Search

STUDY QUESTIONCan risks associated with novelties in assisted reproduction technologies (ARTs) be assessed in a systematic and structured way?SUMMARY ANSWERAn ART-specific risk assessment tool has been developed to assess the risks associated with the development of novelties in ART (EuroGTP II-ART).WHAT IS KNOWN ALREADYHow to implement new technologies in ART is well-described in the literature. The successive steps should include testing in animal models, executing pre-clinical studies using supernumerary gametes or embryos, prospective clinical trials and finally, short- and long-term follow-up studies on the health of the offspring. A framework categorizing treatments from experimental through innovative to established according to the extent of the studies conducted has been devised. However, a systematic and standardized methodology to facilitate risk evaluation before innovations are performed in a clinical setting is lacking.STUDY DESIGN, SIZE, DURATIONThe EuroGTP II-ART risk assessment tool was developed on the basis of a generic risk assessment algorithm developed for tissue and cell therapies and products (TCTPs) in the context of the project ‘Good Practices for demonstrating safety and quality through recipient follow-up European Good Tissue and cells Practices II (EuroGTP II)’. For this purpose, a series of four meetings was held in which eight ART experts participated. In addition, several tests and simulations were undertaken to fine-tune the final tool.PARTICIPANTS/MATERIALS, SETTING, METHODSThe three steps comprising the EuroGTP II methodology were evaluated against its usefulness and applicability in ART. Ways to improve and adapt the methodology into ART risk assessment were agreed and implemented.MAIN RESULTS AND THE ROLE OF CHANCEAssessment of the novelty (Step 1), consisting of seven questions, is the same as for other TCTPs. Practical examples were included for better understanding. Identification of potential risks and consequences (Step 2), consisting of a series of risks and risk consequences to consider during risk assessment, was adapted from the generic methodology, adding more potential risks for processes involving gonadic tissues. The algorithm to score risks was also adapted, giving a specific range of highest possible risk scores. A list of strategies for risk reduction and definition of extended studies required to ensure effectiveness and safety (Step 3) was also produced by the ART experts, based on generic EuroGTP II methodology. Several explanations and examples were provided for each of the steps for better understanding within this field.LIMITATIONS, REASONS FOR CAUTIONA multidisciplinary team is needed to perform risk assessment, to interpret results and to determine risk mitigation strategies and/or next steps required to ensure the safety in the clinical use of novelties.WIDER IMPLICATIONS OF THE FINDINGSThis is a dynamic tool whose value goes beyond assessment of risk before implementing a novel ART in clinical practice, to re-evaluate risks based on information collected during the process.STUDY FUNDING / COMPETING INTEREST(S)This study was called EUROGTP II and was funded by the European Commission (Grant agreement number 709567). The authors declare no competing interests concerning the results of this study.

Background: In the area of artificial learners, not much research on the question of an appropriate description of artificial learner's (empirical) performance has been conducted. The optimal solution of describing a learning problem would be a functional dependency between the data, the learning algorithm's internal specifics and its performance. Unfortunately, a general, restrictions-free theory on performance of arbitrary artificial learners has not been developed yet.Objective: The objective of this paper is to investigate which function is most appropriately describing the learning curve produced by C4.5 algorithm.Methods: The J48 implementation of the C4.5 algorithm was applied to datasets (n=121) from publicly available repositories (e.g. UCI) in step wise k-fold cross-validation. First, four different functions (power, linear, logarithmic, exponential) were fit to the measured error rates. Where the fit was statistically significant (n=86), we measured the average mean squared error rate for each function and its rank. The dependent samples T-test was performed to test whether the differences between mean squared error are significantly different, and Wilcoxon's signed rank test was used to test whether the differences between ranks are significant.Results: The decision trees error rate can be successfully modeled by an exponential function. In a total of 86 datasets, exponential function was a better descriptor of error rate function in 64 of 86 cases, power was best in 13, logarithmic in 3, and linear in 6 out of 86 cases. Average mean squared error across all datasets was 0.052954 for exponential function, and was significantly different at P=0.001 from power and at P=0.000 from linear function. The results also show that exponential function's rank is significantly different at any reasonable threshold (P=0.000) from the rank of any other model.Conclusion: Our findings are consistent with tests performed in the area of human cognitive performance, e.g. with works by Heathcote et al. (2000), who were observing that the exponential function is best describing an individual learner. In our case we did observe an individual learner (C4.5 algorithm) at different tasks. The work can be used to forecast and model the future performance of C4.5 when not all data have been used or there is a need to obtain more data for better accuracy.

