Metadata Schema for the Persistent Identification of Instruments

    You are here


Metadata Schema for the Persistent Identification of Instruments

By Rolf Krahl

Persistent Identification of Instruments WG

Group co-chairs: Louise Darroch, Markus Stocker, Rolf Krahl, Ted Habermann

Recommendation: Metadata Schema for the Persistent Identification of Instruments

Authors: Rolf Krahl, Louise Darroch, Robert Huber, Anusiriya Devaraju, Jens Klump, Ted Habermann, Markus Stocker, RDA PIDINST WG members

Impact: This PIDINST Schema is the recommended schema for the persistent identification of instruments by institutions or bodies using or managing instruments.  The Schema is PID provider independent and has been tested with ePIC and DataCite by numerous institutions including Helmholtz-Zentrum Berlin für Materialien und Energie GmbH, British Oceanographic Data Centre, EUDAT.

Recommendation package DOI: 10.15497/RDA00070

Citation: Krahl, R., Darroch, L., Huber, R., Devaraju, A., Klump, J., Habermann, T., Stocker, M., & The Research Data Alliance Persistent Identification of Instruments Working Group members. (2021). Metadata Schema for the Persistent Identification of Instruments. Research Data Alliance.


Instruments play an essential role in creating research data but they are often only identified in scientific literature using free text.  Through the use of globally unique, persistent identifiers (PIDs) such as the Digital Object Identifier (DOI), it is now common practice to establish traceable links between scientific literature and the data that generated it.  Such cross-linking has received considerable attention in the community in recent years and has been generalized to other entities, including people, organizations, funders, etc.  Given the importance of instruments and associated metadata to the assessment of data quality and data reuse, globally unique, persistent and resolvable identification of instruments is also crucial.  The Research Data Alliance Persistent Identification of Instruments Working Group (PIDINST WG) developed and demonstrated a solution for publishing persistent identifiers for instruments at established PID providers.  This solution comprises a metadata schema for the persistent identification of instruments (PIDINST Schema).


Output Status: 
RDA Endorsed Recommendations
Review period start: 
Monday, 10 January, 2022 to Thursday, 10 February, 2022
Group content visibility: 
Public - accessible to all site users
Primary Domain/Field of Expertise: 
Natural Sciences, Natural Sciences/Physical sciences, Natural Sciences/Earth and related environmental sciences
Primary WG Focus / Output focus: 
Domain Agnostic: 
  • Michael Gryk's picture

    Author: Michael Gryk

    Date: 25 Jan, 2022

    I was wondering if you had considered adding properties for the geographic position of scientific instruments?

  • Rolf Krahl's picture

    Author: Rolf Krahl

    Date: 29 Jan, 2022

    It has been considered, see the corresponding issue on GitHub. There were some issues with the idea:

    • it was not clear what properties would be needed. Would a simple geo coordinate be sufficient? For some use cases yes, for others not. What about altitude? What about orientation, e.g. the direction, the instrument points to?
    • In many use cases, where location would be relevant, the instrument is not stationary. How would updates be managed in the PID record? How do you record the history of previous positions?

    In the end, there was no satisfactory proposal to represent the location in a way that would work for all relevant use cases, so we concluded not to add it. But as an outcome of the discussion, in particular regarding the issue with non-stationary instruments, we added the relationType WasUsedIn to RelatedIdentifier, to allow to record the deployment of the instrument in a research activity. It was assumend that if the location is relavent for the use case, it could be recorded in that activity in a community specific format that would work for the particular case.


  • Baptiste Cecconi's picture

    Author: Baptiste Cecconi

    Date: 03 Feb, 2022

    The description of instrument types seems a bit limited in the document. This is a central piece of metadata for instrument discovery, but the specification only states that we should use an InstrumentTypeIdentifier, without expliciting recommended lists of values to be used. It is a selection criterion in many interfaces:

    Have there been any study of possible instrumentType lists to be recommended in this context ? 

  • Rolf Krahl's picture

    Author: Rolf Krahl

    Date: 08 Feb, 2022

    There is no single instrument classification scheme that would apply to all types of instrument being used in all scientific domains.  There are controlled vocabularies or ontologies to classify instrument types for some scientific domains.  But these are specific to their respective domain and are not applicable beyond.  The examples you point to seem to come from satelites and earth observation.  Although I'm not an expert in that field, I assume that the type classification used there is certainly useful in that doamin.  But I don't seem to find any DNA sequencers, seismometers, or transmission electron microscopes in the list of instrument types there.

    So, if you are in a domain that uses a well established instrument type classification, it is indeed recommended to use these terms in the instrument PID record.  The PIDINST metadata schema provides the means to do that and to link the terms in an unambiguous way using the instrumentTypeIdentifier subproperty of InstrumentType.  If there is no such type classification in your domain, you may need to ressort to a free text description of the instrument type.  The schema also allows that.

    The PIDINST schema needs to be applicable to all scientific domains. Therefore we cannot prescribe a controlled list of values for the instrument types, simply because such a comprehensive list does not exist.


  • Francoise Genova's picture

    Author: Francoise Genova

    Date: 10 Feb, 2022

    A follow-up to Baptiste Cecconi's comment and Rolf Krahl's answer: It would be useful that the point made by Rolf on the fact that there is no generally applicable list of Instrument Types but it is recommended to use the one(s) relevant to your domain if they exist is included in the Recommendation if it is not yet there.

  • Rolf Krahl's picture

    Author: Rolf Krahl

    Date: 10 Feb, 2022

    There is a paragraph in the introduction to Section 2 stating that the schema needs to be domain agnostic and thus cannot include community specific technical details.  The issue that we cannot prescribe a controlled list of values for the instrument type is not explicitely mentioned for now.

    I'd suggest we could add a paragraph at the end of Section 2.2 explaining that.

submit a comment