FAIR Principles – review in context of 4TU.ResearchData

The Dutch archive for the technical sciences - 4TU.ResearchData - has recently analysed itself in the context of the FAIR (Findable, Accessible, Interoperable and Re-usable) data principles. Below is an overview of the metadata that describes each dataset according to the following attributes: FAIR Principles (as scoring matrix) - 4TU.ResearchData  PolicyGeneral Comments


   To be Findable: 



4TU.ResearchData  Policy

General Comment

(meta)data are assigned a globally unique and eternally persistent identifier.

Yes, DOIs are employed for each dataset

To the extent that anything is eternally persistent

data are described with rich metadata

Yes, multiple metadata fields are included

Rich metadata is a vague term

(meta)data are registered or indexed in a searchable resource

Yes, our data is crawlable crawlable and we provide OAI-PMH sets (eg., NARCIS , Thomson Reuters DCI) as well as (part of our) present (part of our) to DataCite Metadata store. There’s also a (hidden) SPARQL Endpoint.


metadata specify the data identifier


Slightly opaque English ‘metadata include the data identifier’ perhaps

To be Accessible:





4TU.ResearchData Policy

General Comment

(meta)data are retrievable by their identifier using a standardized communications protocol.

Yes, http is used


the protocol is open, free, and universally implementable.

Yes, http is open etc


the protocol allows for an authentication and authorization procedure, where necessary.

Yes, users need to authenticate themselves


metadata are accessible, even when the data are no longer available.

Yes, this is part of our policy

This principle seems to be policy driven rather than technical and sits a little bit awkwardly in this section

To be Interoperable: 





4TU.ResearchData Policy

General Comment

(meta)data use a formal, accessible, shared, and broadly applicable language for knowledge representation. *

Yes, data uses elements of DC, and can also be exposed as ORE RDF/XML

If you are not a metadata expert this language is opaque

(meta)data use vocabularies that follow FAIR principles. 

We use established vocabs where we can, but they might not follow the FAIR principles. If we develop our own, we publish the ontology.

The FAIR facet is a bit vague.

E.g. we use ORCID, which can be seen as a vocabulary.

We do not have an ontology or controlled vocabulary in place yet.

(meta)data include qualified references to other (meta)data.

Sometimes, eg links to publications or ORCID identifiers (orcid is idenfitier)

This is vague. Does it means links to other vocabs and thesauri ?

   To be Re-usable: 


4TU.ResearchData  Policy

General Comment

meta(data) have a plurality of accurate and relevant attributes

Yes, we employ many of fields

How is this different from Principle F2?

(meta)data are released with a clear and accessible data usage license. 

Yes metadata is released with a CC0 licence. But for data, we currently have our own bespoke licence. We are working on changing this.

This is a worthy aim, but difficult to achieve without much more policy development (for us)

(meta)data are associated with their provenance.

The source of data is included in the metadata records, but does not display the file processing and how the final data was created.

It’s difficult to display this type of provenance information in metadata. If free-text documentation counts as metadata than we could say we meet the principle

(meta)data meet domain-relevant community standards

Partially. Difficult to have subject specific metadata when we cover so many different subjects. However, some data formats are tailored for particular domains.

This is difficult to achieve for non subject-specific repositories, especially for descriptive metadata


Source: FAIR Principles - review in context of 4TU.ResearchData 

Related content: 










Add comment

Log in or register to post comments