Managing of passwords in information systems is a very important task, yet nothing seems to be learned from the recent stories. The consequences of bad password management practices have led to the loss of lives, as in the case of suicides after the “Ashley Madison leak”. Password security is simply not taken seriously, despite problems being known since 1979 at least. Interestingly, the PICMET conference on-line system itself implements a bad password management policy as all passwords are stored and re-sent upon request by plaintext email. The objective of this paper is to present the underlying mechanisms that lead to bad password management policies. Memorability and memory decay, complexity, simplicity and other factors are presented and analyzed. A novel password management policy “Psychopass” is proposed, where a password can be created, memorized and recalled by thinking of an action sequence (visual representation) instead of a string of characters. In the experiment it was shown that users tend to better remember passwords under the “Psychopass” policy compared to other password management policies nowadays in effect. The results confirm that “Psychopass” policy is an alternative to the existing password management practices and can improve the resilience to the attacks on information systems.

Framework-based development is currently regarded as one of the most promising software development approaches when it comes to improvements in lead time, productivity and quality. However, many frameworks and projects based on frameworks still report failures, which indicate that there are problems related to both frameworks technology and frameworks usage. The objective of our research was to examine the major drivers that have an impact on a framework's acceptance and a framework's success. We used the technology acceptance model (TAM) and Seddon's information systems success model as our underlying theory. Data collected from an online survey of 389 active framework users was used to test hypothesized models. Data analysis was performed via structural equation modeling. Our findings support the post-adoption version of TAM and the relationship between continuous use and the successful use of systems, where more use also means more net benefits. We found that the successful use of frameworks is mainly dependent on two factors: continuous framework usage intention and the perceived usefulness of the framework. Several practical and theoretical implications can be foreseen including advances in framework development guidelines and insight into the relationship between the acceptance and success of frameworks.

Versioning is an important aspect of web service development, which has not been adequately addressed so far. In this article, we propose extensions to WSDL and UDDI to support versioning of web service interfaces at development-time and run-time. We address service-level and operation-level versioning, service endpoint mapping, and version sequencing. We also propose annotation extensions for developing versioned web services in Java. We have tested the proposed solution for versioning in two real-world environments and identified considerable improvements in service development and maintenance efficiency, improved service reuse, and simplified governance.

This article proposes specific extensions for WS-BPEL (Business Process Execution Language) to support versioning of processes and partner links. It introduces new activities and extends existing activities, including partner links, invoke, receive, import, and onmessage activities. It proposes version-related extensions to variables and introduces version handlers. The proposed extensions represent a complete solution for process-level and scope-level versioning at development, deployment, and run-time. It also provides means to version applications that consist of several BPEL processes, and to put temporal constraints on versions. The proposed approach has been tested in real-world environment. It solves major challenges in BPEL versioning.

This paper explores the relationship between software size, development effort and team size. We propose an approach aimed at finding the team size where the project effort has its minimum. The approach was applied to the ISBSG repository containing nearly 4000 software projects. Based on the results we provide our recommendation for the optimal or near-optimal team size in seven project groups defined by four project properties.

