VTURCS Abstract Submissions
These are the different abstract submissions for the several Spring Research Seminars. We are working to bring up to date this site with previous year submissions.
- 1. Hallway Display
Robert Lewis <ralewis@vt.edu>
Steve Harrison - This project uses a bar code scanner to read student ID cards as well as a 42" plasma television to display teaching assistances that are currently in their offices. When a student's ID card is read, a list of teaching assistances currently in their offices, for which the student has signed up online, will be displayed. There is also a station set up in the Teaching Assistants area, for them to swipe their ID cards to login within the system.
- 2. Creating Interactive Real World Information Systems
Andrew Zimmer <bob273@vt.edu>
Scott McCrickard - Emerging technologies such as Wi-Fi, cell phone signals, and RFID tagging enable accurate assessment of current location. While prior work has focused on some of the technical issues regarding the processing of this data, a remaining challenge is to develop usable interfaces that address real needs of users. This work seeks to create an environment in which users interact in both an informative and social way with the objects in their space and the other people around them. We address three key challenges in this work: the creation of an information architecture for client-user-location data sharing, the automation of the collection and sharing of location-specific and object-specific data, and the design of a forum for sharing information among all people in an environment. Data collection and sharing occurs with the Whereware location service (developed by the author), merged with the TagIT RFID service, developed in our lab, to create an information sharing architecture. Built on top of this construction is an interface for displaying a map of the environment and research posters in the area to enable visitors easy access to information, the ability to leave and respond to feedback, and the option to virtually revisit previously viewed projects. Our continued efforts will result in a widely-deployed, robust information sharing system that will assist visitors in their knowledge collection efforts.
- 3. Binary Search Tree Algorithm Visualizations
Andrew Whitaker <aawhitak@vt.edu>
Dr. Clifford A. Shaffer - One goal of this project was to create a visual tutorial for binary search trees and their variants. There are many binary search tree visualization tools in existence, but few, if any, give the student the ability to step through each tree operation step by step. This feature of the visualization improves students’ in-depth understanding of each tree operation. Instead of specifying data values and sitting back and watching the tree populate, the user must step through each operation.
Another goal of the project was to show students the relative performance of each tree structure. The student would load a specific number of data into each type of tree and compare performance of various tree operations.
Our research has created a Java application that shows, step by step, the details of each binary search tree operation. The application is modular and extendable, and should enable easy addition of other tree types and operations. Since only the BST portion of the project has been completed, there is lots of room for addition and improvement. - 4. Visualization of Eqrthquake Simulation Data using Volume Raycasting
Colin Braley <cbraley@vt.edu>, <>
Professor Yong Cao, - This project, done in conjunction with University of California, Santa Barbara (UCSB), is aimed at better visualizing and predicting the impact of an earthquake in southern California. Researchers at UCSB have used sophisticated sensors to detect movement in the earth’s crust in several areas around California. The researchers have then used a computer simulation to estimate the movement of the Earth’s crust at other points on the surface, and at points below the surface of the Earth.
Now that the researchers have gigabytes of data, they need a way to visualize it in order to fully understand it. This is where Virginia Tech comes in. The goal of our research is to create a computer program to allow users to visualize the data as it changes through time. Furthermore, we wish to allow users to manipulate and rotate around the data intuitively in a three dimensional environment. The software is designed to be as close to real-time as possible, by taking advantage of spatial data structures and parallel processing.
In order to accomplish this goal, we employ a rendering technique known as Volume Raycasting. With this technique, we are able to render out visualizations of the data as a whole, as well as isolate and render specific iso-surfaces of a certain value. Volume Raycasting is used in a variety of fields, most notably Medical Imaging and Computational Fluid Dynamics
Volume Raycasting is a general purpose technique for creating 2d images from discreetly sampled 3d data, which can consist of either vectors or scalars. The data, hereafter referred to as voxels is centered about a point in 3d space. The viewpoint of the user, hereafter referred to as the camera is another point in 3d space. Numerous rays are then generated, one for each pixel in the image, each one leading from the camera and towards the voxel data. Each ray is iterated down, and the voxel value is sampled at each iteration, according to some transfer function. All these iterations added up form a voxel value for each ray. Finally, a mapping is used to map voxel data to colors; this is how the final image is colored. - 5. Cxx for Visual Studio
Jaime Garcia Ramirez <jgarciar@vt.edu>
Dr. Stephen Edwards - Cxx for Visual Studio came about from the need of having test driven development in introductory Computer Science courses. Virginia Tech’s Computer Science curriculum currently puts great emphasis on this subject, and many of the tools used to teach these courses have features that facilitate testing. For example BlueJ, used in introductory Java, is closely integrated with JUnit tests. Eclipse, used both in introductory Java and C++ courses, also has JUnit capabilities as well as Cxx tests. Cxx is a test framework that allows C++ code to be run against a series of user created tests; they work in a similar way to JUnit and the user is able to receive feedback such as failures and run-time errors. The plug-in for Cxx for the Eclipse IDE was created at Virginia Tech to allow students to be able to efficiently use the test framework, integration with the IDE allowed for a shallower learning curve than if the students tested from command-line. Having IDE integrated tools allow the user to be more productive in their tasks, especially if these tools automatically handle repetitive tasks. For this reason the Cxx plug-in for Eclipse is very helpful.
Eclipse however is not the only IDE in the market; another very commonly used IDE is Microsoft’s Visual Studio. Many students express interest of why Virginia Tech does not use Visual Studio for their Computer Science courses. One reason is it lacks tools that provide the same functionality we have now. Thus, I have created a plug-in for Visual Studio that mimics and expands the features of its Eclipse cousin. Features of the Cxx for VS plug-in include automatic generation of skeleton test classes, selective test running, debugger integration, and graphical outputs of test results. These features provide the necessary information for students to be able to write tests rapidly and correctly. The debugger integration allows for the easy tracking of bugs, and selective testing can decrease the amount of time a student spends on debugging. Written as a Visual Studio Integration Package, the plug-in can be distributed to any Visual Studio user, and installation is done by a set-up program. The Cxx for VS plug-in will open the door for more IDE’s to be used for educational purposes. - 6. Rubric Design Tool
Lauren Smith <smithls@vt.edu>
Dr. Stephen Edwards, Dr. Manuel A. Perez-Quinones - This project entailed the design and creation of a rubric tool. This tool allows instructors to create and reuse rubrics. The rubric structure consists of several layers: categories, questions and responses. Categories contain questions and help in the organization of topics for the rubric. Questions are specific inquiries as to the content of an assignment. To evaluate an assignment, a question has a rating and a list of check boxes. The check lists allow the evaluator to “check” each task in the list that has been completed to satisfaction. Checklists give a visual lay out of how well the assignment has been completed and aid in making a decision as to which rating to give. The rating is a set of radio buttons that allow a value to be given to each question. Ratings can be anything from “good” to “bad” or 1 to 100; their purpose is help with assigning grades to the rubric. A response to a question contains the selected answers for checklists and ratings.
