Font Size:
Constraining flexibility: moving entity relationships from structure to data
Last modified: 2011-12-22
Abstract
Early in the development of Heurist (HeuristScholar.org) we developed a strategy for building relationships between heterogeneous entitities. Those relationships - themselves stored as entities - were typed, time-stamped, annotated and directional, and could be created without limitations on the type of entitites linked or the number of links created.
This strategy has served us well in numerous projects, but it places the onus on the researcher to build sensible relationships. With the increasing use of Heurist for complex, multi-contributor resources, including excavation records and historical encyclopaedias, we have needed to move to a more controlled system of entity linkages.
In this paper we will describe our strategy for term-based relationship constraints and demonstrate how it is used in the construction of a complex database of legacy excavation data from the site of Zagora, Greece (used as the case study for our workshop on Heurist). We will also show how constrained relationships are embedded as fields in data entry forms to facilitate data entry.
We believe our strategy of term-based relationship constraints and relationships embedded in data entry forms provides a good model for building coherent networks of entities. However the system is still evolving and we will contrast it with alternatives, such as fixed relationship models defined by database structure, and invite feedback on ways it might be improved.
This strategy has served us well in numerous projects, but it places the onus on the researcher to build sensible relationships. With the increasing use of Heurist for complex, multi-contributor resources, including excavation records and historical encyclopaedias, we have needed to move to a more controlled system of entity linkages.
In this paper we will describe our strategy for term-based relationship constraints and demonstrate how it is used in the construction of a complex database of legacy excavation data from the site of Zagora, Greece (used as the case study for our workshop on Heurist). We will also show how constrained relationships are embedded as fields in data entry forms to facilitate data entry.
We believe our strategy of term-based relationship constraints and relationships embedded in data entry forms provides a good model for building coherent networks of entities. However the system is still evolving and we will contrast it with alternatives, such as fixed relationship models defined by database structure, and invite feedback on ways it might be improved.