یکپارچه سازی تجزیه و تحلیل سیستم و ابزارهای مدیریت پروژه

Currently, computer-aided tools for system analysis are distinct from project management tools. This study proposes and prototypes a model that integrates these two aspects of the Information System Life Cycle (ISLC) by automatically mapping system analysis objects into project management objects. To validate the feasibility of our model and without loss of generality, the conversion of Data Flow Diagrams (DFD) objects into Gantt and Pert diagrams is demonstrated in this study. Experiments with the prototype confirm that integrating common tools for system analysis and standard tools for project management, during system development, helps improve system building tasks and their management. In addition, project managers using the proposed mapping approach can better assess project duration and system performance parameters such as response time and data traffic. We address implications of our work to both academics and practitioners, discussing directions future research might take as well as opportunities and prospects for commercialization of the proposed approach.

Computer-Aided Software Engineering (CASE) tools support the analysis, design, construction, and implementation stages of the Information System Life Cycle (ISLC) [1], [2] and [26]. Although Project Management (PM) tools support management and control of ISLC development tasks [4] and [5], there is hardly any integration to date between CASE and PM tools. Thus, ISLC modeling approaches, such as Data Flow Diagrams (DFD) or Unified Modeling Language (UML) [6], even when automated, are used in the early analysis stage primarily for visual documentation. The “database of specifications”, laboriously elicited and gathered during the creation of modeling diagrams, is hardly ever applied again for project management purposes, even though this information is valuable for project managers who are involved in the construction and implementation stages. In fact, due to lack of integration along the ISLC, the specifications database is often either overlooked altogether or collected again as if their creation earlier never took place. Moreover, standard methods for system analysis and development usually make no reference to methods for project management [7]. Given the current object orientation of both system analysis and project management methodologies, this paper contends that a model integrating these two aspects by automatically mapping system analysis objects into project management objects is not only desirable but also feasible. The need for such mapping emerges from the next section that contains a theoretical and practical review of DFDs, Functional Hierarchies (FH), Pert and Gantt diagrams, and the corresponding computer-aided tools. The focus of this study on data-flow, Gantt, and Pert diagrams are done, without loss of generality, for demonstration purposes of our model to validate the model and its feasibility. Based on Section 2, it becomes clear that a major problem in system development is the gap between the analysis phase, which deals with the required functionality (i.e. the Requirements Engineering phase [8]), and the construction phase, which deals with the actual program coding. 3 and 4, respectively, describe and prototype a model that maps system analysis objects, based on DFDs or FHs, into project management and control objects, in the form of Gantt and Pert diagrams. On the basis of this mapping, it is also possible to run a Critical Path Method (CPM) analysis with respect to “DFD component chains” paths. Section 4 shows that bridging analysis and construction based on our approach is more objective than without our approach because raw DFD components are stored in the specifications database. Because of its automation, the proposed mapping can shorten project duration, contribute to efficient use of resources, and help estimate the required development effort and cost as well as the expected system performance, complexity, and quality. In the concluding section, we sum up the study and address the implications of our work to academics, in terms of directions for future research, and to practitioners, in terms of opportunities and prospects for commercialization.

The current research demonstrates the derivation of project management objects directly on the basis of raw system analysis and design objects. By so doing we provide an integrating layer, which combines standard project management and control tools with common system analysis and design tools. Applying such an integrating layer in software development projects enables, due to the possibility of executing the CPM on paths constructed of “software component chains”, improved assessment of such essential elements as development duration, development cost, response time, and data traffic volumes. CPM execution can enable monitoring and control over various topics uniquely related to software projects. Without such an integrating model, estimating development cost and time is difficult, no matter which methodology is applied, for the following three reasons: 1. First, estimation must be based on quantitative data, which are available only after providing the results of analysis to members of the development team who, for the most part, are not involved in the analysis stage [1], [2], [24], [25] and [26]. 2. Second, methods for evaluating development time and cost are hardly an integral part of system analysis and design tools. The same holds for methods for estimating the response time of the system under development. 3. Third, accepted models such as COCOMO [27], are based on a presumptive assessment of archaic factors such as program size and LOC. Today’s fourth generation software provides little in terms of certainty or accuracy. Estimation of project cost and time is, therefore, based usually on subjective judgment of a veteran development manager or consultant on the basis of comparisons with previous similar projects, or breaking down the project into sub-components, and evaluating each separately [28]. Other than these drawbacks of conducting development in the absence of an integrating layer, the mapping model presented in this study has the following advantages: 1. The mapping model enables effort and cost estimation for information system development to be an integral part of conventional analysis and design methodologies, i.e. DFD and FH. This is due to the possibility of deriving assessments directly from system analysis and design raw data, as opposed to relying on aggregates or other secondary sources. 2. The mapping model enables extension of the development effort and cost assessment process, beyond the mere aspects of <Time> and <Resources and Budget>, to include also <Complexity and Quality>, yielding expected system response times, and data volume traffic. 3. The mapping model bases the estimation process on the general project management and control tools Gantt and Pert, which are widely used and supported (e.g. MS-Project). 4. Use of Gantt and Pert diagrams enables dynamic control of the estimation, based on reports regarding actual progress. This provides for projected-to-actual comparisons of cost, and moment-to-moment updates of CPM calculations, as opposed to the static control methods customary in the field of system building. 5. Use of Gantt and Pert charts allows “drilling down”, into the system code design, including master routines and system major service routines. This differs from current methods used for information systems development, which are limited primarily to the area of functionality. 6. The potential of a detailed drill-down concerning system code design provides for engaging and integrating the technical team (development managers) as early as the analysis stage. This results in a more reliable and accurate estimation base for the entire system development project. It is noteworthy that conversion of a DFD model to Pert/Gantt diagrams is actually a representation of a knowledge model as a semantic network. Even though DFD models have been used for demonstration purposes in this paper, the current research suggests that the demonstrated conversion capability is applicable to any other model, which can be represented in a network format. Stated differently, we have reason to believe that the integration of system analysis and project management tools, modeled and prototyped in this study, is not limited to the DFD approach, but can be applied to any “network-based” modeling such as the UML. Therefore, a promising direction for further research may focus on integrating UML diagrams with project management tools. The fact that conversion of a DFD model to Pert/Gantt diagrams is actually a representation of a knowledge model as a semantic network opens opportunities for commercialization for practitioners. Gaines and Shaw [29] present a formal visual language that enables such a representation and show how this language can represent a knowledge model in a document, either within a word processor or within a Web browser. Based on their work, the prospects are good for commercially expanding our mapping model, in the future, to support virtual development teams on the Web, regardless of geography and time constraints. In sum, this study showed the feasibility and validity of translating system analysis objects into project management objects. Our mapping, which aims to integrate common information system analysis and design methodologies with standard project management and control tools, can potentially improve estimation, planning, and control of software development projects, in terms of cost, time, and management. In such projects, where so many things may go out of control, any raw information gained during the analysis phase will make a contribution in later phases of system building and project monitoring.