The development of the rubric entailed many design considerations; dilemmas involving everything from the hierarchy of the rubric to how rubrics were to be saved had to be resolved. The specifics of the rubric were determined and a prototype produced. The prototype created is being integrated with Moodle, a course management system, to offer another option for grading assignments. Many issues arose during development, including the versatility of changing the sequence of categories and limitations involved in integration with Moodle. Despite the limitations, the development of this rubric is unique, because there is no current system that allows for the customization and reuse of rubrics and their entities. - 7. Dynamically Instrumenting OS Kernels
Anthony Romano <chz@vt.edu>
Dr. Godmar Back - Dynamic instrumentation can be applied to a variety of problems, including bug analysis, visualization, and performance tuning. Although a great deal of effort has been given to user-land instrumentation frameworks and analysis tools, kernel level instrumentation remains incredibly specialized and difficult to use effectively.
This project focuses on the implementation of dynamic code analysis tools built on top of the qemu system emulator. Besides detecting race conditions, uninitialized data accesses, stale pointer violations, and data invariants, these tools motivated the creation of a target operating system aware instrumentation framework, thus enabling the rapid development and succinct expression of new dynamic analysis utilities. - 8. Analyzing Relevant Words in Web Page Paths
Gregg Tabot <gtab@vt.edu>
Dr. Lenwood Heath - The Internet is easily the world's largest network. With its complex architecture of interlinking web pages and its paths, it can be difficult to make sense of any obtainable information. A particular set of data is useful when analyzing web pages and their paths and that is the textual content of the web page. Analyzing the text of the web page not only gives you data to keep track of but it also can be used to identify the subject of the web page. The purpose of my project was to keep track of the “relevant words” of a web page while analyzing how those words appear throughout the path of several web pages.
In order to better analyze the data, I came up with a key term called “relevant words.” Relevant words are words that have been deemed significant in context to the actual web page. In order to become relevant, the ratio of a word's occurrence on the web page is compared to the ratio of its occurrence in data collected by the Linguistic Data Consortium's “Web 1T 5-gram Version 1” web collection data set. This data set is intended for statistical language modeling and was generated from approximately 1 trillion words from publicly accessible web pages. If the word occurrence ratio found on the web page was over 100 times that of the data set, the word was considered relevant. The higher the relative ratio, the more relevant the word is regarded. - 9. LibX 2.0 A composable publish-subscribe framework for client-side mashups.
Tobias Wieschnowsky <frostyt@vt.edu>
Dr. Godmar Back - The goal of LibX 2.0 is to allow it's users to redesign the way they view and use the Internet, by providing a new kind of web localization/mashup script framework. What makes LibX 2.0 unique and more powerful then other scripting frameworks for mashups is the composition of LibX scripts out of reusable modules and the easy publish/subscribe features for script management.
Splitting scripts into modules that each have clearly defined functionality and then allowing the reuse of these modules in other scripts, eliminates a lot of code duplication that normally occurs when writing similar scripts or scripts that share some part of their functionality. Modules allow for quicker creating of such scripts since the duplicate parts only need to be written once and the module can then be reference in the other scripts instead of recreating the code.
Another unique features of LibX 2.0 is the way scripts are managed, using a simple publish/subscribe system. This offers a number of advantages for both users and authors of web localization scripts and mashups. Authors can easily and quickly distribute hot fixes and updates by simply updating the script. For users this has the advantage that their scripts are always up to date and they never have to go out and search for the most recent versions of the scripts they use. We also support the publication of packages of scripts that can be subscribed to just like a single script.
More features of LibX 2.0 include a tuple space model for module communication as well as a sandbox and a white list for security purposes. - 10. CATspace: Using Facebook to Share Assignments Among Professors
Tarish Smith <tarish@vt.edu>, Idris Eltahir <dingis5@vt.edu>, Ricardo Quintana-Castillo <quintana.ricardo@gmail.com>
Dr. Manuel A. Perez-Quinones, Dr. Stephen Edwards - The CATspace research project involved designing and prototyping a new social networking application. We chose to use Facebook, because its social network is already established around academia, and its application API allows external applications to be integrated with Facebook. The CATspace application allows computer science educators to post, exchange, rate, and reuse complete course assignments (assignment write-ups, grading rubrics, solutions, etc.). Our team met weekly to collaborate on the design and iteratively refine the prototype. We implemented that data model using Ruby on Rails. Additionally, we used FBML, CSS, and AJAX to implement the user interface. Navigation uses tagging, tabs, sorting options, and a CATspace news feed and profile notifications providing links to the affected assignment. Both a partial and an expanded view of each assignment is available, along with an option to download the assignment or install it on the sister program – Web-CAT. The next stage will include implementing the ability to instantly edit uploaded assignments, add additional search options, develop the community features, and launch the application.
- 11. Using Augmented Reality to Garner More Information About the Real World
Sergio Bernales <bernales@vt.edu>
Joe Gabbard - Augmented reality has been a minor part of Computer Science applications for some time now, but has immense potential. Now that we have micro-sized powerful hardware, this can change. Using GPS to know where one stands, a digital compass to know where one is facing, accelerometers to know one's pitch, and a camera-enabled PDA combining all these parts, we have a portable Augmented Reality device providing a virtual representation of our surroundings. High quality graphics can be overlaid onto the video feed of the PDA showing properties of the world not seen with the naked eye. This can enhance activities like a simple walk on campus or improve ideas of future projects at certain locations by allowing people with these devices to give feedback about what they see. It can easily be used by anyone because all it takes is pointing the PDA's camera towards the spot one is interested in and as you get closer the graphics or images would update accordingly. With its ease of use, portability, and endless possibilities of types graphics, images or simple signs that could be overlaid on a PDA video feed, this device could provide robust communication opportunities.
- 1. Automated Analysis Tools for Algorithm Visualization
Sean Ponce <ponce@vt.edu>
Dr. Clifford A. Shaffer - Algorithm visualizations provide visual representations of abstract concepts relating to algorithms and data structures. They have been used by Computer Science educators since the 1980s. Today, numerous visualizations are available through the internet. However, a vast number of these visualizations are of low-quality due to the limited understanding of what makes a good visualization. Without indicators to objectively identify which visualizations benefit students the most, rating them is largely subjective. Until now, there has not been an organized collection of visualizations available to educators, nor a ranking scheme to indicate which visualizations are of good quality. This project seeks to gather information about currently available visualizations. By creating a Visualization Wiki, we hope to start a community of developers and users that will lead to more effective use of algorithm visualization in the future.
