Thresholded Debugging of XPath Queries

Abstract

We have recently designed/implemented a method for debugging Fuzzy-XPath queries which produces a set of alternative Fuzzy-XPath expressions with higher chances for retrieving answers from XML files. The main goal of the present paper consists in the introduction of a new fuzzy command inside the Fuzzy-XPath debugger which comfortably relies on our implementation based on fuzzy logic programming. So, when <<[FILTER=r]>> precedes a fuzzy query the debugger lazily explores an input XML document for dynamically disregarding as soon as possible those branches of the XML tree leading to irrelevant solutions (i.e., with a chance degree degraded below r), thus allowing the possibility of efficiently managing large files without reducing the set of answers for which users are mainly interested in. Hence, advice that this dynamic thresholding technique embedded into the core of the Fuzzy-XPath debugger has two advantages: • firstly it permits to concentrate on significant answers (i.e., alternative queries which do not excessively deviate from the original one) without disturbing the attention with useless information, and • secondly, the computational behavior of the debugging process is highly improved (both in time and space) since a great amount of work is avoided when discriminating useless branches of the XML tree.