============================= VOLUME 21 NUMBER 1 March 1997 ============================= 1. A Situational Informatical Dynamics: The Case of Situation-contextual and Time-contextual Non-additive Influences Magoroh Maruyama, Aomori Koritsu Daigaku, Goushizawa Aza Yamazaki 153-4 Aomori City, 030-01 Japan Fax: Japan+17764-1544 pp. 5-18 Keywords: causal loops, conjunction, disjunction, individual case studa, informational dynamics, interlocking, motivation, non-additive concepts, parajunction, stress Abstract: Individual decision making and organizational processes do not follow rules or patterns which are a priori predictable or situation-independently and time-independently programmable. We must look into situation-contextual and time-contextual texture instead of theoretical skeletonization and formula-fitting. There are many types of situational informatical dynamics. In this article the concept of non-additive influences is used as an example. A case is analyzed in detail. ------------------- 2. The Definition of an Integrated Software Life-Cycle Tool Rick Leonard, PEI Electronics, 110 Wynn Drive, Huntsville, Alabama 35812, USA pp. 19-29 Keywords: computer aided software engineering, convergent metrics, life-cycle models, software engineering, software metrics Abstract: For Computer Aided Software Engineering to fulfill its promise, every aspect of the software product should be accomplished within an integrated software engineering environment. Although integrated software engineering environments have been attempted in the past, several areas of concern still exist. Too often, the life-cycle models and their resulting products are incongruous between all phases of the software life-cycle. With this incongruity comes a loss of previously captured information which, in turn, results in a larger, overall work effort to complete the software product. Incongruity between life-cycle phases also causes inconsistent evaluation metrics and, as a result, improper life-cycle management. In addition, techniques for reconciliation of errors with previous life-cycle products are practically non-existent. This paper attempts to overcome the aforementioned problems by providing the background, methodology, and technical analysis to demonstrate an interactive software engineering environment capable of addressing the entire software life-cycle in an integrated, automated fashion. In addition, a related metric set is identified that converges with actual results as the life-cycle progress from one phase to the next. ------------------- 3. Qualitative Reasoning and a Circular Information Processing Algebra Franjo Jovic, Institute of Computer Engineering & Faculty of Electrical Engineering, University of J.J.,Strossmayer in Osijek, HR-31000 Osijek, Istarska 3, Croatia pp. 31-47 Keywords: qualitative modeling, circular algebra, quantitative/qualitative transformation, information, qualitative noise Abstract: Quantitative versus qualitative system modeling. Qualitative algebra. Events spaces and measure of their complexity. Ordinality and cardinality of models. Continuous Goedel numbering. Qualitative correlation algebra. Model sensitivity amplification by a circular quantitative/qualitative transformation. Qualitative noise. ------------------- 4. Hierarchical Classification as an Aid to Browsing Royce Rose and Caroline M. Eastman, Department of Computer Science, The University of South Carolina, Columbia, South Carolina 29208, USA pp. 49-57 Keywords unsupervised machine learning, hierarchical classification, browsing Abstract: An approach to browsing large chemical reaction databases is presented. The method that is described builds on earlier work in which unsupervised hierarchical classification was used to extract generalizations of reaction classes from reaction databases for use in reaction knowledge bases. The method described in this paper involves classification based on both semantic and topological features. It supports the creation of deep hierarchies in which succeeding levels represent increasing degrees of abstraction. The creation of a hierarchy allows the user to quickly locate interesting items or classes of items by performing a tree traversal as opposed to sequentially scanning a hit list. In addition, the depth of the resulting hierarchy is determined interactively by the user. ------------------- 5. Towards Recursive Models-A Computational Formalism for the Semantics of Temporal Presuppositions and Counterfactuals in Natural Language Stefano Mizzaro, Department of Mathematics and Computer Science, University of Udine, Via delle Scienze, 206, Loc. Rizzi, 33100 Udine, Italy, Tel: +39 (432) 55.8456, Fax: +39 (432) 55.8499 E-mail: mizzaro@dimi.uniud.it WWW: http://dimi.uniud.it/textasciitilde mizzaro pp. 59-77 Keywords: artificial intelligence, computational linguistics, natural language processing, semantics, pragmatics, temporal presuppositions, counterfactuals, linguistic knowledge, extra-linguistic knowledge, ontology, content, computational models, recursive models, TOBI Abstract: The linguistic phenomena of temporal presuppositions and counterfactuals, situated on the boundary line between semantics and pragmatics, are common to many languages, and the computational treatment of such phenomena is difficult because of their nonmonotonic aspect. These phenomena are presented through a corpus of examples; they are studied emphasizing the various types of knowledge underlying them; and the fragment of language that encloses such phenomena is defined in a way not dependent from a specific language. Then, Recursive Models, a formalism for modeling the semantics of utterances ontaining temporal presuppositions and counterfactuals, are