Attributes sanitization in object-oriented design to improve database structure

Modelling using the Entity Relationship Model was introduced more than thirty years ago. Until late 1990’s, object-oriented introduced class diagram. However, designing a good database is still a serious issue. Some of the issues are very difficult to handle such as consistency checking between system design and database design, redundancy of data, mismatch of the data structure with the user’s needs in the database and unused data in the database. In this thesis, a new technique called UInData is introduced as an alternative method for designing database based on attribute sanitization. The proposed technique will extract class behaviour from class diagram to produce schema table which will then be compared with the user interface to normalize the structure. Attributes sanitization is introduced to remove the unused attributes and to provide final schema table. An experiment using three case studies has shown that some improvements of designing optimal database have been achieved in term of data sanitization and data accessibility. Attribute sanitization was applied in LAS, SPKS and MPBP database. Data sanitizations have removed 2.2%, 14.1% and 24.5% from defined attributes which are not used by user interface. Meanwhile the results shown in data accessibility for these three cases have shown that LAS was reduced by 50%, SPKS have not reducing of data accessibility and MPBP was reduced by 20% when UInData is used as compared to using ordinary object-oriented. Therefore, the UInData is a good alternative technique to improve database structure

Name dc:type.qualificationname

phd

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

Universiti Tun Hussein Onn Malaysia

Year dc:date.issued

2014

Author dc:creator

• Saringat, Mohd Zainuri

• QA76 Computer software

Language dc:language

en