PurposeOne way to improve the quality of business and software processes is to make them compliant with standard. This paper seeks to explore the suitability of the BPMN and the teaching method for process modeling, which uses SEI CMMI‐SW standard as input and produces software process models as output.Design/methodology/approachThe research approach consists of two parts. First, the students' responses during practical lectures of process modeling were gathered and studied. Then, the teaching process was modified and students' responses were studied again. The second part involves the survey about students' attitude towards BPMN usability and ease of use.FindingsThere are two key findings of the study. First, using the SEI CMMI‐SW standard in early phases of software process modeling is not very effective, because it overloads the students with abstract requirements, which are included in the standard. The more effective way is to use the standard in later phases of process modeling, in the phase of assessment and improvement of the process model. The second key finding is that there is a positive attitude among students towards BPMN usage for process modeling.Research limitations/implicationsThere is no evidence that this teaching approach can be generalized beyond the population of information science students and academic environment.Practical implicationsThe improved teaching method for business and software process modeling can be used in similar environments (practical lectures of process modeling). Although there is no evidence, quality managers from organizations can also benefit from the lesson learned and avoid a less suitable process‐modeling approach.Originality/valueThe paper is useful to all roles in education and businesses, who deal with process models and standards (teachers, business and software analysts, quality managers, process designers).

One of the tasks of data mining is classification, which provides a mapping from attributes (observations) to pre-specified classes. Classification models are built by using underlying data. In principle, the models built with more data yield better results. However, the relationship between the available data and the performance is not well understood, except that the accuracy of a classification model has diminishing improvements as a function of data size. In this paper, we present an approach for an early assessment of the extracted knowledge (classification models) in the terms of performance (accuracy), based on the amount of data used. The assessment is based on the observation of the performance on smaller sample sizes. The solution is formally defined and used in an experiment. In experiments we show the correctness and utility of the approach.

Estimating software size is a difficult task that requires a methodological approach. Many different methods that exist today use distinct abstractions to depict a software system. The gap between abstractions becomes even greater with object-oriented artifacts developed in unified modeling language (UML). In this paper, a formal foundation for the representation of functional size measurement (FSM) methods is presented. The generalized abstraction of the software system (GASS) is then used to formalize different functional measurement methods, namely the FPA, MK II FPA and COSMIC-FFP. The same model is also used for object-oriented projects where UML artifacts are mapped into the GASS form. The algorithms in symbolic code for those UML diagrams that are crucial for size estimation are also given. The mappings defined in this paper enable diverse FSM methods to be supported in estimation tools, the automation of counting steps and a higher-level of independence from the FSM method, since the software abstraction is written in a generalized form. Both improvements are crucial for the practical use of FSM methods.

This article analyses two most commonly used distributed models in Java: Web services and RMI (Remote Method Invocation). The paper focuses on regular (unsecured) as well as on secured variants, WS-Security and RMI-SSL. The most important functional differences are identified and the performance on two operating systems (Windows and Linux) is compared. Sources of performance differences related to the architecture and implementation are identified. The overheads related to the usage of security and the influences of JCE (Java Cryptography Extension) security providers on the performance of secured remote invocations are identified. Finally, the impact of distributed models on design and implementation of distributed applications is identified and guidelines for improving distributed application performance in design and implementation stage are provided. The paper contributes to the understanding of functional and performance related differences between Web services and RMI and their secure variants, WS-Security and RMI-SSL.

A systematic approach to software size estimation is important for accurate project planning. In this paper, we will propose the unified mapping of UML models into function points. The mapping is formally described to enable the automation of the counting procedure. Three estimation levels are defined that correspond to the different abstraction levels of the software system. The level of abstraction influences an estimate's accuracy. Our research, based on a small data set, proved that accuracy increases with each subsequent abstraction level. Changes to the FPA complexity tables for transactional functions will also be proposed in order to better quantify the characteristics of object-oriented software.

