This paper focuses on management of active metadata. Examples of technical metadata are table and column definitions in a database schema, configuration files or programs used for user-interface customization, and definitions that support validation of input data.
#Yasu rule engine software
Process-related or technical (“active”) metadata support software operation. Examples of descriptive metadata are the semantic descriptions of a table's columns and controlled vocabularies used to encode clinical observations. We use the following classification of metadata 4–6:ĭescriptive metadata support human users concerned with the software's application domain. Background The Importance of Metadata Management
#Yasu rule engine code
It is also pertinent to “knowledge-base” applications, where the distinction between data and program code is blurred. Although we illustrate this approach in the clinical-trials domain, our approach is relevant to other metadata-driven applications, such as clinical patient record systems. We describe here a rule-based approach to metadata management that lets a system designer define complex constraints that prevent inconsistencies from occurring when changes are made to the metadata or automatically cascades a series of corrective or maintenance actions to maintain consistency. For such systems, it is therefore desirable to build a software layer that minimizes the possibility of incorrect metadata, in other words, a metadata management system (MMS). If, by accident, a metadata element is defined incorrectly, subtle or major malfunctions manifest in the system's operation for example, errors may creep into the data, as discussed shortly. In highly functional metadata-driven systems, the interrelationships within the metadata become complex, and metadata maintenance becomes challenging. While harder to create, metadata-driven software ultimately requires fewer modifications as the domain evolves. At runtime, software uses this information to perform various tasks such as data validation, interface generation, and customization.
One way to meet such requirements is to use a generic (“entity-attribute-value” or EAV) data model 1, 2 and an architecture driven by metadata, which is loosely defined as “data that describe data.” 3, 4 Basically, one captures information such as domain-specific descriptions, application conditions, parameters, and methods in a repository.
Database applications for the management of scientific and clinical data should be easily modifiable and adapt to scientific advances.
0 kommentar(er)
