Section: New Results

Managing Dynamic Linked Data

Participants : Angela Bonifati, Denis Debarbieux, Joachim Niehren, Tom Sebastian.

Bonifati, Goodfellow (former PhD student at the University of Strathclyde, UK), Manolescu and Sileo (former PhD student at the University of Basilicata, Italy, directed by Bonifati) studied XML view maintenance in the presence of updates [1] . Their approach relies on algebraic operators for propagating source updates to the target XML view, e.g. in a typical scenario of GAV (global-as-view) schema mappings. Their algebraic approach is set-oriented as opposed to tuple-oriented methods presented in the literature. Moreoever, it leverages structural identifiers and structural join algorithms. As such, it proved to be more efficient than existing methods for updating materialized XML views.

Debarbieux, Gauwin (former PhD student in the team, now Assistant Professor at the University of Bordeaux), Niehren, Sebastian and Zergaoui (CEO at Innovimax) focused on using early nested word automata in order to approximate earliest query answering algorithms for nested word automata in a highly efficient manner [9] . This approximation can be made tight in practice for automata obtained from XPath expressions. An XPath streaming algorithm based on early nested word automata has been implemented in the FXP tool. FXP outperforms most previous tools in efficiency, while covering more queries of the XPathMark benchmark.