Hide metadata

dc.date.accessioned2013-03-12T08:02:58Z
dc.date.issued2006en_US
dc.date.submitted2006-09-08en_US
dc.identifier.citationSand, Gunhild. The evaluation of UML-based development methodology in a large project. Hovedoppgave, University of Oslo, 2006en_US
dc.identifier.urihttp://hdl.handle.net/10852/9529
dc.description.abstractUML and UML-based development methods have become a de facto standards in the industry when modelling object-oriented systems. Mostly the literature covers examples on developing relatively small systems, but it is believable that certain problem aspects in UML-based development only show up in projects at a certain size. A project with many co-workers for example requires cooperation among many teams, and changes in existing code will often be an activity in projects producing large amounts of code. Within UML-based development methodologies use case modelling is an important activity starting early in the development effort. Use cases specify what a system must do from the users' perspective. When used appropriately they provide a means to validate stakeholder concerns early in the project. They are not only a requirements tool, but can drive the definition of the system architecture and the development and delivery of the final system. The literature reports problems with the technique, often resulting in use cases that are hard to understand. Such problems can result in ambiguous documents leading to misunderstandings and discrepancies easily resulting in errors. To later gain from positive effects when UML-based development methods are used in large projects, there is a need for evaluating and understanding how UML is used in such large projects and to identify problem aspects. ABB is a global company with 110 000 employees. The company is a global leader in power and automation technologies. ABB decided to use UML-based development in the company's system development projects to enable certification according to the safety standard IEC 61508. A UML-based development method was developed and introduced in a large, international project with 230 co-workers, most of them newcomers to UML. The project developed embedded hardware and software. Some subprojects developed from scratch, but most of the work was based on reuse of old code. ABB started a case study to identify benefits and problems whit the use of UML, and became members of the national project SPIKE (Software Process Improvement based on Knowledge and Experience). The Simula Research Laboratory started in cooperation with ABB a project to evaluate how ABB's UML-based development methodology was used in the form of interviews, questioners and document analysis. The master thesis covers a case study in the form of document analysis with the focus on documents produced in the phases "requirements analysis" and "analysis and design" to identify problems from the work on use case modelling. Data from different subprojects variable in size, type of development and grade of legacy development are collected. The results show that unclear system boundaries resulted in problems because clear rules about the environment where not defined in advance. These problems showed up in the wrong identification of actors and use cases, also forcing other parts of the model to be changed. Difficulties with imprecise language like the use of ambiguous names and words resulted in problems with an ambiguous model. Maintenance became necessary to obtain a consistent model. Detailed descriptions about design in the high level specifications resulted in more complicated maintenance than would have been necessary with a less detailed representation. Among the individual projects one project performed additional rework, because the method was not defined in advance and were introduced at a later stage. Adjustments according to the new method were necessary, where the resulting model did not separate well between "analysis and design" and "detailed design". Difficulties where also identified for the subprojects that specified legacy systems because of the lack of a defined legacy process. The method described development from scratch, with little focus on how the legacy development could be performed. The resulting documents from "analysis and design" were rewritten many times. The problems with an undefined system boundary were most clear in the projects developing form scratch.nor
dc.language.isonoben_US
dc.titleThe evaluation of UML-based development methodology in a large projecten_US
dc.typeMaster thesisen_US
dc.date.updated2006-09-27en_US
dc.creator.authorSand, Gunhilden_US
dc.date.embargoenddate10000-01-01
dc.rights.termsKLAUSULERING: Dokumentet er klausulert grunnet lovpålagt taushetsplikt. Tilgangskode/Access code Cen_US
dc.rights.termsforeveren_US
dc.subject.nsiVDP::420en_US
dc.identifier.bibliographiccitationinfo:ofi/fmt:kev:mtx:ctx&ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&rft.au=Sand, Gunhild&rft.title=The evaluation of UML-based development methodology in a large project&rft.inst=University of Oslo&rft.date=2006&rft.degree=Hovedoppgaveen_US
dc.identifier.urnURN:NBN:no-13116en_US
dc.type.documentHovedoppgaveen_US
dc.identifier.duo44326en_US
dc.contributor.supervisorBente Andaen_US
dc.identifier.bibsys060834900en_US
dc.rights.accessrightsclosedaccessen_US


Files in this item

FilesSizeFormatView

No file.

Appears in the following Collection

Hide metadata