TAAC'2013: Section "Software Engineering"

Unified Modelling Language (UML) is a widely used and accepted software modelling language. However, UML still suffers from many issues such as formality, lack of precision and low support for abstraction. This paper highlights some of the aforementioned issues with comparison to Language for Pattern Uniform Specification (LePUS3).

Recently the approach of patterns recognition has wide application in astronomy, medicine, robotics, remote sensing, etc and it is reduced to classification of the input data set of symbols, text, speech, images, and so on. Systems of linear decision functions are one of the classical methods applied for this purpose and some of the other methods are reduced to it. Its effective and efficient skills acquisition is possible though problem solving technologies focusing the learner's construction learning. The paper deals with a technology for generating individual test-like exercises for visual modeling based on such method.

A success of file distributing peer-to-peer applications such as BitTorrent motivated researches to use this idea for video streaming applications. Due to a real-time data generation nature, heterogeneous behavior of peers and underlying network the main challenge is to develop a robust topology structure and a fast dissemination algorithm. This paper presents a simple, but efficient and completely distributed multicasting algorithm that is called Tailcast. It is based on an idea of building tailed tree topology, which guarantees low stretch and reliability of the network. As compared with the existing tree-based algorithms a time penalty for the peer churn event does not linearly grow with the network size. The results demonstrate robustness and efficiency of this solution.

The task of machine-code decompilation is to reconstruct a high-level-language (HLL) representation, such as C code, from a binary executable file. In present, we can find several such tools that support different target architectures (e.g. Intel x86, ARM, PowerPC). Each of these architectures can be classified either as RISC (reduced instruction set computing) or CISC (complex instruction set computing). However, none of the existing decompilers support another major architecture type — VLIW (very long instruction word). In recent years, VLIW processors are once again on the rise because of their high performance. Therefore, there is a demand of a VLIW-supporting decompiler. In this paper, we depict the unique features of the VLIW architecture and we present several approaches how to handle them in the decompilation process.

In this paper we solve on GPUs massive problems with large amount of data, which are not appropriate for solution with the SIMD technology. For the given problem we consider a three-level parallelization. The multithreading of CPU is used at the top level and graphic processors for massive computing. For solving problems of cluster analysis on GPUs the nearest neighbor method (NNM) is developed. This algorithm allows us to handle up to 2 millions records with number of features up to 25. Since sequential and parallel algorithms are fundamentally different, it is difficult to compare the computation times. However, some comparisons are made. The gain in the computing time is about 10 times. We plan to increase this factor up to 50-100 after fine tuning of algorithms.

This paper describes an approach that can be used to build flexible in setting up a decision making information system for personal lending. The peculiarity of such an information system (IS) is that banking subproduct is automatically selected for clients; meets to the fullest extent the requirements of the client and at the same time is the most profitable for the bank. Such IS allows to optimize the charges of the Bank and minimize Time-To-Yes.

The graph theory taught generally in discrete mathematics, mathematical modeling and similar courses is applied mainly for structural and informational modeling in many subject domains. For this purpose both teachers and learners need a problem-solving technology with powerful tools for preparing and performing test-like exercises. The paper is aimed to fill this gap presenting the teacher's side of a technology for generating individual test-like exercises for modeling with graphs.

According to UNESCO, 2002 "An instructional technology can be used to support traditional forms of learning or transform them into innovative and creative ones". The student-centered activities that allow 3D geometry instruction/learning incorporating 2D geometry knowledge and experimentation are especially important as 3D spatial knowledge is difficult for most of the students. The paper focuses on the instructor's tool of a technology for generation of individual exercises of 3D geometry tasks.