OOD for Linguistic Systems
An exploration of the validity of object oriented design techniques in modelling linguistic systems
Abstract
Object orientated design (OOD) lies at the heart of almost all computer software development. Describing and constructing components as objects with properties and methods gives an extensible model which encourages a normalised solution. Implicit in the concept of normalisation is the assumption that the system is understood sufficiently to allow the designers to remove duplication and identify the properties and methods of the key components. Methodologies exist such as UML1 to encourage properly normalised design. Unlike typical software developments, linguistic systems are at best a partial model of a highly complex system. This complexity militates against successful normalisation. Conversely, the intuitive nature of OOD for components encourages the development of objects which encode the nature and behaviour of the entities they model. Understanding how such objects can be designed within the many interrelated contexts of language is key to the long term success of computational linguistics.2Beginning with the identification of macro linguistic contexts and the relationship between them, this proposal seeks to construct a relationship framework within which linguistic entities may be successfully modelled using OOD techniques.
1 Ambler, Scott William (2004). The Object Primer: Agile Model Driven Development with UML 2. Cambridge University Press
2 Sascha Konrad – Betty H.C. Cheng Automated Analysis of Natural Language no. 3844 in Lecture notes in computer science pp48-56 Springer Verlag 2006
Abstract
Object orientated design (OOD) lies at the heart of almost all computer software development. Describing and constructing components as objects with properties and methods gives an extensible model which encourages a normalised solution. Implicit in the concept of normalisation is the assumption that the system is understood sufficiently to allow the designers to remove duplication and identify the properties and methods of the key components. Methodologies exist such as UML1 to encourage properly normalised design. Unlike typical software developments, linguistic systems are at best a partial model of a highly complex system. This complexity militates against successful normalisation. Conversely, the intuitive nature of OOD for components encourages the development of objects which encode the nature and behaviour of the entities they model. Understanding how such objects can be designed within the many interrelated contexts of language is key to the long term success of computational linguistics.2Beginning with the identification of macro linguistic contexts and the relationship between them, this proposal seeks to construct a relationship framework within which linguistic entities may be successfully modelled using OOD techniques.
1 Ambler, Scott William (2004). The Object Primer: Agile Model Driven Development with UML 2. Cambridge University Press
2 Sascha Konrad – Betty H.C. Cheng Automated Analysis of Natural Language no. 3844 in Lecture notes in computer science pp48-56 Springer Verlag 2006
You must be logged in to make comments on this site - please log in, or if you are not registered click here to signup
