Author identification: The benefit of being able to identify researchers uniquely

To identify the author of a scientific document has long been an aim for many reasons: researchers studying an article want to find other papers of the same author, authors want to claim their authorship, universities want to count the number of papers of an author or to detect plagiarism and so forth. Usually the author's name is used to track the authorship, but problems arise by notation:

• authors may write articles under different names (reasons: marriage, pseudonyms, long/short name variants as in India)
• searchers may try different spellings
• publishers may use notation variants (e.g. diacritical marks or different abbreviations)

and by identification:
• papers of an author are assigned erroneously to a different author of the same name;
• name variants of an author are assumed to be different identities
• papers are spread over many different journals, publishing houses, and institutions with non-connected registries
• contact information becomes invalid when the author changes professional affiliation

The Euroscience Science Publishing Working Group therefore recommends the development of a digital International Author Identifier (dIAI) Service, as recently proposed by Raf Aerts. This service should be Open Access, with a suitable international structure and stable funding to guarantee long-term stability, integrity, and political independence.

The main obstacles for the realisation of this service are not of technical nature but are:

• national legislation with regard to privacy of personal data varies; often publication of a list of names etc. is not allowed, which has to be respected
• national pride and habits make international agreements and services difficult
• the attractiveness of the service grows with the filling of its database (but how to get to a sufficiently large database for a start?)
• commercial publishers are reluctant to open their databases for external non-profit services
• authors do not know about (or do not care for) the new useful service

Current attempts for author registries

The linking of a registered author to the documents for which he/she claims authorship is essentially a matter of trust and authenticity: The registry trusts the registrant and the reader trusts the registry. Authors in science seldom try to deny an authorship or make a false claim. Assuring authenticity means the author of a scientific work can be identified and thus properly cited and credited.

Currently, there are many parallel author registries, with more or less restricted author numbers and documents to relate to:

Commercial services as part of their future publisher's toll-access bundle of services; e.g. Blackwell's Author Services that comprise identification by email address and a list of the author's articles in any Blackwell journal (see demo).

A Free Access service may attract the widest possible number of authors. Making use of the ever-increasing Open Access space of documents can lead to an attractive service. One example is AuthorClaim running the Academic Contributor Information. The software implements the idea of author-driven identification. A precursor has been running for the RePEc digital library since 1999. Authors describe themselves through their name, workplace, homepage, and some papers. AuthorClaim then searches its document database for documents with one of the author's name variations and offers them for approval/denial of authorship, thus filling the publication list and the denial list. AuthorClaim is intended to become an interdisciplinary system. It includes already PubMed, CiteSeer, DBLP and E-LIS. Authors cannot upload their own documents or create new document descriptions. It is thus a complement to existing digital libraries, not a competitor.

In Citation Upload services, authors can upload the citation data of their papers and are rewarded by a list of their uploaded citations with metadata in various formats. These services benefit from the growing Open Access world. As an integral part, they should contain an author registry and identification service. An example is citeulike, Everyone's Library, an Open Access service for citation management, founded by Dr. Cameron, now sponsored by Springer. It comprises an author registry, powerful add-on-services such as uploading of authored paper citations, picking own paper citations from a large number of commercial/non-commercial databases, connection, and alerting services. Thus an author can build up a pretty comprehensive publication list and export it as bibtex.

Citation services on top of networks of (Open Access) repositories: The German Research Foundation DFG is funding a new project developed and served by the DINI German Initiative for NetworkInformation e.V., the OAN Open Access Network, that will serve as synergistic add-on service for Open Access repositories. An integral part of it is to install the service Distributed Open Access Reference Citation Services (DOARC, see its demonstrator).

There are also various research field-oriented author services: PhysNet (since 1996), the Physics departments and documents network, offered by the European Physical Society EPS, developed and operated by ISN, is served by a worldwide set of Physics institutes and offers a coordinated bundle of services (department address geographical coordinates, document lists, search for persons, documents, topics, journals, jobs, e-learning, etc.). An analogous service for mathematics is Math-Net, also an International Information and Communication System, served by the IMU. Part of it is Persona Mathematica. The Federated World Directory of Mathematicians uses its data as part of its input. A newer AI service in Germany is from BID , installed as a wiki. It comprises links to professional homepage information of presently about just 190 persons of the Library / Information / Documentation area.

National registry actions:

• the recently launched Norwegian Frida service, linking Human Resources data to bibliographic data
• the Registry for the German Dissertations at Universities DissOnline of the German National Library which can be used as an identification handle for authors, provided they made their Ph.D. in Germany
• the Dutch NARCIS National Academic Research and Collaborations Information System, funded by the Surf Foundation; it uses the national database of research (METIS), giving most Dutch researchers a Digital Author Identity number
• GEPRIS, a registry for authors involved in projects of and served by the German Science Foundation DFG

A pragmatic approach

Technically, a tool for unambiguous identification of researchers would in its most simple form come along as a database with at least two kinds of fields in each data set: (alpha-)numeric author identifier ("IAI number") and name variants. To cover all languages and scripts used for publication worldwide, the database should be set up with the Universal Character Set (ISO/IEC 10646, Unicode).

With respect to the author’s publication list and contact details (e.g. professional e-mail address and web page), the dIAI service could follow the general approach of the early Math-Net / PhysNet concept: let the information remain locally at the professional institution of the author – and let them take care of updating it – but have them add internationally agreed metadata (the author's IAI number) which can then be grazed by national, international, or field-specific search engines.

The filling of the dIAI database could start with a single paper of an author who is contacted through the mentioned e-mail address and then allowed to claim the authorship of other papers by connecting the assigned IAI number to each publication on an on-line list. Researchers who have already experienced name-related disadvantages would make excellent first movers. The database itself would be linked to a small search engine that retrieves a list of all matching name variations (and some publications) for each name query in the first round. In the second round, the user selects one match and receives the corresponding IAI number that can be used with the appropriate search engines in order to retrieve a synoptic publication list. By using the IAI number in national, research field, or publishers' author identification services when adding or confirming authorship of a paper, the author allows for a "holistic" listing of publications or citations, regardless of them being spread over several providers.
The dIAI system would increase the performance of add-on services, such as keeping an author's publication list complete and up to date, generating citation and download statistics, finding related papers of other authors, etc.

The organisation of the dIAI would be best achieved by an international organisation defining the metadata, providing a registry for the IAI numbers assigned by the multitude of worldwide distributed learned societies and research institutions, and detecting and removing double countings. Thus ,the heavy load of keeping the entries up to date stays with the local research entities. This organisation scheme follows the well-proven one of PhysNet.

Summarising, international author identification in science is a benefit for effective research work, especially in linguistically fragmented Europe. Therefore Euroscience should engage to initiate its technical realization, international organisation, and support.

About the authors

Eberhard R. Hilf, CEO; ISN Institute for Science Networking Oldenburg GmbH at the Carl von Ossietzky University, Germany

Bernd Kappenberg; biochemist and MA in European Studies; Product Stewardship Counsellor, Cefic aisbl (European Chemical Industry Council), Brussels, Belgium

Hans E. Roosendaal; professor for strategic management, Dutch Institute for Knowledge Intensive Entrepreneurship (NIKOS), School of Management and Governance, University of Twente, the Netherlands