The use of the so-called Digital Object Identifier (DOI) suggests the existence of "digital objects". However, in developing the concept of the DOI, a notion dominated by the technical dimension seems to have been the prevalent factor. The DOI is mainly about being able to find and address a resource that exists on the Internet and has a URL, even if the domain of the URL has changed. Yet, it is not associated with a precise or even binding definition of what a digital object is. Thus, the resources that can be accessed via URLs and DOIs are extremely heterogeneous in terms of their content. The only thing that all corresponding instances have in common is that they are static or ad hoc data or files that are consistent from a technical point of view and can therefore be interpreted unambiguously, called up, displayed and/or downloaded via browsers.

However, to create true interoperability, it is essential to add a content dimension to the technical dimension of the concept DIGITAL OBJECT. Moreover, this content dimension must be subject to a binding structuring system that would be roughly comparable to the controlled vocabularies used, for example, by libraries to index the content of analogue and digital holdings. Specifically, it would have to be clearly outlined which entities can be given the status of a digital object, what the logical internal structure of these objects should look like and which naming scheme should be used. An essential prerequisite for the useful application of a corresponding concept would certainly be the finest possible granulation of datasets. Only then is effective networking in the sense of genuine interoperability guaranteed, making the corresponding measures a supporting pillar in the implementation of the FAIR criteria.

The development and implementation of a corresponding system or procedures based on it cannot be achieved by individual projects. Rather, decisions on science policy are essential here, which could be made, for example, by the Council for Information Infrastructures (RfII). The currently (2020) emerging NFDI consortia could also work towards the development and establishment of such conventions.

As a trial, VerbaAlpina will define the morpho-lexical types and concepts collected in the project, together with their associated metadata, as digital objects. For this purpose, the respective type and concept data are stored in individual text files that can be accessed via URLs and subsequently also DOIs. These files will also be collected by the university library (UB) of the LMU. In addition, the data is recorded in schematised metadata formats (DataCite for generic metadata, CIDOC-CRM for in-depth content indexing), which will eventually lead to the retrievability of the digital objects via the library catalogues as well.


Literature: Hui 2012; Hui 2016; Krefeld 2018b .