ModRef Project: Modelling, References, Digital Culture

Digital Humanities - Semantic Web - CIDOC-CRM Triplestores

ModRef Project (Modelling, Repository, Digital Culture) gathers a set of projects from laboratory Labex Past in Present: history, heritage, remembrance - Labex Les Passés dans le Présent: histoire, patrimoine, mémoires (See from the University of Paris Nanterre and involves several organisms, such as:

ModRef goal is to provide Labex's projects with a digital expertise as well as figuring out a Proof of Concept (POC) concerning "linked open data" and modelling using references, in order to encourage discussions over issues related to data migration to the web semantic by creating and exploiting "triplestores" (collections or datawarehouses of RDF files). The CIDOC-CRM norm (see has been chosen since it is currently the reference for the semantic description of museographic or cultural heritage data. An OWL implementation of the CIDOC-CRM by the University of Erlangen-Nuremberg is available at the following address:

Three projects have been selected for the Proof Of Concept:

Table. Comparing data for the Proof Of Concept of ModRef
CDLI ObjMythArcheo-LIMC BiblioNum-BDIC
Languages English French-English French
Size (Texts) 300 Mo 100 Mo 100 Mo
Data Number 313 332 objects - 105 000 exposed 17 424 objects - 8250 exposed 77 collections - 62 392 files
Logical Structure Database of type spreadsheet Relational database XML-EAD
Elements number of logical structure 1 table of 61 attributes 59 tables 146 XML-EAD elements

Moving data into triplestores involves different steps:

Hence, the main issues are (1) moving from non-structured or semi-structured data (notebook, books, html) to structured data (spreadsheets, relational databases, XML files) and then, (2) moving those structured data into semantic data (RDF files) in order to improve the sharing, the exchange and the discovery of new knowledge.

On the other hand, various projects around the world work on the migration of data into triplestores (CIDOC-CRM or not), such as: