Randy K. Smith

Value-Oriented Requirements Prioritization in a Small Development Organization

Requirements Engineering, especially requirements prioritization and selection, plays a critical role in overall project development. In small companies, this often difficult process can affect not...Requirements engineering, especially requirements prioritization and selection, plays a critical role in overall project development. In small companies, this often difficult process can affect not only project success but also overall company survivability. A value-oriented prioritization (VOP) framework can help this process by clarifying and quantifying the selection and prioritization issues. A case study of a small development company shows a successful VOP deployment that improved communications and saved time by focusing requirements decisions for new product releases on core company values

Cyclomatic Complexity and Lines of Code: Empirical Evidence of a Stable Linear Relationship

Researchers have often commented on the high correlation between McCabe’s Cyclomatic Complexity (CC) and lines of code (LOC). Many have believed this correlation high enough to justify adjusting CC by LOC or even substituting LOC for CC. However, from an empirical standpoint the relationship of CC to LOC is still an open one. We undertake the largest statistical study of this relationship to date. Employing modern regression techniques, we find the linearity of this relationship has been severely underestimated, so much so that CC can be said to have absolutely no explanatory power of its own. This research presents evidence that LOC and CC have a stable practically perfect linear relationship that holds across programmers, languages, code paradigms (procedural versus object-oriented), and software processes. Linear models are developed relating LOC and CC. These models are verified against over 1.2 million randomly selected source files from the SourceForge code repository. These files represent software projects from three target languages (C, C++, and Java) and a variety of programmer experience levels, software architectures, and development methodologies. The models developed are found to successfully predict roughly 90% of CC’s variance by LOC alone. This suggest not only that the linear relationship between LOC and CC is stable, but the aspects of code complexity that CC measures, such as the size of the test case space, grow linearly with source code size across languages and programming paradigms.

Enhancing the Cocomo estimation models

In software engineering, team task assignments appear to have a significant potential impact on a projects overall success. The authors propose task assignment effort adjustment factors that can help tune existing estimation models. They show significant improvements in the predictive abilities of both Cocomo I and II by enhancing them with these factors.

Feature location by IR modules and call graph

When different types of test are performed on software, from unit test, to component test to system test many bugs can be detected and recorded in bug reports. Developers must then fix them one by one. However, an important job before fixing bugs is to locate them in source code. Given a large scale software project with hundreds of bugs, it is a tedious job to locate the problems in source code. Feature location is a solution of this problem. Feature location seeks to identify pieces of source code corresponding to a specific feature, where a feature is defined as a function in software. Since bugs have the same attributes as features, they can be treated as features. In this paper, we provide a technique to achieve feature location. The approach uses a combination of lexical information and structural information. We combine Latent Semantic Indexing with Call Graphs to on a small test case to assist in feature location. Comparing our approach to an approach that uses LSI shows improved accuracy ad effectiveness.

Security for fixed sensor networks

Sensor networks consist of resource-constrained sensors operating in a variety of environments. Given the severe constraints on these sensors, it is a particularly challenging problem to choose and design valid security protocols for such networks. This problem has recently given rise to new research addressing the security issues. This paper presents an overview of the important works, specifically the new mechanisms and protocols, which have been introduced or are still under development in this area.

Variable selection and ranking for analyzing automobile traffic accident data

Variable ranking and feature selection are important concepts in data mining and machine learning. This paper introduces a new variable ranking technique named Sum Max Gain Ratio (SMGR). The new technique is evaluated within the domain of traffic accident data and against a more generalized dataset. In certain cases, SMGR is empirically shown to provide similar results to established approaches with significantly better runtime performance.

PerpSearch: An integrated crime detection system

Information technologies such as data mining and social network analysis have been widely used in law enforcement to solve crimes. Recent research indicates that geographic profiling also plays an important role in facilitating the investigation of crimes. However, lack of integration makes those systems less helpful in practice. In this paper, we propose an integrated system called PerpSearch that will take a given description of a crime, including its location, type, and the physical description of suspects (personal characteristics or vehicles) as input. To detect suspects, the system will process these inputs through four integrated components: geographic profiling, social network analysis, crime patterns, and physical matching. Essentially, geographic profiling determines “where” the suspects are, while other components determine “who” the suspects are. We then process the results using a score engine to give investigators a ranked list of individuals. To date, we have implemented a prototype of the system based on current Alabama law enforcement data.

A Web-based process and process models to find and deliver information to improve the quality of flight software

Aerospace systems demand high-quality software engineering processes to deliver high-quality products. Although most aerospace organizations have high-quality processes, many of these processes fail to deliver to the engineer the organizations wealth of information and experience nformation and experience that can further contribute to the quality of software products and engineering processes. In this paper, we present an interactive, Web-based process support tool that delivers the information in a flight software engineering process as well as associated standards, lessons learned, and background information. The tool is based on an underlying formal model of the software engineering process activities and artefacts. This underlying model provides a semantic basis for context-based search and for reasoning about the engineering process. The result is an information portal to search for and deliver process and project-specific information to support the development of flight software.

Combining lexical and structural information for static bug localisation

In bug localisation a developer uses information about a bug present in a software system to locate the source code elements that must be modified to correct the bug. Researchers have developed static bug localisation techniques using Information Retrieval techniques such as Latent Semantic Indexing (LSI) to model lexical information from source code. In this paper we present a new technique, LSICG, that combines LSI to model lexical information and call graphs to model structural information. A case study of 21 bugs in Rhino demonstrates that our technique provides improved performance compared to LSI alone.

Improved variable and value ranking techniques for mining categorical traffic accident data

The ever increasing size of datasets used for data mining and machine learning applications has placed a renewed emphasis on algorithm performance and processing strategies. This paper addresses algorithms for ranking variables in a dataset, as well as for ranking values of a specific variable. We propose two new techniques, called Max Gain (MG) and Sum Max Gain Ratio (SMGR), which are well-correlated with existing techniques, yet are much more intuitive. MG and SMGR were developed for the public safety domain using categorical traffic accident data. Unlike the typical abstract statistical techniques for ranking variables and values, the proposed techniques can be motivated as useful intuitive metrics for non-statistician practitioners in a particular domain. Additionally, the proposed techniques are generally more efficient than the more traditional statistical approaches.