The wiki is great for sharing visualizations, but there is little support to gather data and statistics about the visualizations. Using automated tools, it is possible to formulate a question, then generate a graph in a matter of minutes. We can answer our research questions using the current data on the wiki, or previously downloaded data that existed on the wiki. This allows for easy analysis over time, as the wiki grows and changes. - 2. Speech and Gesture Interface for Gigapixel Interaction: A Computer Vision Approach
Blake Sawyer <basawyer@vt.edu>
Dr. Francis Quek - The goal of SGIGI is to create an intuitive and robust system for interacting with a gigapixel display using computer vision techniques. The system will utilize both a speech and gestural command set. Gesture recognition will make use of Hidden Markov models. HMM's give accurate results with little processing time, but require large data sets for training. The system will use stereo-vision to track the position and angle of a user's arm for deictic direct manipulation. Direct manipulation combined with an intuitive command set will make the interface disappear. Current research at Virginia Tech includes a wand pointing device that allows for easy direct manipulation, but requires an expensive infrared Vicon system. This system will use two overhead cameras and a computer to calculate the position of the arm.
SGIGI will address many areas of research. Processing two video frames to track and recognize gestures requires a lot of time and memory and the system must take this into consideration. A user interface for a gigapixel must be able to track and recognize gestures in real time.
Human-Computer Interaction will become important in how the system is designed. Users may or may not be able to point, gesture and speak commands at the same time. It will be interesting to see how a user can adapt to this new interface and perhaps improve their work methods on a gigapixel display. - 3. Extending the Pintos Operating System to Real Machines
Anthony Romano <chz@vt.edu>
Dr. Godmar Back - Pintos is an instructional operating system designed for operating system courses. By actually adding various facilities to an existing, but simple, OS, students gain a better grasp of fundamental operating system concepts than by a less hands-on approach. However, due to the nature of Pintos's architecture, execution on standard hardware is impractical; it is primarily booted through PC emulators such as bochs and qemu. This is unrealistic and ultimately limits the possible projects that can be implemented without significantly modifying the base operating system. My research aimed to improve Pintos such that booting on real computers would be feasible in a lab environment.
My research resulted in several key changes to Pintos. The bootloader no longer assumes to be booting an IDE device. A generic block layer abstracts disk device assignments away from specific hardware details. Specifically, the OS can now utilize non-IDE disk devices. Furthermore, the system now recognizes disk partitions. The PCI bus is now supported, allowing for PCI-based device drivers to be developed. Finally, I have introduced a basic USB stack, along with appropriate drivers, enabling Pintos to run directly from a USB flash drive. Students may now easily load a disk image onto a flash drive and boot the OS on a physical PC. - 4. Using Content Syndication to Distribute Academic Publications
Jessica Whitley <jaydub@vt.edu>
Dr. Manuel A. Pérez-Quiñones - Professors often struggle to keep their publications lists up to date. The publication lists often appears on their CV, personal website, graduate student websites, research center websites, funded project pages, and the departmental website. It is easy to see why it is difficult to maintain all of these updated and consistent. But, it does not need to be that way. Digital information, such as publication lists, is easy to store, replicate, transfer and manipulate. This flexibility allows for content syndication, the idea that information is created at a source location but easily and openly disseminated to other locations. Each new location that uses the information presents it in a suitable format and provides links back to the original information. This is done without any human intervention. Blogs using RSS have made content syndication very popular. RSS is also being used to improve the user experience of humans interacting with electronic workflows.
In this project, I have explored one more possible use for content syndication. I have developed a centralized information management coupled with flexible and personalized information display. This project features one such implementation as a way to study how this architecture could be used to support a publication report center at the Center for Human Computer Interaction. The list of publications is maintained in a central database at the CHCI. The central publication repository provides the usual functionality for maintaining, browsing, and searching publications. It also provides personalized views of each faculty or student and views for particular funded projects.
With content syndication, however, the central repository also provides for ways for the professors to have the publications list updated in their own personal websites. This is done by having a small JavaScript embedded in the professor’s page. The JavaScript is dynamically generated and it includes into the web page the list of publications for that professor. This is dynamically done at the time the professor’s page is loaded, so it is always up to date. The system also provides RSS syndication for inclusion into other sites. We expect to deploy this in the CHCI website this summer. - 5. Extending JES To Aid Classroom Interaction
Aaron Lindsay <aaron.lindsay@vt.edu>
Dr. Manuel A. Pérez-Quiñones - The Jython Environment for Students (JES) is a development environment used in the Media Computation course at Virginia Tech that enables students to easily learn how to manipulate media, especially images. During class students are often encouraged to experiment with image manipulation. However, there is currently no good way for the professor to quickly and easily monitor the students' progress. My research involves creating an extension to JES that allows students in the course to easily send their pictures to the professor's computer. This allows the professor to review the pictures the students create, and show examples to the class.
My design involved creating a multi-threaded server that runs on the professor's computer and a matching client that is used by the students to send their manipulated images to the professor. I implemented the server and client in Java and integrated them into the existing JES platform by writing wrapper functions in Jython. This makes the functionality of my extension available as just another command in the JES interface. My server / client pair makes it quick and easy for students to send their pictures to their professor for review. - 6. Interactive News Resources
Lauren Smith <smithls@vt.edu>
Dr. Andrea Kavanaugh, Dr. Manuel A. Pérez-Quiñones - This work is part of the Digital Government research project at Virginia Tech. I have created a news portal website with functionality that will act as a prototype for future websites to be used in the local community of Blacksburg and the New River Valley. The website will have several coordinated public displays placed in areas around Blacksburg and the NRV. The displays will allow citizens to become more informed on local issues and news. Citizens will be able to read articles from the internet on the plasma displays placed around town. These display screens show news articles organized by popularity, determined by the frequency which each article is read on the news portal website. People in the county would then be able to see what articles are most important to their fellow citizens. This interaction is key to increasing the informed population.
The goal of this project is to create a functional way of ordering news articles for a website so that users can quickly see the most popular local news stories. The website works similar to many social filtering sites; users determine which news articles appear first by clicking the links to the articles. This allows the users to quickly go to articles have piqued the interest of others. These news feeds can be viewed on both the website and the display screen.
On the webpage for the news display, I have turned an RSS feed page into an interactive, responsive page. The RSS feeds from several different sources are read and combined. They are then sorted by the amount of visits each has received. A separate class records, retrieves and stores the number of clicks each news story receives. When the user visits the website and reads a news story, the website redirects to an intermediate page to take note of the visit. It then immediately takes the user to the requested page, and the next time the entire RSS feed page is visited, the page will reflect the visit to the page. Any changes in popularity status are also reflected in the plasma display. - 7. Map Interaction on Mobile Devices
Doguhan Uluca <duluca@vt.edu>
Dr. Scott McCrickard, Miten Sampat - This study aims to find the correct interaction approach between a user and a map, using a mobile device, specifically in a walking user scenario. Two automated approaches, Auto-Pan and Auto-Zoom, were designed to understand what level of flexibility, intuition, and focus an interface must offer to be successful. In a building or a campus wide scenario the pilot study suggests that the Auto-Zoom approach would be the correct interaction approach. This implies that users prefer to be presented with more choices on a sub-conscious level and that they complement the missing features of a user interface without realizing it. With these observations in mind future designs could be fine tuned to be less complex and seamless in the way they interface between the device and the user.
