As alternative I'll to try to explain the results obtained by looking more closely at the algorithms implemented. It is relatively easy to create an algorithm that gives good results for questions in a small domain. General purpose algorithms are a lot harder to come by. By comparing not only the results but also the way we obtained them we should be able to filter out the more general solutions. This will also point us to shortcomings that would remain invisible if we were only to compare results. In the next section potential improvements are presented based on these shortcomings.