13.07.2015 Views

Software Design 2e - DIM

Software Design 2e - DIM

Software Design 2e - DIM

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

364closely at the representation parts of the methods surveyed. The conclusions of immediateinterest are as follows.<strong>Design</strong>ing with objectsnnnUse of DFDs. Wieringa observes that the use of DFDs (and of any ‘close relatives’)is incompatible with object-oriented structuring because of the enforced separationof data storage from the data processing that is implicit in the DFD model. As wasobserved in the previous section, the encapsulation of data and related operationsforms a core element of the object model.Use of Finite State Diagrams. These are seen as being the form that is most suitedfor use in object-oriented modelling because they allow ‘the specification of localstate, state transitions, and behaviour’.Conceptual system decomposition. Despite the variety of approaches to this foundamong the methods in this survey, Wieringa observes ‘overwhelming agreementthat the decomposition must be represented by a class diagram, component behaviourby a statechart, and component communications by sequence or collaborationdiagrams’. However, he also observes that there are quite extensive syntactic(and interpretative) variations in the use of these forms by different methods.Wieringa goes on to discuss the consequences of this last point in terms of notationalneeds of methods and, indeed, identifies this as forming part of the rationale for developmentof the UML.16.2.4 Survey 4: Johnson and Hardgrave, 1999The form adopted for this fourth study differs quite substantially from those employedfor the previous three studies we have reviewed. One key difference is that it was performedusing a survey technique employing two separate groups of subjects: experienceddevelopers and trainees. Also, the focus was upon the respondents’ attitudes andpreferences with regard to object-oriented methods and tools, rather than upon theforms of the specific methods.Comparison frameworkSeparate survey forms were used for each of the two groups and a total of 160 subjectstook part in the survey (102 experienced developers and 58 trainees). Since the surveywas conducted on-line, there are some methodological questions about sampling andrepresentativeness, as indeed the authors acknowledge.The authors also observed that the degree of comparison that could be achievedwas limited. This was chiefly because ‘a theory explaining attitudes and behaviourtoward, and use of OOAD methods does not currently exist’, and hence they arguedthat the survey offered an inductive approach that would ‘use facts to develop generalconclusions’ and lay some of the groundwork for the development of such a theory.In terms of its comparative element, this survey was chiefly concerned with thedegree of training provided for developers, their familiarity with and preferencebetween, object-oriented methods, and their attitudes towards them.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!