- 8. Rubric-based Hand-grading in Web-CAT
Jaime Garcia Ramirez <jgarciar@vt.edu>, Crystal Weil <ckweil@vt.edu>
Dr. Stephen Edwards - Web-Cat is an automated assignment grading system that already provides a rich interface for course staff to manually “mark up” student code with comments, suggestions, and point deductions, and already lets instructors choose their own balance between the amount of credit allocated to hand-grading and the amount devoted to the automated analyses provided by the assessment plug-ins they have chosen to use. However, many instructors prefer to use more structured strategies for hand-grading rubrics as the basis for performing even the automated portions of an evaluation.
This research has created a prototype with several helpful grading features including drag-and-drop comments, filtering, and summarizing of material. There are three tabs of content: file, summary, and rubric. The file tab contains the student’s submission which the reviewer can make comments to. The summary tab organizes the reviewer’s comments by file or by comment type allowing the reviewer to make an informed decision on the overall grade or feedback he or she wishes to leave for the student. The rubric tab contains the grading and overall commenting of each rubric category. Each rubric definition can be collapsed or expanded depending on the reviewer’s preference. The grading tool can be customized for each assignment or class by changing the rubric categories and the list of default comments. Ease of use and customizability were major focuses in this research. - 9. Loch Project: Underground Cave Mapping and the Benefits of Immersive Environments
Philip Schuchardt <ohc@vt.edu>
Dr. Doug Bowman - Underground cave systems can have complex structures, which have many stacked levels and connections between those levels. Traditional 2D cave maps cannot easily describe or model these structures. The purpose of this research was twofold: first, to explore the use of 3D models for visualization of underground cave structures, and second, to determine the benefits of immersive environments for viewing such models. We collect cave survey data and create a 3D model; the cave models are then loaded into an immersive environment, the VT-CAVE. We asked a set of domain experts similar questions in more immersive and less immersive conditions. We evaluated subjects accuracy, comprehension, and speed in performing these tasks. The results show that the immersive environment significantly improves accuracy, performance, and comprehension over the non-immersive environment, and that 3D visualization overall is a good match for the cave data.
- 10. Displaying and Establishing Relationships in a Case Study Library.
Gregory R Smith <sgr07a@vt.edu>
Laurian Vega, Dr. Scott McCrickard - One of the more cumbersome problems in establishing a case study library, to teach students about design, is the difficulty in automating case creation. One way that case creation may be able to be automated is by providing designers a useful set of design tools. These tools will track artifacts created by the designer throughout the course of a project. This will allow the designer to record crucial junctions, make note of what phases of design take a particularly long time, and streamline future design. If an environment is provided that a designer wants to use hopefully he or she would be more likely to use it and by doing so provide educators a repository of design data. This design data will provide educators the backbone for complete case studies that can in turn be used in the class room.
One of the more important aspects of teaching design is conveying how and when the included artifacts evolved. This became the focus of our research. A website was created in which a user could create cases (projects) and then enter artifacts which he then could track and modify. Tools were created to create relationships between these artifacts. Particularly in depth relationships were established to relate claims to claims and scenarios and claims. The goal is to allow both the designer and student to observe in which way certain aspects of a design changed and for what reason.
As discussed this is only a single part of the suite of tools that will be offered to designers through this environment. Further development of these tools and design environment is planned for the summer of 2007 with human trials set for the fall. To what amount these tools aid aspiring designers, and whether or not our design contributes further to the usefulness of design case studies, are questions we will explore. - 11. Developing a Community-Supported Programming Language Reference Website
Vincent Fang <vinfang@vt.edu>
Dr. Godmar Back - The goal of this research is to develop a website that aids programmers in learning new languages by comparing code taken from a language with which they are already familiar to code samples from the language they wish to acquire. Programmers who already know what feature they want to code but are unsure about the syntax in another language will benefit from this website. For instance, a Java programmer who knows how to use System.out.println() could learn how to use the C++'s cout object. To enable this resource, we needed to create a framework that would allow users to not only access information, but that would allow for contributors to add and modify content as well. Sites such as WikiPedia have used this idea to great success. Not only do those sites provide a wealth of information to the users, their content can be edited directly by the contributors who believe that the webpage is incomplete or inaccurate. Similar to how an active community of contributors is helping to create a large online encyclopedia, we hope to create such a community for our programming language reference site. We chose MediaWiki, the software used to create the WikiPedia website, for our project. By itself, however, MediaWiki does not enforce the necessary structure that would facilitate our comparison-based approach, nor does it provide built-in facilities for presenting code examples. Therefore, we modified MediaWiki to support a hierarchical structure for our website. We designed an AJAX-based interface to allow contributors to easily change this hierarchy, and provided the necessary database support for the Wiki to reflect this structure. We also implemented extensions to MediaWiki to help users navigate through our website. Finally, we incorporated specific extensions such as syntax highlighting and tooltips. Our current prototype supports a basic navigational infrastructure for users, and it allows contributors to manage this navigational structure, but we have not started adding content. After adding initial seed content, we plan to open this resource to the community in the near future.
- 12. Improving HOMER with Scaling in Immersive Virtual Environments
Curtis Wilkes <cbwilkes@vt.edu>
Dr. Doug Bowman - Immersive virtual environments provides users with a realistic way to view and interact with 3D virtual environments, however, many times tasks involving interaction are difficult due to the imprecision of hardware and the inability of users to accurately position input devices in 3D space. One way to improve upon existing interaction techniques is to add scaling which is shown to improve usability by dynamically changing the sensitivity of the interaction technique. In this study we improve upon HOMER, a well-known existing manipulation technique, by adding scaling to it. This affords the user the ability to seamlessly switch between both coarse and fine-grain interaction during a given task. This union offers the user the freedom to accomplish both long and short distance manipulation tasks with higher levels of precision without compromising speed. We present results from a user study that shows that scaling significantly improves the usability of the traditional HOMER.
- 13. A Comparative Study of Clustering Algorithms for Stepping Stones and Pathways
Nicholas O. Andrews <nandrews@vt.edu>
Dr. Edward A. Fox - The Stepping Stones and Pathways (SSP) algorithm takes as input a collection of documents, and outputs a grouping of these documents where each cluster represents a topic ("stepping stone") along with the identified connections between these topics ("pathways"). Two application scenarios have been identified: dynamic collections, as in the results of Google search, in particular the results of multi-aspect queries; and static collections, as in digital libraries like the ACM, where SSP can assist in exploratory searches and in finding nonobvious connections between topics.
