University of Twente Student Theses

Login

Patterns in data conversion

Muis, Allard (2009) Patterns in data conversion.

[img] PDF
797kB
Abstract:When a legacy information system is replaced by a new information system, the data inside the legacy system needs to be moved to the new system. This process is called data conversion. In the data conversion the data is exported, transformed to fit the new systems data structure and imported in the new system. Making a design for a data conversion is a complex and time consuming task. Designers feel the need to improve the efficiency of the design process. They would like to share and reuse experience of data conversion design in other projects. There is currently no structured way of doing this. This thesis proposes the use of design patterns in data conversion design in order to enable designers to reuse experience. These data conversion patterns describe a common, recurring problem in data conversion design and give a solution to resolve the problem that has proven to work effectively in the past. Five data conversion patterns are provided. The patterns are structured in a pattern language. This language groups patterns that give a solution for the same problem. Using this, a designer can quickly find multiple alternative solutions for a data conversion design problem. New patterns can easily be added to the pattern language. A designer needs to decide which solution fits the context of a particular problem best. In order to make this decision easier, every pattern is weighted on eight quality attributes. Depending on the context and situation, a good score on some attributes is more important than others. The quality attributes also give very clear insight in the tradeoffs of a pattern. The data conversion patterns were evaluated by a discussion with several experienced designers. They recognize the problems and solutions described in the patterns and support this solution for enabling reuse is data conversion design. Comments of the experts were used to improve the patterns further. Data conversion patterns allow designers to share design experience with each other and reuse this experience in other projects. Using patterns designers can make better, explicit design decisions based on experience recorded in the past. This will lead to improved efficiency of the design process, resulting in better design while needing fewer resources. The research presented in this thesis forms a good basis for further research. The quality attributes could form a basis of assessing design quality. The effect of data conversion patterns should be tested in an experimental setting. Expanding the number of data conversion patterns will increase the usefulness of the patterns.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science MSc (60300)
Link to this item:http://purl.utwente.nl/essays/61511
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page