The COEL database and database software, a combined reference and research tool created by historians for historians, is presented here through Screenshots illustrating the underlying theoretical model and the specific situation to which that has been applied. The key emphases are upon data integrity, and the historian's role in interpreting and manipulating what is often contentious data. From a corpus of sources (Level 1) certain core data are extracted for separate treatment at an interpretive level (Level 3), based upon a master list of the core data (Level 2). The core data are interdependent: each record in Level 2 is of interest in itself; and it either could or should be associated with an(other) record(s) as a specific entity. Sometimes the sources are ambiguous and the association is contentious, necessitating a probabilty-coding approach. The entities created by the association process can then be treated at a commentary level, introducing material external to the database, whether primary or secondary sources. A full discussion of the difficulties is provided within a synthesis of available information on the core data. Direct access to the source texts is only ever a mouse click away. Fully query able, COEL is formidable look-up and research tool for users of all levels, who remain free to exercise an alternative judgement on the associations of the core data. In principle, there is no limit on the type of text or core data that could be handled in such a system.

Relational Database Systems currently dominate the marketplace and thus the use of Database Management Systems by historians. This technology can constrain the thinking ofits users and limit the representational and analytical power of the applications built using it. This has led to research into other models of Database Management which are less restricting. An introduction toone ofthese approaches, Object OrientedDatabase Systems, ispresentedhere in a historical context with the purpose ofillustrating its power for historical research. An example of this power is given with a description of the authors research into the development of a workbench system utilising Object Orientedprinciples for Nominal Record Linkage.

This article considers the fundamental differences between two emergent approaches to historical database work, the ‘model-oriented’ , traditional, results-driven, relational database approach, which is quite clearly what the majority of historians want, need and will continue to adopt, and the ‚source oriented’, ‚database as edition’ way exemplified by Kteioi. These approaches are seen as legitimate alternatives belonging to different approaches to historical research. Their relative strengths and weaknesses are discussed, including the issues ofsuitability to particular types ofresearch, functionality, userfriendliness, relative learningcurves andresults. The article ends by assessing future prospects, particularly as exemplified by current 'user-oriented' software development in the Netherlands.