The goal of this work is to improve the clustering accuracy of SSP. Requisite properties of the clustering algorithm are (1) overlapping (or fuzzy) clusters, where overlapping documents can be used to find connections between topics; (2) structure discovery, so that the number of clusters need not be known a priori (supplied by the user); and (3) descriptive labels for the clusters. The current algorithm, based on suffix trees, produces such a clustering in subquadratic runtime, but its accuracy scales poorly in practice as the size of the collection or the length of the documents increases. Alternative algorithms are evaluated according to their suitability to the two applications described above, and open issues are discussed. - 14. Battery-Sensing Intrusion Protection System (B-SIPS)
Lee Clagett <lclagett@vt.edu> John Paul Dunning <jpvt40@vt.edu> Micheal Gora <gora@vt.edu> Tim Buennemeyer <timb@vt.edu> Theresa Nelson <TNelson@vt.edu>
Dr. Joe Tront, Randy Marchany - This research introduces a supporting model for a unique Battery-Sensing Intrusion Protection System (B-SIPS) for mobile computers, which alerts when power changes are detected on small wireless devices using an innovative Dynamic Threshold Calculation algorithm. B-SIPS enabled hosts are employed as sensors in a wireless network and form the basis of the intrusion detection system (IDS). This detection capability is scalable and complementary with existing commercial and open network IDSs. Irregular and attack activity is detected and reported to the intrusion detection engine for correlation with existing signatures in a database for forensic investigation by a security manager. An analytic model is employed to examine the smart battery characteristics to support the theoretical intrusion detection limits and capabilities of BSIPS. This research explores the modification of the smart battery polling rates in conjunction with the variance of malicious network activity. Using the results from a previous study of optimized static polling rates to create minimum and maximum thresholds, a dynamic polling rate algorithm was devised. This algorithm allowed the smart battery to gauge the network illicits attack density and adjust its polling rate to efficiently detect attacks, while conserving battery charge life. Lastly, a trace signature methodology is presented that characterizes unique activity for IEEE 802.15.1n(Bluetooth) attack identification.
- 15. The Gigapixel Project: Ultra-High Definition Information Visualization
Scott Conner <skottie@vt.edu>, Ronald Forbes <ron4bz@vt.edu>, David Machaj <dmachaj@vt.edu>
Michael Della Noce, Dr. Chris North - Recent advances in display hardware have allowed the resolution of monitors to increase substantially. As this trend continues, the way software is designed and used will be vastly revolutionized. Specifically, two popular kinds of applications were focused on: video games and geospatial imagery visualization. By developing prototypes of these particular types of applications, the impacts that the Gigapixel display, a fifty monitor touch screen display with a combined resolution of 16000 x 6000 pixels, had on performance and usability were observed. To accomplish the task of distributing rendering across this cluster of display nodes, the Chromium and VR Juggler packages were evaluated and compared. They both proved to work well in different situations with different rendering demands. By using these packages, both the video game and geospatial imagery visualization applications were successfully developed. Additionally, with the newfound understanding of these distributed rendering frameworks, many directions can be taken in the future to further enhance development of software on ultra-high resolution displays.
- 16. Capturing Truthiness: Mining Truth Tables in Binary Datasets
Clifford Conley Owens III <ibcliffo@cs.vt.edu>
Dr. T. M. Murali, Dr. Naren Ramakrishnan - We introduce a new data mining problem: mining truth tables in binary datasets. Given a matrix of objects and the properties they satisfy, a truth table identifies a subset of properties that exhibit maximal variability (and hence, complete independence) in occurrence patterns over the underlying objects. This problem is relevant in many domains, e.g., bioinformatics where we seek to identify and model independent components of combinatorial regulatory pathways, and in social/economic demographics where we desire to determine independent behavioral attributes of populations. Besides intrinsic interest in such patterns, we show how the problem of mining truth tables is dual to the problem of mining redescriptions, in that a set of properties involved in a truth table cannot participate in any possible redescription. This allows us to adapt our algorithm to the problem of mining redescriptions as well, by first identifying regions where redescriptions cannot happen, and then pursuing a divide and conquer strategy around these regions. Furthermore, our work suggests dual mining strategies where both classes of algorithms can be brought to bear upon either data mining task. We outline a family of levelwise approaches adapted.nto mining truth tables, algorithmic optimizations, and applications to bioinformatics and political datasets.
- 1. minimUML - A Minimalist Approach to UML diagrams
Jaime Garcia Ramirez <jgarciar@vt.edu>
Dr. Manuel A Perez-Quinones - minimUML is a minimalist UML diagramming tool that was specifically targeted to the 1st year computing education community. It is built with Java using Swing and runs on all platforms (Windows, OS X, Linux). The initial implementation left out certain user interface components which are needed for preparing more complex UML diagrams.Through my work over the last semester the tool now contains new features: a custom zoom, an options dialog box with several user preferences, and an export plug-in architecture. Both the plug-ins and the options have a flexible software architecture that can easily be modified to add more functionality. The tool also provides a base for the study of the interaction between .NET components from Microsoft, and Java code from Sun. In the ever increasing world of Tablet PC's, it has become necessary to create an interface that will allow for the easy integration of .NET components into Java programs. For minimuml, we wanted to support Tablet PC's ink when the tool runs on Windows Tablet PC Edition and just regular mouse interactions when it runs on other platforms. For this purpose a special Managed DLL can be used to bridge the gap of Ink in Java AWT components when running on a Tablet PC. This DLL is capable of storing ink from a .NET ink collector and render it to an AWT component. This support for ink is currently being added to the minimUML tool. Support for inking in Java applications is however currently limited to AWT and is not available in Swing components. This presents some limitations that we cannot overcome in our current implementation.
- 2. Large Screen Public Kiosk
Blake Sawyer <basawyer@vt.edu>, Michael Ziray <mziray@vt.edu>
Dr. Manuel A. Perez-Quinones - The ability to display and search for information in various public settings and situations for the community of Virginia Tech has until recently been restricted to a computer. If a student needs to look up a faculty member's office hours and office location, he or she needs access to a computer to do so. If one wishes to know what the weather is going to be, check out the latest campus news or events, or if they're new to Virginia Tech and want to know about the campus, again they need access to a computer. Most people do not carry a computer around with them however. Our research attempts to remedy this situation by providing a user friendly and customizable information kiosk for students, faculty and the public. The information kiosk uses large displays and a simple knob as an input device. Users of the system can navigate through a plethora of different information displays and types including text (such as a news article), a web page or web application (including webcam views of campus), video (which could give a presentation or show areas of interest on campus), visualization of an RSS feed, or even simply an interactive map of the campus. All this can be accessed with three actions from the user: clockwise turn, counterclockwise turn, and push. With these simple inputs, we can accomplish a number of advanced interactions. The system supports an unlimited number of information modules and menus that are fully customizable and dynamic. To accomplish the customization and dynamic loading of the information modules and menus, the system reads a description of the menu and the modules to display from an XML file. The modules themselves can also load information dynamically from the internet or from the local hard drive. Some information modules include such things as web pages, maps, directories, RSS feeds and video. The web pages include support for full web page display, including javascript and various plugins such as Flash. It uses Apple's WebKit, which uses the same renderer in Apple's Safari web browser.
