International Journal of Software Engineering and Knowledge Engineering

More than twenty years ago the idea of producing software systems from reusable software components was proposed. Since that time many changes have taken place in Computer Science and Software Engineering, but software systems are still built as one-of-a-kind craftsman efforts. A method for software construction using components is rationalized using experience from software components, program transformations, system architecture, industrial large systems, automatic programming and program generation. Experience with the method is discussed. The limiting factors of the method that prevent the widespread use of reusable software components are identified.

Domain modeling can be applied to collections of reusable designs and reusable software. Libraries of such collections can be used when applying a refinement by transformation technique to software construction. There already exist systems that can automatically or semi-automatically perform program transformations that constitute refinements. An important question is how to organize such libraries so that transformation tools may feasibly use them.

We show that transformation of a high level program with constraints on the transformations is an NP-complete problem; however, appropriately organized libraries are tractable. Moreover, we define a property which a library of transformations can have, ensuring that any consistent high level program can be transformed into an executable form. Finally, we give approximations which reduce the complexity of transformations for libraries which do not have this property.

The most important aspect of this work is that it implies certain rules should be followed in constructing libraries and the domains that are placed in them.

We present a model of the data structure domain that is expressed in terms of the GenVoca domain modeling concepts [7]. We show how familiar data structures can be encapsulated as realms of plug-compatible, symmetric, and reusable components, and we show how complex data structures can be formed from their composition. The target application of our research is a precompiler for specifying and generating customized data structures.

Domain analysis, while still relatively new, has matured to the point where it is beginning to appear in production settings. This raises the problem of how to select among proposed approaches. Domain analysis approaches possess some similarities but differ in many ways — some subtle, some significant, some uninteresting. Comparing them is not simple.

This paper presents criteria for comparing and contrasting domain analysis approaches. The criteria were derived by studying five approaches that seem representative. The criteria attempt to provide a common conceptual ground for comparing and contrasting different ways to undertake domain analysis. The criteria are studied in terms of external factors that determine how and why an organization should use domain analysis. The criteria are first presented independent of any approach. They are then applied to the five approaches, to illustrate characteristics the approaches possess, and to show what some of the approaches' merits are. The paper concludes with some thoughts on the implications of the criteria for the future of domain analysis, considering in particular the likelihood of a single, unified approach emerging.

Refining a specification S₁ means to provide another specification S₂ which contains all the information given in S₁ but with more detail. In this paper, we use logical implication from lower to higher levels of logic specifications to give a definition of refinement between these levels. This guarantees that any property of the higher level is also verified at the lower one. The definition of the relation "is refinement of" is given for specifications which are general first-order theories and it is proved to be transitive. A relevant aspect is that the different levels of logic specifications are in general not immediately comparable, because they can use different vocabularies. For this reason, the concept of transcription is introduced formally in our definition. Then the particular case of Horn specifications is considered. Horn specification semantics can be given by the methodology of least models. This may suggest definitions of the concept of refinement different from the one based on logical implication from lower to higher levels. However, conceptual problems can arise depending on the kind of the refinement definition chosen. Perhaps the most interesting effect is that the property of refinement transitivity may be lost. A possible way to restore the transitivity is provided.

A management discipline of software metrics facilitates their consistent measurement and usage during the life cycle of software systems, by maintaining knowledge and data related to their evaluation tasks and application domain knowledge. To investigate environment of the management discipline, we represent software engineering environments by an activity flow model and a micro cycle model. A unified management view of software metrics is established by their collectivity and comprehensiveness in two-dimensional flows of the activity flow model. Then, we propose the Software Quality Manager (SQM) as a realization of the management discipline of software metrics. The proposed software quality manager makes use of two knowledge bases: (1) knowledge about application domain in the form of reusable software components and (2) knowledge about software engineering. The three-stage architecture and preparation process of the proposed system are also presented. Finally, we propose how to support the measurement and analysis task for database management systems within the context of the software quality manager.

Three sample information retrieval systems, archie, autoLib, and WAIS, are compared as to their expressiveness and usefulness — first, in the general context of information retrieval, and then as prospective software reuse repositories. While the representational capabilities of these systems are limited, they provide a useful foundation for future repository efforts, particularly from the perspective of repository distribution and coherent user interface design.