Database reverse engineering (DBRE) attempts to recover the technical and semantic specifications of the persistent data of information systems. Despite the power of the supporting tools, the discovery of a single foreign key through the analysis of half a million lines of code may require several hours of meticulous analysis. Considering that an actual schema can include several thousands of implicit constructs, DBRE can prove very costly for medium to large projects.
Even in the parts of the process that can be automated, tools must be used with much caution. While they can help the analyst, the latter needs to understand their strength and weakness.
On the other hand, reducing the quality requirements of the result can lead to much higher costs when the recovered conceptual schema is used as the basis for reengineering the information system or to migrate its data to datawarehouses or to other applications.
Hence the dilemma: how can we automate the highly knowledge-based and interactive DBRE process without impairing the quality of the resulting products'
|Title of host publication||Proc. of Data Reverse Engineering Workshop 2000 (DRE'2000)|
|Publication status||Published - 2000|