- 3. Alternative Input Devices Using Motion Capture Systems
Matt Witherspoon <spoon@vt.edu>
Dr. Chris North - The goal of this project was to develop an alternate input device for a large high-resolution display. Large high-resolution displays can sometimes pose additional problems to interact with when using typical input devices such as a mouse. My research focused on utilizing a motion capture system from Vicon Peak as a means to achieve a new input device. Using a wand that I constructed, the Vicon Motion System tracks the wand by using reflectors on the wand. Software that I created uses this information to compute where the user is pointing on the physical screen and moves the mouse cursor accordingly. Calculations take into account bezels on the screen resulting in increased accuracy. This device gives users a lightweight input device that allows them to move freely in the environment while providing an effective interface for large-scale interaction.
- 4. Computational Prediction of Gene Functions
David Badger <dbadger@vt.edu>
Dr. T. M. Murali - Genomes of more than 200 organisms have already been sequenced. A fundamental roadblock to the successful use of these genomes in biological research is that the biological functions of 40-60% of the genes are unknown. Genes of unknown function might support important cellular functions. Discovering the functions of these genes will provide critical insights into the biology of many organisms. In addition, discovering these functions will improve our ability to annotate genomes sequenced in the future. The large scale and diversity of available data requires the development of techniques that can automatically utilise these data sets to make quantified and robust predictions of gene function that can be experimentally verified. In this project, we compare two powerful techniques for predicting gene function: Hopfield networks and Support Vector Machines (SVMs). In our framework, both methods rely on a functional linkage network whose nodes are genes and whose edges are pairs of genes that may share the same function.Hopfield networks work by systematically propagating functional labels across the edges of the network in order to minimise an appropriate quadratic programme. SVMs operate on 0-1 vectors representing the nodes (and their adjacencies) of the network. An SVM solves a different quadratic programme to find the largest margin hyperplane that separates nodes belong to different classes. We compare the two methods using cross validation and support for the predictions in the literature.
- 5. Automated Testing for a Programming Language
Aaron Tokhy <atokhy@vt.edu>
Dr. Stephen Edwards - Test cases built around the functionality of individual classes reveal bugs which may be overlooked, resulting in better code quality. Implementation of test case generation for existing languages is limited and ineffective due to the restrictions held by these languages. The goal of Sulu, an object-oriented programming language, is to successfully integrate a language with a testing framework, resulting in effective automated testing. Various black box and white box test generators will be evaluated over a set of artificially mutated versions of a class. The implementation and evaluation of automated testing may prove useful in determining if such a language could possibly reduce debug time.
- 6. LibX
Michael Doyle <doylem@vt.edu>, Robert Ellis <roellis1@vt.edu>
Dr. Godmar Back - When doing projects, with Computer Science or engineering related work, you may need to search through various websites. This can range from a simple word in a dictionary to articles from various computing journals or even to whom an IP address belongs. Our plug-in has a method, which is easy to use and access, to search though numerous collections. A large list of searches could easily make it too crowded to navigate, but through a preference system, which we designed, the problem of crowding can be solved. The algorithm used to add searches was also simplified so that future work on the plug-in will be simpler. This was all added to an existing plug-in named LibX. LibX is a plug-in for the popular open-source browser Firefox. LibX was designed using JavaScript and Xml files to assist in making searches using a school's library catalog easier. LibX can easily be adapted to work with a school's library catalog making it a versatile tool. As a way to showcase LibX a screen cast has been taken to show LibX. LibX has many uses and will see more functionality added in the future. Examples of this functionality could include more search options along with new options through the preference menu.
- 7. Gigapixel
Franck Malapelle <francky@vt.edu>
Dr. Chris North, Robert Ball - The goal of this undergraduate research was to analyse user performance in a very large high-resolution context. The 24-LCD display running on linux machines in the GigaPixel lab was used to perform the experiments. A satellite high-resolution map of Houston, TX was the subject of my experiments. The data mining of real estate websites was already done at the beginning of the semester, therefore the first part of the research was to analyse the map. Like in Google Earth the user can zoom in the map using the scroll of the mouse. The more the user zooms in, the more information related to houses is displayed (Price, number of bathrooms, number of bedrooms, square footage). The experiments focused on the viewport size of the display and the effect of curvature on user performance. The users had to perform 8 isomorphic tasks, on 8 different display size, varying from 3 to 24 screens. In most of the cases the larger the size, the better the performance. Also, a curved display increases performances compared to a flat display.
- 8. Ultra High-Definition Gaming
Andrew Sabri <plastk@vt.edu>
Dr. Chris North - Advances in technology and display hardware have allowed the resolution of monitors and video games to incrementally improve over the past three decades. However, little research has been done in preparation for the resolutions that will be available in the future if this trend continues. We developed a number of display prototypes and acheived resolutions of up to 10240x3072 to explore the different aspects of gaming on large, high-resolution displays. By running a series of HCI experiments, we were not only able to evaluate the benefits of these displays for gaming, but also identify potential user interface and hardware issues that can arise. Building on these results, various interface design solutions were developed to better notify the user of passive and critical game information as well as to overcome difficulties with mouse-based interaction on these displays. Different display form factors and user input devices are also explored in order to determine how they can further enhance the gaming experience. In many cases, the new techniques developed can be applied to single-monitor games as well as solve the same problems in other real-world high-resolution applications.
- 9. Intelligent Library Navigation using Location-aware Systems - The Newman Project
Brian Sciacchitano <bsciacch@vt.edu>, Chris Cerwinski <ccerwins@vt.edu>, Ivan Brown <ivbrown@vt.edu>
Dr. Scott McCrickard - The goal of this project was to develop an application that could be used in the Newman library at Virginia Tech to aid in the process of navigation and book retrieval. Finding resources in an indoor facility can be a difficult task, especially for patrons who are not that familiar with what the facility has to offer. The Newman Library has over 2 million items of physical media and many other forms of media such as film, microfiche, etc. over six floors. With an organization system that may not seem clear to casual visitors, these items may appear impossible to find. Current methods of indoor navigation involve the use of static maps, directions posted on walls and other traditional systems. However, as we see greater mobility of always-connected computing devices, we believe a better solution can be developed. As campuses and facilities begin to be covered with wireless internet access, we can access the web seamlessly from any location. In this project, we developed a system on a handheld device to notify people of relevant information as they navigate through the library looking for resources. Working closely with Annette Bailey, the Digital Assets Librarian at the Newman Library, we were able to ascertain what would be useful by library staff and patron alike. By displaying maps with directions and introducing several other features to aid library exploration, we believe we can reduce search and retrieval times as well as enrich the user experience. By integrating these maps with a book searching feature, we can provide the user with an efficient path drawn straight to where the book can be found. Using the Newman Project, the library patron saves time and energy and gets the most of their library experience.
