← Back to Open Access

Contents

What are persistent identifiers?

Persistent identifiers (PIDs) are long-lasting references to digital objects of various types. An identifier is generally any label used to name an object uniquely. URLs are examples of an identifier, but also serial numbers, and personal names. But PIDs can do more than naming an object uniquely. PIDs were invented to address challenges arising from the expanding internet which is constantly changing and makes it hard to maintain a persistence record of digital objects. Often, the address of a digital object changes over time and the information is not reachable any longer, leaving ‘404’ tombstones.

PIDs are labels that locate, identify and share information about digital objects persistently without ‘rotting away’, because a PID may be connected to a set of metadata describing an object rather than to the object itself. They allow different platforms to exchange information consistently and unambiguously and thus provide a reliable way to track citations and reuse.

There are multiple PID systems for different digital objects, operated by different service providers, and used by different research disciplines. We most commonly distinguish between two varieties of PIDs: PIDs for objects (publications, data, software etc.) and PIDs for people (e.g. researchers). Some PIDs you may have already encountered are an ISBN number on a book, the DOI of a CERN Yellow Report, the identifier given for preprints on arXiv or the ORCID iD on an INSPIRE author profile.

The European Commission project FREYA developed guides that provide a general overview of different PID types for multiple research artifacts. You can consult these guides here in order to get help for choosing the right PID for your use case. The guides were last updated in May 2020.

Why use persistent identifiers?

Persistent identifiers (PIDs) are critical for building and maintaining reliable and robust links between objects, people, communities and infrastructures. This is especially critical for having reliable citations in scientific literature. Citations help researchers to verify scientific content and give credit where it is due, and in both cases, PIDs play a crucial role.

PIDs eliminate name ambiguity and point persistently on the location of a digital object. They ensure that people can be referred to unambiguously and thus solve the challenge of name changes and common names (the so-called ‘John Smith’ challenge). Credit can be assigned correctly and accumulated throughout a researcher’s career.

Without PIDs, the problem of ‘broken links’ resulting in frustrating HTTP errors remains. Even after decades, clicking on the link of an object should still resolve into useful information about what used to be there.

PIDs help to make digital objects:

  • Discoverable
  • By identifying them uniquely and reliably.
  • Accessible
  • By resolving to the specific object reliably and consistently, even if it moves to a new location.
  • Useable
  • By pointing directly to a particular version / to a specific state of an object.
  • Intelligible
  • By exposing the provenance of objects, by connecting them and thus by improving accuracy and the flow of information.
  • Interoperable
  • By providing provenance and transparency, which makes objects trustworthy.
  • Assessable
  • By an interconnected network of specifically identified objects.

For scientists at CERN, the main benefit of PIDs are the discoverability and profile building opportunities offered. For example, your profile on INSPIRE benefits from using ORCID iDs, as more of your contributions can be assigned unambiguously to your account. Also, you can receive credit for your research data or software once you have published them on services like CERN Open Data or Zenodo. In addition, discovering new literature is easier when the content is persistently available and you can actually access the literature.

For repository managers, the main benefit is the longevity and trustworthy nature of PIDs. For software developers, the main benefit of PIDs is their trustworthiness and the accompanying metadata requirements contribute towards interoperability. Additionally, PIDs help digital repositories to comply with funder requirements as minting PIDs aligns with the FAIR principles and Open Science guidelines.

Persistent identifiers for objects

There are many persistent identifiers (PIDs) for digital objects such as publications, data, and software. The most commonly seen PID systems for objects in the field of scientific information are the following:

Archival Resource Key (ARK)

ARKs are no formal standard, but all ARK systems follow the same structure (they all start with the ‘ark:’ for example) and workflows. There is also no central resolver. Organizations can sign up to become Name Assigning Authority Numbers (NAANs) and run their own resolution infrastructure for ARKs. The system is run by the California Digital Library with dozens of NAANs worldwide through a combined ARK/DOI infrastructure EZID.

Example: ark:/13030/tf5p30086k

arXiv Identifier

The document server for preprints arXiv developed their own system of PIDs and changed their identifier scheme in 2007. Until 2007, the arXiv identifier included subject-classification information, but was limited to 999 submissions per month in a single archive. As arXiv gained more popularity, they changed the structure of their identifiers. It allows arXiv to handle now more than 9999 submissions per month.