The standard Function Point Analysis (FPA) method maps complexity to an ordinal scale such that function types (functions) labelled ''high'' complexity may have a very different underlying complexity. This deficiency is the subject of this article. An improved FPA method is presented that breaks down a very complex function into an equivalent set of functions with normal complexity. The improved FPA method is not a new method, it is an improvement of the standard FPA method. In the process of defining an improved FPA method, a model was built, which helped us find the connection between the standard FPA method and the Common Software Measurement International Consortium-Full Function Point (COSMIC-FFP) and Mark II Function Point Analysis (MKII FPA) methods. These methods were used to define the improved FPA method algorithm, because they show a bigger functional size for very complex functions. In the end, an evaluation of the improved FPA method was made with the results of the ISBSG Repository Data Disk.

A server infrastructure in web servers, message servers and other parallel systems use a variation of two software architectures for providing concurrency: threaded or event-driven. This paper analyzes the performance limitations of concurrent applications implemented in Java. Both architectures have been evaluated and compared with various design patterns, which combine the best practices from both architectures. For each architecture the suitability for handling a large volume of client requests, the efficient management of a server load, the influence of client request structures, and the physical size of a client request, have been studied. The discussed Java APIs are core technologies for high-level APIs, used in developing web and distributed applications. The research also includes performance comparison on various platforms and discusses performance variation on various versions of a Java runtime. The paper contributes to the understanding of Java-based server architecture capabilities. Core server software architectures and required Java libraries are compared, the reasons for the limitations are identified and guidelines for choosing proper combinations are given.

Software size is an important attribute in software project planning. Several methods for software size estimation are available; most of them are based on function points. Albrecht introduced function points as a technologically independent method with its own software abstraction layer. However, it is difficult to apply original abstraction elements to current technologies. Therefore researchers introduced additional rules and mappings for object-based solutions. In this paper several mapping strategies are discussed and compared. Based on the similarities in compared mappings, a common mapping strategy is then defined. This mapping is then tested on the reference application portfolio containing five applications. The aim of the test scenario is to evaluate the impact of the diverse detail levels in the class diagrams on software size measurement. Although the question of how to perform quality size measurements in object-oriented projects remains unanswered, the paper gives valuable information on the topic, supported by mathematics.

This article compares different approaches for developing Java distributed applications which have to communicate through firewalls and proxies, including RMI over open ports, HTTP-to-port, HTTP-to-CGI, HTTP-to-servlet tunneling and web services. A functional comparison of approaches has been done, as well as a detailed performance analysis with overhead analysis and identification of optimizations. Therefore the paper contributes to the overall understanding of different approaches for developing Java distributed applications in circumstances, where the communication through firewalls and/or proxies is inevitable. The paper also contributes to the understanding of performance related issues.

This article compares binary and XML object serialization on Java and Microsoft .NET platforms from the performance and size perspective. It uses three different types of objects and different number of objects to make a comparison which reflects real-world circumstances. The article has the following contributions: (1) it compares binary and XML serialization between Java and .NET to compare the efficiency of both platforms; (2) it compares binary and XML serialization within the platforms to compare the differences between the two serialization types; (3) it studies the reasons for performance differences and provides possible performance optimizations.

In the paper, we present an algorithm that can be applied to protect data before a data mining process takes place. The data mining, a part of the knowledge discovery process, is mainly about building models from data. We address the following question: can we protect the data and still allow the data modelling process to take place? We consider the case where the distributions of original data values are preserved while the values themselves change, so that the resulting model is equivalent to the one built with original data. The presented formal approach is especially useful when the knowledge discovery process is outsourced. The application of the algorithm is demonstrated through an example.

In medical decision making (classification, diagnosing, etc.) there are many situations where decision must be made effectively and reliably. Conceptual simple decision making models with the possibility of automatic learning are the most appropriate for performing such tasks. Decision trees are a reliable and effective decision making technique that provide high classification accuracy with a simple representation of gathered knowledge and they have been used in different areas of medical decision making. In the paper we present the basic characteristics of decision trees and the successful alternatives to the traditional induction approach with the emphasis on existing and possible future applications in medicine.