- 10. VTAssist - A Collaborative Social Effort to Assist People on Wheelchairs through Notifications
Sarav Bhatia <saravb@vt.edu>, Spencer Rinehart <anubis@vt.edu>
Dr. Scott McCrickard - VTAssist - A Collaborative Social Effort to Assist People on Wheelchairs through Notifications Sarav Bhatia, Spencer Rinehart Center for HCI, Dept. of Computer Science, Virginia Tech The goal of our project was to assist people with wheelchairs with location accessibility. People using wheelchairs often face the problem of lacking accessibility information on locations they want to access. This may come from locations that they are familiar with but has undergone a change or from a location that they might be accessing for the first time. We worked with the Assistive Technologies Lab at Virginia Tech to conduct our requirement analysis and design the system. VTAssist collects information on location changes and notifies users of the updates. These notifications are delivered when the users enter a location, or when a change is made for a point of interest that the user is subscribed to. The increasing ubiquity of wireless internet access and mobile computing platforms such as handhelds and TabletPCs has led more users to spend increasing time with their devices to access or post information to social networks. In our project we render and introduce a social network that could harness such ubiquity to impact the lives of people on wheelchairs. Community users like Virginia Tech can post changes that they see to location attributes such as elevators, restrooms, entry/exits, water fountains, and other such points of interest with the help of an effective and quick feedback feature. The information change when delivered as notifications saves users time and energy, gain independence and flexibility factors that become critical when they are exposed to sudden changes to location attributes.
- 11. Improving the Flow of Information Between Citizens and Local Government
Andrew Mike <amike@vt.edu>
Dr. Andrea Kavanaugh, Dr. Manuel A. Perez-Quinones - In the last several years, the Internet has evolved to a point where it has become the primary method for information dissemination. Local governments rely on the Internet to reach its citizens, and citizens in turn have begun to disseminate their opinions and ideas using the Internet. In the last year, I have researched two Internet technologies that I believe may prove useful in improving the flow of information between local governments and their citizens: RSS and Phone-based user interfaces. RSS, or "Really Simple Syndication", is a method of website syndication that plays a vital role in several information distribution technologies, while RSS is being used effectively at the national and state level, I found that at the local level is not used as much. I present a summary of how RSS is being used in different government websites. The other technology I explored is VoiceXML, a markup language designed to facilitate development of voice-based interactive menu systems. I have developed a "comment line" for citizens to communicate their opinions to local government officials. In my system, users leave comments on a series of predefined issues. Government officials get to see on the web the number of messages left on each topic and can hear individual messages over the web. This allows citizens to provide feedback using a "commonly available technology" (the phone) and reduces the effort on the part of government officials to receive citizen's feedback.
- 12. SeeVT-ART: Opportunistic Navigation and Multimodal Interfaces
Ben Hood <bhood@vt.edu>, Scott Kelly <smkelly@vt.edu>
Dr. Scott McCrickard - The goal of this project was to design a system that would effectively enhance the experience of alumni or other guests navigating the art on display in Virginia Tech's new Skelton Conference center. The center contains an extensive collection of art pieces donated by Virginia Tech alumni, This collection was previously unapproachable as it is unlabeled and distributed all over the complex. We worked with Leigh Lally of the architecture department to create an appropriate system. We decided that the problem would best be handled by an opportunistic navigation system utilizing multimodal output. We used the location aware technology of SeeVT and implemented our system on a handheld device. Most systems implemented to handle this sort of problem would be some form of a guided tour. In order to maximize user satisfaction, we decided to use a method of opportunistic browsing instead. Users are able to explore the space at their leisure and the system relays information as it is needed. In order to accommodate the unique information requirements of opportunistic browsing, a multimodal output interface was developed. We designed an interface combining together the information systems of maps, thumbnails, and descriptive text. In the near future, audio clips would also be incorporated into the system. The result was a versatile system allowing us to display as much or as little information as needed by the users. This facilitates the opportunistic browsing through allowing us to display information necessary to identify a piece then displaying a detailed description of the item as requested by the users.
- 1. Math Handhelds in the Classroom: Visualizing the Data
Meg Dickey <mdickey@vt.edu>
Deborah Tatar - Educational software has been developed to improve students ability to visualize and comprehend difficult subject matter. However, technology developers should also be concerned with the development or detriment of community and social skills the technology might affect. Dr. Tatar previously worked at a company called SRI before coming to Virginia Tech. There she helped develop software, called Net Calc, that ran on palm pilots and helped students visualize calculus concepts. The software worked and the students made notable improvement on calculus tests. The question I am exploring is the technology's impact on the class room interaction and social dynamics. To do this I am examining the logged machine data, the video data, and developing a visualization tool that will describe their interactions.
- 2. Redesigning Requirements Analysis: Simplification and Reusability
Melissa Grant <megrant@vt.edu>
McCrickard - The goal of this project is to develop a way of performing requirements analysis tasks in an understandable and reusable way. By providing checklists to simplify decision making, "wizards" to automatically calculate complicated data, and pictures that will provide a visual representation of the system the user is currently designing, a basic approach was created for a process that is usually difficult and time-consuming. The system is strengthened by using knowledge about scenarios, Hierarchical Task Analysis, Norman's Stages of Action, and claims. Users can enter information and receive suggestions about how to best design their notification system, as well as how to effectively increase their system's usability.
- 3. LINK-UP: Empirical Testing Module
Anderson Ray Tarpley III <atarpley@vt.edu>
McCrickard - The LINK-UP system provides a web-based interface to guide the usability engineering process for a notification system. As part of the usability engineering process, designers generate a huge collection of untested interface features. In the past, designers went through elaborate measures to design a user study and process the results. LINK-UP provides a solution to this problem by automating a process that helps notification system designers empirically test interface features. Specifically, the Empirical Testing Module within the LINK-UP system helps designers create a user study and upload and process the results of that user study. Since LINK-UP covers the entire usability engineering process, these empirical results can be compared to other metrics developed by the designer for validation of the interface features. In addition, empirical test results are appended to the interface feature's record for other designers to use.
