We use cookies to ensure that we give you the best experience on our website. By continuing to browse this repository, you give consent for essential cookies to be used. You can read more about our Privacy and Cookie Policy.

Durham e-Theses
You are in:

A domain-specific language based approach to component composition, error-detection, and fault prediction

Ingham, James (2001) A domain-specific language based approach to component composition, error-detection, and fault prediction. Doctoral thesis, Durham University.



Current methods of software production are resource-intensive and often require a number of highly skilled professionals. To develop a well-designed and effectively implemented system requires a large investment of resources, often numbering into millions of pounds. The time required may also prove to be prohibitive. However, many parts of the new systems being currently developed already exist, either in the form of whole or parts of existing systems. It is therefore attractive to reuseexisting code when developing new software, in order to reduce the time andresources required. This thesis proposes the application of a domain-specific language (DSL) to automatic component composition, testing and fault-prediction. The DSL ISinherently based on a domain-model which should aid users of the system m knowing how the system is structured and what responsibilities the system fulfils. The DSL structure proposed in this thesis uses a type system and grammar hence enabling the early detection of syntactically incorrect system usage. Each DSL construct's behaviour can also be defined in a testing DSL, described here as DSL-test. This can take the form of input and output parameters, which should suffice for specifying stateless components, or may necessitate the use of a special method call, described here as a White-Box Test (WBT), which allows the external observer to view the abstract state of a component. Each DSL-construct can be mapped to its implementing components i.e. the component, or amalgamation of components, that implement(s) the behaviour as prescribed by the DSL-construct. User-requirements are described using the DS Land appropriate implementing components (if sufficient exist) are automatically located and integrated. That is to say, given a requirement described in terms of the DSL and sufficient components, the architecture (which was named Hydra) will be able to generate an executable which should behave as desired. The DSL-construct behaviour description language (DSL-test) is designed in such a way that it can be translated into a computer programming language, and so code can be inserted between the system automatically to verify that the implementing component is acting in a way consistent with the model of its expected behaviour. Upon detection of an error, the system examines available data (i.e. where the error occurred, what sort of error was it, and what was the structure of the executable), to attempt to predict the location of the fault and, where possible, make remedialaction. A number of case studies have been investigated and it was found that, if applied to the appropriate problem domain, the approach proposed in this thesis shows promise in terms of full automation and integration of black-box or grey-box software. However, further work is required before it can be claimed that this approach should be used in real scale systems.

Item Type:Thesis (Doctoral)
Award:Doctor of Philosophy
Thesis Date:2001
Copyright:Copyright of this thesis is held by the author
Deposited On:01 Aug 2012 11:32

Social bookmarking: del.icio.usConnoteaBibSonomyCiteULikeFacebookTwitter