The primary motivation to reuse database components is to reduce the time and the effort required when building a conceptual model, actually a database. Because the quality of software systems is enhanced by reusing quality software artifacts, similarly, reusable database components can influence logical and physical database design and database maintenance, as well as quality. At the same time reusable database components can be compared with pattern technology, understanding pattern as a reusable database component. The pattern technology guides the user by observations and measurements to the identification of potentially useful patterns. Such patterns are suggestions, not prescriptions and are subject to change; they should be adopted, reformed and improved if needed. Usually they are not connected to a specific functional or business application domain. They can be used in many different environments, including medicine. According to different definitions, an intelligent system is a "power tool for thinking" but on the other side it is only a kind of information system with built-in knowledge to support decisions made by human experts. Similarly we could conclude for intelligent medical systems and introduce the database reusability in this environment in order to increase the quality of an intelligent medical system. The problem of database reusability is presented in detail, especially its integration in an intelligent medical system. Finally the results of such an integration and the benefits for medicine are discussed.

Distributed object architectures and Java are important for building modern, scalable, web-enabled applications. This paper is focused on qualitative and quantitative comparison of two distributed object models for use with Java: CORBA and RMI. We compare both models in terms of features, ease of development and performance. We present performance results based on real world scenarios that include single client and multi-client configurations, different data types and data sizes. We evaluate multithreading strategies and analyse code in order to identify the most time-consuming methods. We compare the results and give hints and conclusions. We have found that because of its complexity CORBA is slightly slower than RMI in simple scenarios. On the other hand, CORBA handles multiple simultaneous clients and larger data amounts better and suffers from far lower performance degradation under heavy client load. The article presents a solid basis for making a decision about the underlying distributed object model.

This paper is focused on the performance analysis, comparison and optimization of distributed object middleware for Java 2: RMI (Remote Method Invocation), CORBA IDL (Interface Definition Language) and RMI-IIOP (Remote Method Invocation on Internet Inter-ORB Protocol). The paper presents the following contributions to the research on distributed object performance. First, a detailed performance analysis is provided with the comparison. These results help to understand how the models perform. Second, an overhead analysis has been done, which explains why there are differences in peRformance. Third, optimizations and improved performance for RMI-IIOP and CORBA IDL are presented. These show considerably better performance in all areas compared to the original versions.

The ability of a new technology to reuse legacy systems is very important for its economic success. This paper presents a method for integrating legacy systems within distributed object architectures. The necessary steps required for integration are defined. It is explained how to define object interfaces. A detailed overview of how to implement the wrappers is given. The paper also answers the question which distributed object model is most suitable for legacy integration. Therefore a decision model is defined and the evaluation results are presented.

Most known models for SPI (e.g. SEI CMM, ISO standards and other methods derived from those mentioned) are primarily suited for large or medium organizations, but with some tailoring they provide substantial support also for the SPI in small organizations considering their specific characteristics. In the article a case of such tailoring – the PROCESSUS model – is presented. The baseline of the methodology is the integration of the CMM and the ISO 9001 together with the ISO 9000-3. According to the integrated model and the study of different lifecycles, a set of standard procedures (SP) and standard documents (SD) was defined. Each standard procedure provides guidelines on how to perform related activities, who is involved, which documents are supposed to be used/derived within the procedure etc. The set of SP and SD is the essential help for SPI conduction – for the purpose of small companies the optimal use of suggested documents and the disposition of roles was defined. The SP-SD set is also described in the article. Copyright © 2000 John Wiley & Sons Ltd

Java has an important role in building distributed object oriented web enabled applications. In the article an analysis of two distributed object models in context of Java language is presented. Several aspects of RMI and CORBA such as features, maturity, support for legacy systems, learning curve and ease of development are compared. A special emphasis is given to the performances. Different testing scenarios give a complete overview about real world performances of both architectures. Based on the comparison results, recommendations for selecting the most appropriate architecture for a given problem domain are presented. Therefore the paper contributes to the understanding of the distributed object architectures and to the study of Java RMI and CORBA performances.

