Debugging while interpreting fuzzy XPath queries

We have recently introduced “dynamic thresholding” techniques into our debugger of XPath queries which produces a set of correct XPath expressions with better chance degrees for retrieving answers from large XML files in a very efficient way. In this paper we focus on a new command called DEBIN intended to automatically interpret all these correct queries for the retrieval of their answers. The interest of the new command resides in the fact that users can retrieve now new information not necessarily reported by the execution of their initial queries, thus collecting useful novel answers (very often accompanied with a greater “retrieval status value” or satisfaction degree) associated to correct queries which slightly deviate from the original ones. In this paper we justify why, apart for automatically removing redundant solutions and sorting them, the use of appropriate filters seems to be mandatory when managing queries with the DEBIN command, in order to maintain efficiency and to avoid the generation of useless knowledge, since both the set of correct queries produced by the debugging process, as well as the set of answers obtained after interpreting them, can be enormous. Additionally, we justify why DEBIN should retrieve answers of correct queries, with worse chance degree but better satisfaction degree.