Vineet Sinha

The role of context in question answering systems

Despite recent advances in natural language question an-swering technology, the problem of designing effective user interfaces has been largely unexplored. We conducted a user study to investigate the problem and discovered that overall, users prefer a paragraph-sized chunk of text over just an exact phrase as the answer to their questions. Fur-thermore, users generally prefer answers embedded in con-text, regardless of the perceived reliability of the source documents. When users research a topic, increasing the amount of text returned to users significantly decreases the number of queries that they pose to the system, suggesting that users utilize supporting text to answer related ques-tions. We believe that these results can serve to guide future developments in question answering user interfaces.

Using dependency models to manage software architecture

This demonstration will present a new approach, based on the Dependency Structure Matrix (DSM), which uses inter-module dependencies to specify and manage the architecture of software systems. The system is decomposed into a hierarchy of subsystems with the dependencies between the subsystems presented in the form of an adjacency matrix. The matrix representation is concise, intuitive and appears to overcome scaling problems that are commonly associated with directed graph representations. It also permits succinct definition of design rules to specify allowable dependencies.A tool, Lattix LDM, will be used to demonstrate this approach by loading actual open source Java applications to create DSMs that can represent systems with thousands of classes. We will show how algorithms can be applied to organize the matrix in a form that reflects the architecture and highlights problematic dependencies.We will demonstrate how design rules can be used to specify and enforce architectural patterns such as layering and componentization. We will examine the evolution of architecture by creating dependency models for successive generations of Ant, a popular Java utility. Finally, we will explore the application of this approach to the re-engineering of Haystack, an information retrieval system.

Haystack: a platform for creating, organizing and visualizing semistructured information

1. OVERVIEW In this demonstration we present Haystack, an environment that allows users to easily manage their documents, e-mail, appointments, tasks, and other information [1]. Haystack uses a semistructured data model to describe the connections between different documents in a user’s corpus as well as the metadata concerning each document. This amalgamation provides users with a unified framework for managing all of their information, e.g., documents, e-mails, etc., through a single interface. Furthermore, Haystack’s user interface exposes general tools for navigating the various kinds of information found in users’ corpora.

Understanding code architectures via interactive exploration and layout of layered diagrams

Visualization tools that target helping developers understand software have typically had visual scalability limitations, requiring significant input before providing useful results. In contrast, we present Strata, which has been designed to actively help users by providing layered diagrams. The defaults used are based on the package structure, and user interactions can allow for overriding these defaults and focusing on relevant parts of the codebase.

Interactive exploration of compacted visualizations for understanding behavior in complex software

In this demonstration we present Chrono, a tool that creates sequence diagram based visualizations. Since the diagrams produced by traditional sequence diagramming tools become large and unmanageable when dealing with complex code bases, Chrono focuses on removing less relevant information, condensing diagram components, and allowing for interactive exploration.