Draft of ENRESSH Minimum Data Model

Mandatory elements

Justification: As it's the base element for all the other elements it should be mandatory to provide.

Disclaimer: High variation on source systems in terms of publication inclusion criteria, e.g. are non-scholarly publications included as well (professional and general/popular books, articles, reports etc.) or are conference presentations or short abstracts included. To some extent this can be traced by the type -element and what kind of types are included. Scope of what is actually recorded as a publication could be provided by for example provenance information in about element or otherwise controlled/validated in metadata exchange. Limitations to inclusion criteria are not recommended though, as comparability and analysis of records are more dependent on e.g. publication types to be consistent.

Justification: For technical purposes e.g. harvesting and updating the record, a unique id is needed for the metadata to be traceable.

Disclaimer: None?

Justification: Metadata should be provided on the publication type to have at least some structure and information on the publications. This allows the data to be comparable and used as a basis for analysis and research.

Disclaimer:  Source systems almost never have the exact same semantics / classifications for publication types e.g. what is included in a review article type might differ based on the source system definition. Mapping publication types from data model to another always leads to loss of information. Three different approaches can be had on the mapping (1) try to achieve the most accurate possible mapping between data models (2) use the so called lowest denominator of publication types; a limited set of broad categories (3) use lowest denominator publication types, but include refinements where possible and use the broad categories externally.

Justification: In terms of proper identification of the publication, a publication title should be provided. Although in theory, this could be enriched from elsewhere, but for accurate information it should be provided from the source systems to get title in the original publishing language. Data exchange can be done using multilingual strings to provide information on alternative languages as well.

Disclaimer: As publication title is provided in a free form text string, there is no guarantee that the title would match either the original publication title or title provided in other databases for the same publication.

Justification: For publishing date a year of the publication is needed for comparison and analyzing purposes. This allows the data to be e.g. analyzed between time periods.

Disclaimer: Publication year is not always definitive, but it should be based on the time when publication was published as a version with full bibliographic information available.

Justification: Author information is highly relevant to be able to both identify and analyze the bibliographic metadata for research purposes. Without author data, there would be no ties to the persons and a research dimension, that are the people behind the publications, would be lost. To collect this information on the source systems, the metadata quality should be high as the different

Disclaimer: As the format and spelling of names might vary heavily between source systems, the use of identifiers e.g. ORCID is highly recommended. But as many researchers are still missing the unique identifiers or the systems are not supported by it, the author information should still be provided in text string format as well.

Justification: Organizational author information is relevant as it ties the publication via affiliation information to a certain organization. Thus, an analysis or research can be made using the affiliation information. It also provides better legitimacy for the publication as the affiliated organization is known. The information about organizational units should be made as close to the source system as possible and it should be made mandatory for that reason.

Disclaimer: At this time there is no reliable identifier in place for organizations. By providing information in both human and machine readable formats, there would still be no up to date information on e.g. organizational units as they tend to change as time passes and organization structures change.

Justification:

Disclaimer:

Optional elements

ENRESSH-VIRTA pilot

Core

• Publication title
• Publication year
• Authors
  • Organization authors / affiliation
• Publication type
• Field of science

Optional (co-publication identification)

• ISSN/ISBN
• Source title
• Volume
• Number
• Page numbers

Supporting documents

VIRTA European pilot - the data contents

https://openaire-guidelines-for-cris-managers.readthedocs.io/en/latest/cerif_xml_publication_entity.html