Bruno Alabiso. Transformation of data flow analysis models to object oriented design. In OOPSLA Conference, Special Issue of SIGPLAN Notices, pages 335--353, San Diego, CA, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....(defined by common routines) among objects. This commonality could be represented as a generalized object with each of the sharing objects containing routines that specialize the generalized object. 2.2. 4 Other approaches From Data Flow Diagrams to Object Oriented Design Alabiso [3] proposes a method for transforming Data Flow Diagrams (DFDs) into an Object Oriented Design notation. Although some ideas can be borrowed from it, this method cannot be directly applied to the problem of reverse engineering or reengineering source code into the object oriented paradigm. This is ....

....the procedural paradigm Substantial research has been done in reverse engineering within the procedural paradigm. Benedusi et al. [5] propose a method to produce a hierarchy, at different levels of abstraction, of Data Flow Diagrams (DFDs) from Pascal code. This work, in conjunction with Alabiso s [3], poses the following question. Would it make sense to reverse engineer a DFD and from it an object oriented model as an approach to identifying objects and their responsibilities in a procedural program To answer this question, Wilde [64] asks, will the abstracted DFD tell much about what ....

Bruno Alabiso. Transformation of data flow analysis models to object oriented design. In OOPSLA Conference, Special Issue of SIGPLAN Notices, pages 335--353, San Diego, CA, 1988.

....have arisen, that all adopt extensions of the ER modelling method and combine it with DF models, state machine models or with pre postcondition style specifications. In addition, there has been a lot of interest in the possibility to integrate object oriented modelling with Structured Analysis [1, 2, 38, 44, 47]. The possibility to combine Jackson System Development (JSD) with an object oriented approach has also roused interest [4] The possibility to combine JSD modelling with ER modelling is briefly discussed by Sutcliffe [42] but is not studied there in detail. These attempts at integration can be ....

B. Alabiso. Transformation of data flow analysis models to object oriented design. In N. Meyrowitz, editor, Object-Oriented Programming Systems, Languages and Applications, Conference Proceedings, pages 335-- 353. ACM Press, 1988. SIGPLAN Notices, volume 23.

....models. 1 Introduction In recent years, there has been a rising interest in the possibility to combine different conceptual modelling methods. For example, the possibility to integrate object oriented modelling with Structured Analysis has received the attention of a number of researchers [1, 3, 20, 26, 28]. The possibility to combine Jackson System Development (JSD) with an object oriented approach has also roused interest [5] The possibility to combine JSD modelling with Entity Relationship (ER) modelling is briefly discussed by Sutcliffe [25] but is not studied there in detail. A comparison of ....

B. Alabiso. Transformation of data flow analysis models to object oriented design. In N. Meyrowitz, editor, Object-Oriented Programming Systems, Languages and Applications, Conference Proceedings, pages 335--353. ACM Press, 1988. SIGPLAN Notices, volume 23.

....VDM speci cations. In Section 4.5 we come back to this restriction in more detail. There is a one to one mapping between the input to the system and the output from the system. One to many mappings and many to one mappings are a common problem when interpreting DFDs, described in more detail in [Alabiso88] 4 . However, we are not entirely satis ed with the solution proposed by Alabiso, and since in our experience most of the DFDs with one to many or many to one mappings should be regarded as design products and not as speci cation products, we feel that a restriction to one to one mappings is ....

....Alternatively, the analyst may supply a mini speci cation for each non primitive data transformer not obeying the restriction of a one to one mapping between input and output. To simplify the formal description the data ows must have unique names at each level in the hierarchy of DFDs. 4 In [Alabiso88] this problem is called I O uncohesiveness. I O uncohesiveness occurs if either a data transformer must consume several pieces of input data before generating output data, or if a data transformer generates pieces of output independently of all other inputs and outputs. Alabiso describes a ....

[Article contains additional citation context not shown here]

Bruno Alabiso. Transformation of Data Flow Analysis Models to Object Oriented Design. In OOPSLA'88 Proceedings, pages 335-353, ACM, November 1988. 19 pages.

....are much more course grained than ours. Therefore, it is not possible to build sophisticated transformation tools which allow for the propagation of incremental changes between requirements definition and design. Other approaches try to find data abstractions analyzing the structure of the DFDs ([1], 2] 14] 15] These methods tend to create very complex diagrams which are difficult to understand or the transformation process is not well described. There are also some authors which totally reject (simple) SA as a front end for object oriented design ( 7] It is our opinion too that a ....

Alabiso, B. `Transformation of Data Flow Analysis Models to Object--Oriented Design' Proceedings of OOPSLA '88 (1988).

....and target tools may handle the same or different levels of abstraction. The transformational approach to software development [13] concentrates on transformations from a higher level of abstraction to a lower level of abstraction, such as transformations from analysis to design models (see, e.g. [1, 6]) However, we may also want to interface source and target models on similar levels of abstraction. If we want to move entire software artifacts from an outdated tool to a new target environment (defined as immigration in [17] only unidirectional information transfer is adequate. Similarly, if ....

Bruno Alabiso. Transformation of data flow analysis models to object oriented design. In Proceedings of OOPSLA'88, pages 335 -- 353, September 1988.

