The MSE community makes great steps towards findable, accessible, reusable, and interoperable (FAIR) data. In the context of FAIR data, giving digital things a digital identifier in the form of a uniform resource identifier (URI) [1] is part of the best practice. A well-known scheme for such URIs is the digital object identifier DOI [2] which can in principle be used to identify any digital object on the web but is best known for identifying digital documents. A DOI provides a unique and uniform resource identifier to a digital object. Through the dereferencing service operated by the International DOI foundation, an easy access to the digital objects is achieved while providing full data sovereignty to the owner of the digital object. A drawback of the DOI is that it is not well suited for human recognition.

We propose the digital material identifier DMI, which adopts the advantages of DOIs while providing a friendly human readable identifier scheme for materials that are described by already existing specifications. Such specifications could be standards, agreements, reports, datasheets, and others. The first version of the material identifier is built on top of the Persistent Uniform Resource Locator (PURL) service [3]. PURL is an established tool in the linked data, semantic web, and ontology community to assign persistent names to resources that may move their location over time. Furthermore, the digital material identifier relies on the W3C recommendations for cool URIs [4]. The form of the digital material identifier is:

http://purl.org/dmi/{issuer_type}/{issuer_acronym}/{specification_acronym]/{material_acronym}

where {issuer_type} distinguishes between standardization bodies, associations, academic projects, and other issuer types. The {issuer_acronym} is chosen at time of registration of the issuer to the DMI service while {specification_acronym} and {material_acronym} are assigned by the issuer and should reflect well known labels for the specification and the material respectively. For example, a DMI for the aluminum material EN AW 2618A according to DIN EN 573-3:2019-10 would then look like this: http://purl.org/dmi/SB/DIN/EN_573 3_2019 10/2618A. This form is compact, unique, and human- and machine usable at the same time. By accessing this URL, the request would be redirected to an URL at the choice of the issuer, preferably a website that gives additional details about the material. The level of detail displayed on the issuer’s website should vary with the subscription level of the caller and support content negotiation to serve human clients as well as machines.

The DMIs benefit from the immense effort (and in the case of standards also from the community agreement) that has been put into formulation of the specification while leaving the intellectual property rights of the specification in the hands of the issuers. Such identifiers will be of great benefit for datasets dealing with the description of materials and for datasets making reference to materials. Furthermore, the DMI shows the pathway on how to extend identifier concepts that are based on existing specifications to a broader scope: http://purl.org/mci/SB/DIN/EN_ISO_6892 1_2020 06/R_m will be easily recognizable to the domain expert as the tensile strength $R_m$ from a room temperature tensile test performed in accordance with DIN EN ISO 6892-1:2020-06. The main upcoming challenge now is to involve the community for the adoption of DMIs as formal identifiers for materials.

[1] https://www.rfc-editor.org/info/rfc3986
[2] http://doi.org
[3] http://purl.org
[4] https://www.w3.org/TR/cooluris/