A few months ago, we identified a major performance degradation issue with Longwell and we fixed it (basically one way of querying the triple store with another).
Unfortunately, the two queries are not idempotent in all situations and this resulted in a seriously broken Longwell functionality. Mark Wilkinson unfortunately ran into this problem. I've spent the entire day to find where the problem (performing binary search backtracking on all the longwell revisions to understand what broke it) was and fixed it in the latest Longwell trunk in svn. Please do an 'svn update' to get the latest fixes and perform a 'mvn clean' before running './longwell' again. There is still a major issue in longwell about facets that are mostly numeric but have some intrusive non-numeric values (such as Mark's dataset that he gave me access to) that end up completely empty and without any warning. We've been hit by that before and we'll fix it real soon. Hope this helps. -- Stefano Mazzocchi Digital Libraries Research Group Research Scientist Massachusetts Institute of Technology E25-131, 77 Massachusetts Ave skype: stefanomazzocchi Cambridge, MA 02139-4307, USA email: stefanom at mit . edu ------------------------------------------------------------------- _______________________________________________ General mailing list [email protected] http://simile.mit.edu/mailman/listinfo/general
