Denormalizing for speed
Adhering to the rules of normalization is a good thing. Unfortunately, normalization groups your data for correctness, but not necessarily for functional access. Often you find that performance suffers when you've taken your database to a fully 3rd-normal form. This presentation, last given in 1996 at IUC (and possibly given at a prior conference - I can't remember), demonstrates a safe form of denormalization within a Synon/2E environment. The techniques used focus on preserving data integrity and moving the high cost of data access to the data maintenance phase rather than incurring that cost during the data retrieval phase. Note that these techniques are best suited for low maintenance/high retrieval scenarios. I strongly encourage you to understand the concepts behind the techniques, weigh the benefits and costs in your own situation, and then apply these techniques judiciously.
In PowerPoint 97 format.
