Neunghoe Kim

A Quality Attributes Evaluation Method for an Agile Approach

Rapid change of customer requirements has propelled the emergence of an agile approach due to such merits as fast release and simplified documents. An agile approach aims at maximizing productivity and effectiveness. Participants in an agile project usually focus on functionality of a system during the process. However, there are few considerations for nonfunctional requirements (or quality requirements) such as performance, security and accuracy. This causes huge costs, and wastes efforts for later changes. A novel approach is required to strengthen non-functional aspects in an agile approach. We propose a lightweight quality evaluation method that reflects quality attributes to enhance non-functional features in an agile approach. Our approach enables iterative evaluation of quality attributes in each release. It supports participants to continuously consider quality issues. In addition, participants are able to make plans on quality improvements.

Customer Value-based HW/SW Partitioning Decision in Embedded Systems

In launching a product, requirement change is always risk to an embedded system designer. Including requirements change, limited resources (e.g. development time and cost) are also risk factors to designing a system. However, the time to market constraint among several limitations is most critical especially to embedded system products (e.g. cellular phone, PDA, etc.). For that reason, many strategy makers try to find effective way to a HW/SW partitioning decision in early designing phase. Only ideal way to reduce requirements change and satisfy customers is reflecting their value to decision making steps in early phase prior to fixing HW and SW component design. In this paper, Customer Value based Partitioning Decision (CVPD) method is proposed to identify, analyze, and calculate the value of the customerspsila requirements, reflecting the value on the partitioning decision making process.

EURECA: End‐user requirements engineering with collaborative animation

In recent years, software development environments have changed from being driven by professional developers to being driven by technical end users. One of the key issues in end‐user‐driven software development environments is how to guarantee the quality of the software artifacts. Measuring the quality of developed software requires the correct specification of a quality range the software is expected to meet. However, technical end users are non‐professionals in engineering techniques for software requirements, and training a developer to be an expert in requirements engineering in a short period of time is difficult. This paper proposes a new software development life cycle based on reutilizing previously evaluated software requirements assets from completed projects. End‐User Requirements Engineering with Collaborative Animation, a tool that was developed to support the proposed software development life cycle, is described and demonstrated by application to three projects. The results from real cases are used as the basis for a discussion on the efficiency enhancement of requirements work, an increased rate of reusing legacy software requirements assets, and an improvement in the technical end users individual competency level using the End‐User Requirements Engineering with Collaborative Animation. Copyright

Situation-awareness overheating protection solution for mobile devices

Current mobile devices have high performance that eventually cause high power consumption and overheating. To prevent damage to the system and enhance user safety, we propose an overheating protection solution with situation-awareness to reduce power consumption. The experimental results demonstrate the proposed solution can reduce temperature to a predefined threshold.

Probabilistic Approach to Predicting Risk in Software Projects Using Software Repository Data

Although the factors that need to be focused on for a successful software project appear to be difficult to define, risk management has become one of the key activities for achieving such success because significant risk is involved in each software development phase. Software project failures are often a result of insufficient and ineffective risk information regarding the future. To overcome this, software risk prediction should be performed in advance to allow project managers insight into providing more valuable information for decision making, such as scope coverage, resource allocation, and schedule changes. In this research, we propose a risk prediction model from the perspective of quality using a software repository. We evaluated the risk threat level by mapping some defect attributes that exist in the defect lifecycle, defined their risk threat transition states, and applied a Markov chain for predicting the potential risk level. We evaluated the proposed approach using practical real-industry mobile software projects. The experimental results confirm that our approach is applicable to software threat risk estimation.