repositories (Zenodo, Figshare, institutional repos) for datasets and software releases
- Used as a Handle (doi:10.xxxx/yyyy) and as an HTTPS URL (https://doi.org/10.xxxx/yyyy) in metadata, PDFs, and library catalogues
- Embedded in funder and grant records (e.g. Crossref Grant IDs) and in physical sample identifiers (IGSN, now under DataCite)
- Which of these real-world use hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Typical measurements
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- DOI string length - typically 15-80 characters for the 10.prefix/suffix form; suffix length is registrant-defined - characters
- Resolution latency via doi.org proxy - typically sub-second to a few seconds under normal load - seconds
- Assigned DOIs in the system - on the order of 10^8 globally (hundreds of millions), growing continuously - count
- Which of these typical measurements hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Failure modes and hazards
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Broken persistence: DOI still resolves but landing page is gone, paywalled without notice, or replaced by unrelated content
- Duplicate assignment of the same object under different DOIs, or one DOI reused for a different object
- Prefix/suffix collision or malformed syntax that will not resolve
- Metadata drift: deposited bibliographic metadata no longer matches the object
- Tombstone or 'DOI not found' when a registration agency or member disappears without transferring records
- Confusion of versioned objects (article version of record vs preprint vs dataset version) leading to wrong citation
- Linkrot of the displayed URL if publishers print a publisher URL instead of the doi.org form
- Which of these failure modes and hazards hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Regional variation
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Registration agencies are global but market-split: Crossref dominates scholarly journals (especially North America/Europe); CNKI, JST, and others operate regional DOI services in China and Japan
- Display conventions differ: some publishers still print 'doi:' URI form; Crossref now requires https://doi.org/ display
- Legal deposit and open-access mandates in the EU and UK increasingly require DOIs on publications and datasets; practice is less uniform in regions without those mandates
- Some national libraries and repositories mint handles or URNs instead of, or in addition to, DOIs
- Which of these regional variation hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Neighbouring kinds and how to tell them apart
Reported by the breadth pass; each item needs checking against its source before it becomes normative.
- Handle (CNRI Handle System) - A DOI is a Handle with the DOI namespace prefix 10.; every DOI is a Handle, but most Handles are not DOIs and are not governed by the IDF/ISO 26324 social contract
- URL / HTTP URI - A URL locates a current network address; a DOI names the object and is resolved through a registry that can update the location
- ISBN / ISSN - ISBN and ISSN identify the product or serial title as a bibliographic identifier; a DOI identifies a specific digital (or registered) object instance and is resolvable; a book may have both
- PMID / PMCID - PubMed and PMC identifiers are database accession numbers inside NCBI; they are not ISO 26324 DOIs, though a paper often has both a PMID and a Crossref DOI
- ARK (Archival Resource Key) - ARKs are persistent identifiers with a different resolver (N2T/EZID) and no central ISO registration-agency model; persistence depends on the assigning institution
- PURL - Persistent URLs redirect via a PURL server the owner controls; they lack the DOI prefix registry, metadata kernel, and multi-RA governance
- UUID / OID - UUIDs and ISO OIDs uniquely name an entity but do not by themselves provide a global public resolution and metadata service
- Which of these neighbouring kinds and how to tell them apart hold for the sense of digital object identifier this model covers, and on what evidence? provenance
Sources
- ISO 26324:2022 Information and documentation - Digital object identifier system - Formal definition of the DOI system as an ISO standard; identifier syntax and resolution.
- DOI Handbook - Operational definition, resolution, metadata, and registration agency practice.
- DOI display guidelines and DOI facts - How DOIs are assigned and displayed for scholarly publications.
- DataCite DOI registration and metadata - Dataset, software, and research-output DOI practice distinct from journal articles.
What the second pass must settle
- Which current DOI syntax and equivalence rules cover non-ASCII characters, legacy names and ambiguous percent-encoding, and what validation procedure should this model require?
- Which registration agencies and environments must be supported, and how should their different registration states and negative responses map to the proposed evidence states?
- Which agency or community policies distinguish metadata corrections from referent changes requiring a separate DOI, especially for evolving datasets and software?
- What evidence should resolve suspected duplicate assignments or conflicting referent descriptions when authoritative services disagree?
- Which maintenance, transfer and unavailable-content policies apply across the supported agencies, and how long may observations remain usable for each intended task?