....an integrated structured formal method. In [Randell91] a translation back and forth between DFDs and Z specifications is mentioned. However, this translation simply corresponds to generating an abstract syntax representation of the DFDs, and it does not provide any semantics to composite DFDs. [Alabiso88] contains an explanation of how one can manually transform DFDs into an object oriented design. The paper touches upon some of the problematic issues arising in a transformation from DFDs, in particular the notion of I O uncohesiveness which we also encountered in our formalization of DFDs. In ....

Bruno Alabiso. Transformation of Data Flow Analysis Models to Object Oriented Design. In OOPSLA'88 Proceedings, pages 335--353, ACM, November 1988.

....are that can help the modeler to make decisions in performing these tasks. There are several attempts to combine SA and OOA, either by incorporating OOA in SA (Ward [35] or by incorporating SA in OOA (Bailin [4] or by using the output of SA as the input to object oriented design (Alabiso [2], Seidewitz and Stark [31] In this paper I show that these approaches all have problems, because SA and OOA are fundamentally incompatible. However, this incompatibility is argued by comparing the two modeling approaches, so that by its very argument, this paper shows that the two methods can be ....

....the system must interact (these are the external entities in SA) and ffl listing the principal actors that help to solve the problem which the IS is designed to solve [34] 4.1.2 Objects in SA Although SA is function oriented, it cannot help encountering objects. For example, following Alabiso [2], obvious candidates for objects in a DFD are external entities and data stores in a DFD. However, applying the criteria for objects listed above, we can do better than just listing all external entities and data stores as candidate objects, because we can reduce the number of objects considerably ....

[Article contains additional citation context not shown here]

B. Alabiso. Transformation of data flow analysis models to object oriented design. In N. Meyrowitz, editor, Object-Oriented Programming Systems, Languages and Applications, Conference Proceedings, pages 335--353. ACM Press, 1988. SIGPLAN Notices, volume 23.

....Faculty of Mathematics and Computer Science, Vrije Universiteit, Amsterdam. 1702 Roel Wieringa and recommend something that we refer to as subject domain oriented decomposition. The need for integration of structured and object oriented software specification methods has been long recognized [1, 3, 32, 48, 53]. One of the advantages of such an integration is that it allows practitioners raised in the world of data flow modeling and functional decomposition to incorporate useful elements of objectoriented specification in a stepwise, evolutionary manner. Another advantage is that it allows us to pick ....

.... Early integration proposals incorporate object oriented ideas in structured analysis without fundamentally changing structured analysis [53] or incorporate structured analysis in object oriented analysis [3] or simply use structured analysis as a front end to object oriented design [1, 32, 48]. None of these proposals is based upon a thorough analysis of the underlying principles of structured and OO analysis. Without such an analysis, it is not possible to see which elements of structured and OO methods can or cannot be combined and why this is so. The integration proposed in this ....

B. Alabiso. Transformation of data flow analysis models to object oriented design. In N. Meyrowitz, editor, Object-Oriented Programming Systems, Languages and Applications, Conference Proceedings, pages 335--353. ACM Press, 1988. SIGPLAN Notices, volume 23.

....In Section 4.2.2 we come back to this restriction in more detail. ffl There is a one to one mapping between the input to the system and the output from the system. One to many mappings and many to one mappings are a common problem when interpreting DFDs, described in more detail in [Ala88] 4 . However, we are not entirely satisfied with the solution proposed by Alabiso, and since in our experience most of the DFDs with one to many or many to one mappings should be regarded as design products and not as specification products, we feel that a restriction to one to one mappings is ....

....dictionary (DD) and a collection of uniquely identified mini specifications (MSs) The types of all data flows in the data flow diagrams must be defined in the data dictionary. Besides this, the signature of the top level DFD must conform to its topology. SA = HDFD Theta DD Theta MSs 4 In [Ala88] this problem is called I O uncohesiveness. I O uncohesiveness occurs if either a data transformer must consume several pieces of input data before generating output data, or if a data transformer generates pieces of output independently of all other inputs and outputs. Alabiso describes a ....

[Article contains additional citation context not shown here]

Bruno Alabiso. Transformation of Data Flow Analysis Models to Object Oriented Design. In OOPSLA'88 Proceedings, pages 335--353. ACM, November 1988.

....patron and the document being checked out. As described above, object design also must determine the data structures and algorithms to implement the object classes. This is done as in the traditional software design and is omitted in 7 Undoubtedly, other notations, such as Booch s notation [1] and Wasserman s notation [25] may also be used to represent object designs. this paper. 5 Object Oriented Programming In this section, we first describe how the basic concepts of object orientation can be realized in the C programming language [24] We then present how these can be achieved ....

.... no yes yes multiple inheritance yes yes no assertions no yes no run time meta data no no yes static binding when possible yes yes no garbage collection no yes yes memory allocation static automatic dynamic dynamic dynamic class browser no yes yes object persistence no yes yes 6 References [1] G. Booch, Object oriented Development, IEEE Trans. Software Eng. Vol. SE 12, No. 2, pp. 211 221, February 1986. 2] J. Brunet, Modeling the world with semantic objects, Proc. of the IFIP Conference, Quebec, Canada, October, 1991. 3] C. M. Chao and C. Kung, Rapid prototyping of ....

B. Alabiso, "Transformation of data flow analysis models to object-oriented design." Proc. OOPSLA'88, pp. 335 - 353, ACM, Sept. 1988.

No context found.

: B. ALABISO, "Transformation of data flow analysis models to object-oriented design", In Proc. of OOPSLA'88, San Diego, California, 1988.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC