We currently use the Soprano NRL model in order to expand prefix names of the ontologies. It might make more sense to let virtuoso do that job for us. We should probably also run some benchmarks to see which is faster, and by how much. Relevant Code: nepomuk-core/services/storage/* Virtuoso documentation: http://www.openlinksw.com/dataspace/dav/wiki/Main/VirtSetPrefixNames
Git commit e966c39de0c708d4e23c4e730c1946e0d5849fdf by Vishesh Handa. Committed on 15/09/2012 at 23:56. Pushed by vhanda into branch 'master'. Use Virtuoso's inbuilt prefix expansion instead of the NRLModel The NRL Model is terribly slow and consumes as high as 60% of the actual time it takes to run some queries. Arguably, these queries were small and virtuoso ran them quite fast. But it still is a lot of unnecessary time spent on modifying the query. (Based on callgrind) Since virtuoso already needs to parse the sparql query, it seems to make more sense to let it do the job. Plus, it cleans up our code rather nicely. REVIEW: 106474 M +0 -3 services/storage/datamanagementadaptor.cpp M +25 -19 services/storage/repository.cpp M +0 -4 services/storage/repository.h http://commits.kde.org/nepomuk-core/e966c39de0c708d4e23c4e730c1946e0d5849fdf