Every project, including software engineering projects, start with the project plan. A project plan means the prediction of future events with the goal of minimizing the risk during execution of the project. The heart of the problem is to make the project plan as close to reality as possible. Ideally, it should be exactly the same as the post project measured data. In the end we always encounter the problem of "the perfect project plan". Does a perfect project plan exist? If it does not exist, can we be any better than we are now? In the article we will show that we can improve, but the perfect project plan is still somewhere in the future.

General theories describing the performance of artificial learners are of little help when a user is confronted with a selection of datasets and a given artificial classifier. The objective of this paper is to find out the best description of the learning curves produced by a Naive Bayes classification. The performance of Naive Bayes was measured on 121 datasets using k-fold crossvalidation. Power, linear, logarithmic and exponential functions were fit to the data. The exponential function was a better descriptor of the error rate in 44 of 60 useful cases. Average mean squared error is significantly different at P=0,000 from power and linear and at P=0,001 from logarithmic function. The exponential function’s rank is significantly different from the ranks of other models (P=0,000). The results can be used to forecast the future performance of the learner, or to check where on the learning curve the current measurement lies.

Integration of the two best-known quality management models for software production and maintenance, the Capability Maturity Model and the ISO 9001 Model (including ISO 9000-3) is of great interest for both the European and the American software market. The definition of a quality management model which would conform to the requirements of both models and use the advantages of each was a key goal of our research work within the project PROCESSUS.1 Research work was conducted in two steps: first, the detailed comparison of both models was made in order to define their common issues and specifics of each model, and secondly, the integrated model based on the results of the comparison was defined. In this paper the comparison method and results are described. Further, a method for the integration and a framework of the integrated model are presented. The core of our work is actually the content of the integrated model. Due to extensive content of the integrated model examples of some parts only are given in the appendix. The integrated model presented in this paper has already been successfully used in organizations which are cooperating in the project PROCESSUS o these organizations achieved the needed maturity for certification to the ISO 9001 standard and a maturity level near level 3 of CMM.

Adoption of object-oriented technology by the software industry is to a large extent interfered with a lack of appropriate evaluation criteria. Measures of object-oriented systems are necessary in order to realize and achieve expected benefits and advantages of object technology. This paper presents some evaluation criteria, measures and metrics, suitable for object-oriented software development. The implemented prototype of OOMetTool that provides an automatic support for metric data gathering is described and some initial findings are presented.

this paper described an event-driven object-oriented development which enhances reusability and object composition. Simple composition of objects by means of one class named object environment is suggested. Object openness and reusability is increased with designing objects that instead of invoking the behaviour of other objects with sending messages to them inform the object environment about interesting changes of values of its state variables.

Presentation of the potential behavior of the future software system after earliest phases of the software development process is a key problem annotated in this paper. How to do this presentation to a customer team after the requirements analysis or between them? The activation of the system specifications as useful connection between both customer and developer team is described in this paper. Activation gives a dynamical procedural picture of the future system with an optional level of abstraction and with a possibility of tracing a desired event structure. The advantages of described activation of the system specifications are simplicity, well understandability and flexible connection with techniques for generating the system specifications. Improvement of the customer/developer cycle's communication paths in described manner enables the validation of the system specifications on the basis of more than just statically low cross-referenced figures and textual descriptions. All contributions of the subjects in the requirements elieitation process can be confirmed immediately through the activation process.

How to present to the customer team the aim and the potential behavior of the future software system after the earliest phases of software development process - the requirements analysis? Activation of systems specifications as the useful connection between the both customer and developer team is described in the paper. Activation gives the dynamic procedural picture of the future system with the optional level of abstraction and with the possibility of tracing the desired event structure. The advantages of the described activating of systems specifications are simplicity, well understandability and the flexible connection with the techniques for generating system specifications. All this improve the customer/developer cycle's communication paths.

