Processing incremental multidimensional range queries in a direct manipulation visual query environment

We have developed a MultiMedia Visual Information Seeking (MMVIS) Environment designed to support an integrated approach to direct manipulation temporal querying and browsing of temporal relationship results. We address the optimization of queries specified via our visual query interface. Queries in MMVIS are incrementally specified and continuously refined multidimensional range queries. We present our k-Array index structure and its bucket based counterpart, the k-Bucket, as new indexes optimized for processing these direct manipulation queries. In an experimental evaluation comparing our k-Array and k-Bucket solutions to alternate techniques from the literature, we show that the k-Bucket performs generally equal to or better than the other techniques and is the best overall approach for such environments.