Example: arXiv:1207.7214

Digital Object Identifier (DOI)

The DOI is the most commonly used PID system for publications, data, and software across disciplines and is also broadly used at CERN. DOIs are built upon the older Handle system and were first introduced in 1998 with funding from the International DOI foundation (IDF). In 2012, the DOI system even became an ISO standard (ISO 26324).

DOIs are assigned by DOI Registration agencies. However, building upon the Handle system, there is a central free worldwide resolving mechanism for DOI names which means that DOIs are self-sufficient and their resolution does not depend on a single agency. For every DOI name, a standard metadata kernel has to be defined and the registration of DOI names at DOI registration agencies such as DataCite and Crossref typically involves a fee, but the resolution of DOIs is free.

Examples:
CERN Yellow report: 10.23731/CYRM-2019-007
Dataset on CERN Open Data Portal: 10.7483/OPENDATA.CMS.6O84.WLN8
Software on Zenodo: 10.5281/zenodo.821635
Research publication: 10.1016/j.physletb.2012.08.020

See how to get a DOI.

Handle

The Handle system is a non-commercial identifier resolution system used since 1995. It is a general-purpose global name service enabling secure name resolution operated by the Corporation for National Research Initiatives (CNRI).

A Handle identifier consists of a prefix which identifies the ‘naming authority’ in combination with a suffix, which states the ‘local name’ of an object. The location of the object a Handle refers to and metadata that describe the object are stored separately in the Handle system. A server associates metadata with a Handle, and returns that metadata when requested by a call to the Handle Service. This technique is used as the foundation of other higher level systems such as DOI. Commercial Handle licences can be obtained by research bodies and institutional repositories to establish local Handle systems, such as the European Persistent Identifier Consortium (EPIC).

Example: hdl:2381/12775

International Standard Book Number (ISBN)

The ISBN is a numeric book identifier which is intended to be unique.

An ISBN is assigned to each separate edition and variation (except reprintings) of a publication. For example, an ebook, a paperback and a hardcover edition of the same book will each have a different ISBN. The ISBN is ten digits long if assigned before 2007, and thirteen digits long if assigned on or after 1 January 2007. For more details, see ISBN on Wikipedia

Example: ISBN 978-92-9083-543-1 (paperback), ISBN 978-92-9083-544-8 (PDF)

See how to get a ISBN.

International Standard Serial Number (ISSN)

An ISSN is an eight-digit serial number used to uniquely identify a serial publication, such as a magazine. The ISSN is especially helpful in distinguishing between serials with the same title. ISSNs are used in ordering, cataloging, interlibrary loans, and other practices in connection with serial literature. For more details, see ISSN on Wikipedia.

Example: ISSN 2519-8068 (print), ISSN 2519-8076 (online)

See how to get a ISSN.

Uniform Resource Name (URN)

URNs were originally thought to be part of a three-part information architecture for the Internet just like the Uniform Resource Locators (URLs). URNs were introduced in 1994, formalised in 1997 and are now an IETF standard. URN systems are widely used by major national libraries in Europe (just like the CERN Library) as ISBNs for books are part of the URN system.

Generally, there are no licence costs involved for assigning URNs, but a URN registration agency needs to establish an assigning and a resolving infrastructure. However, there is no central governance and no central resolving infrastructure, as the URN ecosystem is very large.

Example: urn:isbn:0451450523

See more Unique identifiers on Wikipedia

Persistent identifiers for people

Authors, researchers and generally creators of works are identified in the world wide web by many different PID systems, depending on the service they are using. The most commonly used persistent identifiers (PIDs) for people are the following:

arXiv Author Identifiers

arXiv seeks to accurately identify and disambiguate all authors of all articles in their archive. Since 2005, arXiv has used authority records that link user accounts with preprints. Since 2009, public author identifiers are offered once a user publishes their authored articles in arXiv for use in other services. arXiv allows users to link their ORCID iD with their arXiv authority records.

