The preliminary experiments show that the incremental approach can deal with an active subset consisting of 100,000 records with 10 attributes each [13]. In comparison, the pioneering work in the area, the Film Finder program [1], could handle a database of 10,000 records with 10 attributes, and some of the standard data structures analyzed in [10] and tested in [11] demonstrated scalability up to 20,000 records with 10 attributes.
The following subsections describe the detailed experimentation (on an implementation made by using the incremental approach) and show the results of the experiments. First, we describe the implementations and the environment for the experiments. The experimental method and the results are presented next. After this we show the derivation of the experimental run time behavior (complexities) obtained from the experiments. Then we test the validity of these experimental complexities and finally state some conclusions.