One of the tasks of data mining is classification, which provides a mapping from attributes (observations) to pre-specified classes. Classification models are built by using underlying data. In principle, the models built with more data yield better results (are more accurate). However, the relationship between the available data and the performance is not well understood. How much data to use, or when to stop the learning process, are the key questions. In this paper we give a suggestion as when to stop the learning process.

In this paper we study an evolutionary machine learning approach to data mining and knowledge discovery based on the induction of classification rules. A method for automatic rules induction called AREX using evolutionary induction of decision trees and automatic programming is introduced. The proposed algorithm is applied to a cardiovascular dataset consisting of different groups of attributes which should possibly reveal the presence of some specific cardiovascular problems in young patients. A case study is presented that shows the use of AREX for the classification of patients and for discovering possible new medical knowledge from the dataset. The defined knowledge discovery loop comprises a medical expert's assessment of induced rules to drive the evolution of rule sets towards more appropriate solutions. The final result is the discovery of a possible new medical knowledge in the field of pediatric cardiology.

In the paper a new outlier prediction method is presented that should improve the classification performance when mining the medical data. The method introduces the class confusion score metric that is based on the classification results of a set of classifiers, induced by an evolutionary decision tree induction algorithm. The classification improvement should be achieved by removing the identified outliers from a training set. Our proposition is that a classifier trained by a filtered dataset captures a better, more general knowledge model and should therefore perform better also on unseen cases. The proposed method is applied on the two cardio-vascular datasets and the obtained results are discussed.

In this paper, we present a procedure for data protection, which can be applied before any model building based analyses are performed. In medical environments, abundant data exist, but because of the lack of knowledge, they are rarely analyzed, although they hide valuable and often life-saving knowledge. To be able to analyze the data, the analyst needs to have a full access to the relevant sources, but this may be in the direct contradiction with the demand that data remain secure, and more importantly in medical area, private. This is especially the case if the data analyst is outsourced and not directly affiliated with the data owner. We address this issue and propose a solution where the model-building process is still possible while data are better protected. We consider the case where the distributions of original data values are preserved while the values themselves change, so that the resulting model is equivalent to the one built with original data.

Software reliability analysis is inevitable for modern medical systems, since a large amount of medical system functionality is now dependent on software, and software does contribute to system failures. Most software reliability models are based on software failure data collected from the project. This creates a problem for the designers since, during the early stage, software failure data are not available. However, a valuable knowledge can be learned from the analysis of previous projects and applied to the new ones. This paper presents the approach that predicts the potentially dangerous software modules under development based on the analysis of the already finished modules using the machine-learning techniques. On the basis of the prediction given by our method software designers are able to devote more testing effort to the dangerous parts of the system, which results in a more reliable medical software system.

The most important part of the majority of recent computer software systems, according to the users view, is the user interface. In this paper authors present some new metrics devoted to the assessment of some user interface characteristics.

In this paper, we present some important aspects of software testing. We explain why software reliability is important, especially for medical software products. The automation of testing procedures is important in order to achieve the desired reliability level. For this purpose, intelligent systems can be used to identify potentially dangerous software modules. We present an evolutionary decision-tree approach to the identification of dangerous software modules in a real-world medical application.

The area of intelligent systems is one of the most important and successful. Some definitions state that in some way, intelligent systems are a form of information system. This was successfully introduced from the research domain of artificial intelligence in information systems practice. According to some references, the hardest single part of building any system, either information, software or intelligent, is deciding what to build. No other part of conceptual work is so difficult as establishing the detailed technical requirements. The authors propose reusability to support and influence requirements engineering. Reusability is actually a very simple activity that is not even listed in most dictionaries. In the software environment, requirements, design models, algorithms, test documents and many other products of the software process can be reused. However, our approach is not concentrated on software reusability, but rather on database and intelligent systems reusability which are not so often reused, but they are the main goal of the MetaBase project (T. Welzer et al., 1998). MetaBase is an object oriented repository which provides a structure for archiving reusable database components of several applications not only medical information systems. All possible conceptual models are classified according to business domain and functional domain of their application domains. The structure of the repository allows not only storage of reusable database components of several applications but also classifying and comparing components from different applications (T. Welzer et al., 1998).

