La mise en œuvre des principes FAIR par le consortium MASA présenté à DH2018, Mexico

Le consortium MASA à participé au colloque internationale Digital Humanities 2018 à Mexico du 24 au 29 juin. Adeline Joffres, responsable de la coopération internationale à Huma-Num, a organisé pour la TGIR un panel inter-consortiums intitulé The Impact of FAIR Principles on Scientific Communities in (Digital) Humanities. An Example of French Research Consortia in Archaeology, Ethnology, Literature and Linguistics. Les consortiums Archives des ethnologues, Cahier, Corli et MASA y ont participé.

La présentation de MASA s’intitulait What is the Cost and the Efficiency of Exposing Archaeological Data in the LOD? en voici le résumé :

There is a very great disparity in the organization of archaeological data and their management and archiving systems, when they exist. The immediate consequence of this state of affairs is the risk of the irreversible loss of a significant amount of inaccessible archaeological data. The MASA consortium has therefore set itself the objective of digitally sharing archaeological data by proposing guidelines to the archaeological community. There is an urgent need both to safeguard existing archival collections and ensure the re-use of old databases, and to ensure that emerging new databases use sustainable systems that will provide interoperability and the long-term reuse of data. The consortium’s work therefore focuses on the classification and digitization of old archaeological archives, the documentation of its collections according to an appropriate structure, the re-use of old databases, the alignment of vocabularies used with standards, the matching of archaeological information systems with the domain ontology for cultural heritage (CIDOC-CRM, ISO 21127:2014), on the online publication of archives, data and syntheses linked with data.

The difficulties to be overcome are many, depending on whether structured databases or more informal batches are being processed, and must reconcile various situations such as:

  • – The digitization of old archives (often resulting from the work of only one archaeologist) which must be classified, safeguarded, made available and documented according to the archaeological value added. This involves producing an overlay to compensate for the absence of a data structure.

  • – The transformation of old databases developed with systems and in formats that are no longer accessible or disappearing, in order to preserve the data themselves and their structure when it exists.

  • – The interoperability of old, structured and still maintained systems, whose redeployment in standard and open formats is necessary but beyond the means available. While this may seem simpler, it is not the case and care must be taken not to delay those who have had advance notice.

  • – The development of new systems which must be designed directly in accordance with existing interoperability standards so that they do not have to be rethought in mid-term.

All the experiments carried out show the heuristic value of the operations necessary for the digital sharing of data, which is a reflective step in terms of both content and information structure. However, the final quality of the information shared according to these four processes varies. The creation of metadata on loosely structured corpora is a definite added value but never achieves the finesse of description of structured information systems. In addition, the use of metadata description standards alone does not offer the semantic enrichment that can be achieved by mapping with reference ontologies, which constitutes a production of knowledge in itself. In fine, exposing archaeological data in the LOD will help to build bridges with other heritage data but also with other themes.

These different processes and their consequences will be explored using a few examples.