Google Scholar Profiles

Google Scholar offers author profile pages for the identification and disambiguation of publications by a certain author. Users can create their own profile in order to track citations to their works and get additional metrics such as the h-index and the i-10 index.

LinkedIn

LinkedIn uses internal identifiers for profile pages to solve the name ambiguity problem. The ORCID iD can be added to LinkedIn profiles.

ORCID iD

The ‘Open Researcher and Contributor ID’ (ORCID iD) is a type of PID that is compatible with the ‘International Standard Name Identifier’ (ISNI). ORCID iDs were designed specifically for the research community. They are issued without cost by the open, non-profit, community-driven organization ‘ORCID’.

ORCID (Open Researcher and Contributor ID) is a non-profit organization supported by a global community of organizational members, including research organizations, publishers, funders, professional associations, and other stakeholders in the research ecosystem.

Researchers registered with an ORCID iD can use the ORCID registry’s transparent linking mechanism to pull in data from other PID platforms such as arXiv, INSPIRE, Crossref, DataCite and Zenodo to create their own verifiable research profile. Publications, funding bodies, websites, datasets etc. can be added automatically to the researcher’s profile if the ORCID iD was used.

ORCID provides an API to support system-to-system communication and authentication. The many relationships identified through the ORCiD registry result in a more satisfactory discovery process for those seeking research findings, as well as more efficient distribution of research funding and stronger collaboration within the research community.

In addition, ORCID is closely connected to other platforms and services such as INSPIRE and Zenodo, which helps you to save time and reduce the number of accounts and passwords to be remembered.

You can also use it for manuscript submissions to most of the major publishers. This means that you are able to access those platforms with your CERN credentials, since you can link your ORCID account with your CERN login.

For LHC papers in particular, the ORCID iDentifiers are being added to the XML list of authors included in the arXiv submission, automatically extracted on INSPIRE, and therefore improve the identification of individual authors in the collaboration.

It is finally important to keep in mind that several CERN member states now require that researchers provide an ORCID record in the framework of their national research assessment system.

See how to get an ORCID.

ResearcherID

The citation database Web of Science offers their own author identification service to ensure that publications are correctly attributed to authors across Web of Science collections. The ResearcherID is used to aggregate publications and track citations of a specific author. The identifier can be used with services like Publons.

See more Unique identifiers on Wikipedia

Persistent identifiers for grants (research awards)

The sources of support for research are varied and the terms ‘grants’ and ‘grant identifiers’ are used inclusively here for ‘grants, endowments, secondments, loans, use of facilities/equipment and even crowd-funding’.

The Grant Identifier initiative, led by Crossref and several funders including Wellcome Trust, JST, European Research Council and NIH, launched a registration service for grant IDs in 2019. Funders can now register research grants with Crossref. Essentially, DOIs are used in this scheme and assigned to every grant awarded. Some organizations are working on a pilot implementation of these DOIs for Wellcome Trust and some other early adopter funding organizations. Grant IDs can be added to records on Zenodo.

See more Unique identifiers on Wikipedia

Persistent identifiers for organisations

Persistent identifiers (PIDs) for organisations are useful to get an overview of the publication output of an organization and track the influence of affiliations.

ROR

Crossref, DataCite and California Digital Library and Digital Science joined forces to create the Research Organization Registry (ROR). ROR was launched in January 2019 to meet the particular need of identifying affiliations on researchers’ publications. It is the default identifier supported in Crossref DOI metadata, DataCite DOI metadata, and ORCID. ROR consolidates other identifiers currently in use (e.g. GRID, ISNI, Crossref Funder ID, Wikidata) at a coarser granularity.

CERN ROR: https://ror.org/01ggx4157

ISNI

ISNI currently provides identifiers for over 700,000 organizations in its registry. The ISNI IDs are maintained in a single database which is curated by a network of members and registration agencies.

CERN ISNI: 0000 0001 2156 142X

Ringgold

ISNI has one agency dedicated to the specific curation of organization identities, Ringgold, which manages around 500,000 ISNIs linked to its own proprietary Ringgold Org ID.

CERN Ringold ID: 30531

See more Unique identifiers on Wikipedia