The use of PROLOG for digital circuit simulation in the education process is discussed. It is noted that the curriculum of the University of Maribor gives electrical engineering students some basic knowledge of computer techniques and FORTRAN or Pascal. On this basis, PROLOG can be presented to the students in order to acquaint them with some fundamentals of this fifth-generation language. The students who have acquired basic knowledge of PROLOG are able to define digital circuits specification programs as well as other simple PROLOG rules. The experience that students gain makes them capable of understanding similar models of conventional engineering as well as the basic working of expert systems. >

The latest promising technology called web services carries a potential to expand cooperation and collaboration in business networks. There are at least three common deployment types of web services and they are all very similar to current business practices. The economics of web services is heavily dependent on fixed costs thus forcing providers to increase the number of users to reach a point of profitability. This may be a decisive factor in using them with ERP solutions where vendors will have to increase their investments to make web services user-friendlier.

The authors present guidelines that should be followed when designing the curricula for educating engineers who will be capable of managing and evaluating quality in the software design and development process. It is pointed out that quality management is impossible without planning. Consequently, the preparation of planning, which is divided into technical, quality, and resource plans, is of primary importance. In addition to the preparation of plans, a quality system that supports the elaboration of plans should be worked out. However, a quality system cannot be established without knowledge of process models, design methodologies and planning techniques, which requires that one observe standards from the group ISO 9000 and structured approaches. The authors present their experience regarding the level of knowledge that is satisfactory for the computer engineering level and for other predominantly electrical engineering levels, how and where to integrate software quality topics into the curriculum, which courseware to use, and how to persuade students that high quality software is not only code-weighted but that also depends on all attributes that make a software of high quality. >

A methodolgy/courseware for introducing such a quality system in software development organizations is described in this paper. The authors used the ISO standardization (ISO 9000-3: Standard for QA in design/development, production, installation and servicing) as the basis of the methodology and refined it with well known related process models. It is not limited to the standard but also concerns human factors regarding the continual improvements toward the quality system within training programs and group-work. The methodology is supported by an integrated software product which enables software process assessments and goal definitions and offers a hypertext based software-process-improvement knowledge base.

An experiment was conducted to highlight the technicality of two quite different but well known software development methodologies: System Analysis-System Design (SA-SD) and Jackson System Development (JSD). There was no attempt to reach a general conclusion about which software development methodology is better, looking from all aspects. Comparison was done by means of a questionnaire which was completed by 12 young engineers after they had finished their seminar work in postgraduate study. The results obtained speak in favor of the SA-SD methodology because it is more understandable to an engineer. >

The concept of the knowledge-based database designer, which is intended for conceptual database design, is dealt with. Because of its importance, it must be well designed and different tools should be combined to make this design more intelligent. The database design should be supported by expert (administrator know-how, CASE tool results) as well as problem-oriented knowledge of the application. >

Object technology has been recognized as an enabling factor and it is largely introduced into MIS departments. However, the benefits and improvements related to object technology are not obvious and could not be realized and achieved without appropriate evaluation criteria and metrics. The paper presents some evaluation criteria, measures and metrics, suitable for object oriented software development. Basic concepts of the OOMetDaGa Environment that provides automatic support for metrics data gathering are described and some empirical data presented.

In medical decision making (classification, diagnosing, etc.) there are many situations where decision must be made effectively and reliably. Conceptual simple decision making models with the possibility of automatic learning are the most appropriate for performing such tasks. Decision trees are a reliable and effective decision making technique that provide high classification accuracy with a simple representation of gathered knowledge and they have been used in different areas of medical decision making. In the paper we present the basic characteristics of decision trees and the successful alternatives to the traditional induction approach with the emphasis on existing and possible future applications in medicine.