proposed, described from both functional (by formal specifications) and structural points of view, and compared with related work. Finally, the adequacy of Recursive Models is empirically verified: TOBI (Temporal presuppositions and counterfactuals: an Ontological Based Interpreter), a system that interacts with the user in natural language using the recursive models, is illustrated. TOBI is not based on a deductive system, but uses the more primitive and flexible notion of model-based evaluation; its architecture, flow of control and internal data structures are presented. ------------------- 6. Informational Graphs Anton P. Zeleznikar, An Active Member of the New York Academy of Sciences, Volariceva ulica 8, SI-1111 Ljubljana, Slovenia Email: anton.p.zeleznikar@ijs.si pp. 79-114 Keywords: graph definition; informational gestalt, graph, operand, operator, star gestalt, transition; parenthesis pairs; primitive parallel formula system; serial-circular formula system Abstract: Informational graph seems to be one of the most basic underlying structures (circuits cite grafman,cummings, frontal lobes functions cite moscow, schemata cite stuss, impressions, etc.) for the concept and possibilities of informing and its understanding. The graph behaves as a regular parallel informational system of formulas (entities) with its own possibilities of informational spontaneity and circularity. By means of informational graph, it is possible to explain the origin of the so-called informational gestalt and, besides, the arising of informational formulas especially concerning the so-called causality in regard to the position of the formula parenthesis pairs. Another view of the graph lies in the moving along the arrows in the graph, that is, a formula construction, when choosing a path and setting parenthesis pairs in the emerging well-formed formula, in a spontaneous and circular way. This approach, together with the arising of the graph itself, can represent one of the keystones of the informational arising of formulas, the vanishing of their parts, and the changing of the structure during the informational moving through the graph. The paper shows how the informational graph can be understood by the phenomenalism of informational gestalts exerting the causal possibilities of formulas with the same length but differently displaced parenthesis pairs. Several examples are formalized. ------------------- 7. A Study in Generating Readable Modula-2 from Prolog Dan Resler, Virginia Commonwealth University, Richmond, VA U.S.A. Email: dresler@vcu.edu AND Danny Crookes, The Queen's University of Belfast, Belfast, Northern Ireland pp. 115-127 Keywords: program translation, program generators, software readability Abstract: This article presents an empirical study into generating readable imperative language programs from deterministic Prolog. An overview of a prototype translator is given, followed by a detailed discussion of the required transformations. Two case studies, complete with listings of both the specifications and resultant programs, are presented. An assessment of the generated code and overall project is also given. ------------------- 8. Design of Approximate Identity Neural Networks by Picewise-Linear Circuits Jiri Kaderka, Department of Microelectronics, TU of Brno, Udolni 53, 602 00, Brno, Czech Republic, Phone: +42 5 43167 134, Fax: +42 5 43167 298 E-mail: kaderka@umel.fee.vutbr.cz pp. 129-133 Keywords: approximate identity neural networks, piecewise-linear circuits Abstract: The design of the approximate identity neural network (AINN) using the piecewise-linear (PWL) approach is presented. AINN realizes linear and non-linear mapping. This net consists of two executive layers. The cells of the first - hidden layer, hidden neurons, are designed according to extended PWL sub-circuits that make the function of the absolute value. The present theory describes a transfer function and a learning algorithm of AINN. The convergence of the learning algorithm is provided numerically. The electronic circuit of AINN is based on PWL sub-circuits and modeled by Saber. Results show possibility of AINNs to approximate the sine and squared functions. The theory of PWL AINN can be extended; approximate networks can be defined in more dimensions. ------------------- 9. An Integrated Testing Framework for Object-Oriented Programs Shih-Sung Liao, Kai H. Chang and Chun-Yu Chen, Department of Computer Science and Engineering, Auburn University, Auburn, AL 36849 pp. 135-145 Keywords: Software testing, Formal methods, Object-oriented programs, Usage profile Abstract: It has been proved that object-oriented technology (OOT) can improve software productivity and quality through object reuse and high level of code modularity. While most OOT research efforts have been devoted to the object-oriented analysis, design, and programming, little attention has been paid to program testing, especially testing beyond class level. This paper presents an integrated object-oriented programs testing framework that distributes its testing effort based on usage and importance. The framework incorporates both formal methods and usage profiles. A usage profile combines the expected data distribution information and the frequencies of anticipated operation sequences of a program. The data distribution information is represented in a grammar form that can be used to generate test data. The frequencies of operation sequences can be derived from state transition diagrams that are enriched by operation probabilities. An object-oriented styled formal specification language, which enables a clearer association between specification and implementation, is used in this framework.