• Digital Humanities and Cultural Heritage
• Other research productsclose
• Open Access close
• European Commission close
• EC|FP7 close
• EU close
• FI close
• CLARIN close

This report was produced in the context of the project Data Service Infrastructure for the Social Sciences and Humanities (DASISH) work package 4.3 Convergence of Data Services. The goal has been to allow the selection and promotion of high-quality deposit services for researchers in the Social Sciences and Humanities (SSH) and to make suggestions for service improvements.

The aim of this task was to analyse and compare the different metadata strategies of CLARIN, DARIAH and CESSDA, and to identify possibilities of cross-fertilization to take profit from each other solutions where possible. To have a better understanding in which stages of the research lifecycle metadata comes to the fore, we looked at several research data lifecycles and business process models. However the current research data lifecycle models have the ‘static’ data object as basis, whereas metadata design , redesign, creation and management can continue to be ‘live’ issues within the research lifecycle. We therefore developed a metadata lifecycle based closely on familiar lifecycle models but extended to support the more dynamic metadata issues. To describe the metadata management of the different infrastructures we took a double approach. We looked on a more general level and outlined the policies and strategies regarding metadata of the three infrastructures. We evaluated these strategies on metadata qua lity issues with the Bruce and Hillmann criteria. On the other hand we looked with more detail how the work on metadata management is done by the individual data repositories. The infrastructures of CESSDA, CLARIN and DARIAH differ in visions, strategies and initiatives regarding metadata issues; similarly there is a difference in metadata management among the various repositories. Despite these differences, cross fertilisation by coordination on common lists of metadata elements, sharing of knowledge, and linking resources would leverage the overall metadata quality. Evaluation of the prototype of the joint CLARIN, DARIAH and CESSDA metadata portal endorses the opinion that more coordination is needed. Metadata quality must be discussed in relation to the activities for which they are used. We suggest that the infrastructures DARIAH and CLARIN prioritise future collaboration about standardisation efforts, which have already been in itialised in dialogue between the CLARIN Standards Committee and the DARIAH representatives. Similar initiatives could be established with CESSDA.

This work presents parallel corpora automatically annotated with several NLP tools, including lemma and part of-speech tagging, named-entity recognition and classification, named-entity disambiguation, word-sense disambiguation, and coreference. The corpora comprise both the well-known Europarl corpus and a domain-specific question-answer troubleshooting corpus on the IT domain. English is common in all parallel corpora, with translations in five languages, namely, Basque, Bulgarian, Czech, Portuguese and Spanish. We describe the annotated corpora and the tools used for annotation, as well as annotation statistics for each language. These new resources are freely available and will help research on semantic processing for machine translation and cross-lingual transfer.