- 4. Observing Collaborative Phenomena in the Laboratory
Will Humphries <whumphri@vt.edu>
Carroll, McCrickard, Neale - To effectively understand the complex set of factors impacting computer-supported cooperative work (CSCW) systems, field methods used over extended periods of time have been necessary (Grudin, 1988). These types of studies can be contrasted with short-term laboratory research that uses a range of usability engineering methods developed to understand single-user systems. Until recently, there has been no inexpensive, short-term, lab-based option for assessing collaboration and multi-user systems. We recently conducted a study with the goal of validating a new laboratory model for observing complex CSCW phenomena. The goal of the research was to maintain ecological validity while capitalizing on the benefits of a laboratory study, namely short time span and minimal use of resources. Field studies with the goal of observing CSCW phenomena can involve months of frequent site visits. The laboratory model proposed here has a timeframe of less than a month. This laboratory model was used to observe activity awareness -- a complex problem representative of difficult problems to research in CSCW systems (Dourish and Bellotti, 1992).
- 5. NotiFly: Evolving Usability Engineering for Developing Adjustable Notification Systems
Raheel Aidrus <raidrus@vt.edu>, Justin Belcher, Ben Congleton, Doug Hall, Shahzad Hussain, Matt Jablonski, Theresa Klunk
McCrickard - This project deals with development of a notification system. This includes an original design, implementation, and testing, with the use and reuse of design and problem claims as a key part of the entire process. "Notifly", the notification system we developed as part of this research, is a real-time price monitor of flight fares from Expedia.com. The concept of an adjustable notification system (ANS), makes our system highly versatile, allowing the user to alter the levels of interruption, reaction, and comprehension based on their needs. This makes our notification system useful to different user types, with varying needs from notifications.
- 6. A Voice Interface for Remote Control of Home Appliances
Amit Nithian <anithian@vt.edu>
Perez-Quinones - Currently, interacting with appliances in a house is limited to physically being in front of the appliance (or the proxy device controller such as the thermostat). The remote control of household appliances is important for two reasons. First, having the ability to control a house remotely can serve as a safety mechanism (i.e. being able to turn off the stove while on the road). Secondly, it can serve as a timesaving mechanism (i.e. recording your favorite show while doing groceries so that you can watch when you come back). In this project, we implemented a remote control for household appliances using the telephone and voice recognition. We settled on phone-based voice user interface because of the increase in cell phone usage, and because we believe that voice provides a more natural interface to everyday appliances.
- 7. Participatory Design in the Creation of Notification Systems
Kevin Pious <kpious@vt.edu>
McCrickard - Participatory design gets the potential users of a system involved early in the design process. Developers poll the users about certain features and whether requirements are being met. I have worked at creating an interface that will allow developers to create participatory design sessions that will fit their notification system best. Users can participate in the sessions in groups or individually. After a session is complete, data is tabulated and the interface presents the developers with a plan of action.
- 8. NEMO
Nipun Jawalkar <njawalka@vt.edu>
North - NEMO (NEtwork MOnitor) is a network traffic analysis tool that monitors all open network connections on the host computer and visualizes the information in an intuitive and easy to understand manner. Using NEMO, any user can see all open ports on their computer, the processes that opened those ports, the remote IP that the port is bound to, and the status of each port. NEMO also has the ability to monitor the number of packets sent and received by each active connection, and display such statistics in the form of a stacked bar graph. Network activity (packets being sent and received) is also shown visually by highlighting the port that is responsible for it. NEMO is thus a simple utility that can be used by anyone to determine what programs on your computer are using the network, how many and what kind of ports they have open, the volume of data they are transferring, and most importantly, who it is that they are talking to.
- 9. Immersive Home Design
Jason Cowden, Craig Mackie <cmackie@vt.edu>, Ryan McMahan
Bowman - The purpose of our project was to explore how immersion in a 3D environment could ease the process of designing a home while still providing geometric accuracy. Since previous research has shown that creating a structure from scratch is difficult in a 3D environment, we divided our design process into two steps. First, we created a desktop tool. The desktop tool allows users to create a simple floor plan for a structure. One can add walls, windows, and doors. The user can also specify parameters such as the height and color of walls. This design can then be loaded into the CAVE. While in the CAVE, the user can make minor modifications to their design, such as adding and removing windows and doors, resizing windows and doors, and changing the color of walls. We are testing the merits of two different types of menu systems. One is a ring menu, which has all currently available options visible to the user, who only has to rotate the menu to the option she wishes to select. The second system uses VEWL, which simulates desktop windows in a 3D environment. Future goals for this project include perfecting the precision and structural accuracy of the user defined designs and extending the functionality of the 3D system.
- 10. CERVi: A Design Knowledge Explorer
Nima Rashidi <nrashidi@vt.edu>
McCrickard - The CERVi visualization tool is a project designed to take concepts of information visualization and apply them to a database of existing design knowledge or what we call "claims". In making the tool valuable to designers, the tool utilizes the idea of displaying Claims that are related by a certain types of established relationships. In order to do this, several categories of relationships where created. By utilizing these relationships , the tool serves as navigator through the design process by enabling designers to effectively be able to select related design claims that already exist. By providing a superior selection mechanism CERVi allows a designer to more effectively use the concept of knowledge reuse in the design of their systems.
- 11. Making System-Usability Analysis Quick and Simple: A Project Using the Reuse Library
Alan Fabian <alfabian@vt.edu>, Andrew Jackson <andrewj@vt.edu>
McCrickard - Our tools make it easier to quickly communicate a design of an interface and receive feedback on it from an HCI expert. If it is easier to communicate the core ideas of a design to an evaluator, and easier for an evaluator to give feedback to the designer, then the analysis of a design can be completed quicker and more effectively. The end result is a better interface for everyone.
- 12. CITIDEL Online Digital Library
Jed Lake <jelake@vt.edu>, Alyssa Sams
Fox, Perez - CITIDEL (Computer and Information Technology Interactive Digital Educational Library) is being built as part of the Collections Track activities of the National Science Digital Library. CITIDEL includes resources from fields including computer science, information systems, information science, software engineering, computer engineering, and many other related disciplines. Users can search or browse for articles from a wide range of professional associations and publishers. CITIDEL is a collaborative effort of Hofstra University, The College of New Jersey, The Pennsylvania State University, Villanova University, and Virginia Tech. In this project, the interface aspect was the primary focus. Previously, the CITIDEL layout was confusing and irregular. Changes included many layout style and template changes to the pages to make them more clear to users. The same template was included on all pages on CITIDEL. The free-form search was also focused modified, and the instructions were refined to make them easier for users to understand. User testing is planned to assess the effects of the changes made.
- 13. Managing Courses with CourseSite: The Professor's Perspective
Travis Cherry
Perez - As the academic world becomes increasingly reliant upon technology, it is essential that course management systems adapt accordingly. Both students and professors alike find that these new systems greatly enhance and simplify their class experience, as they have immediate access to course calendars, assignments, announcements, grades, and much more. There are several great template-based systems in existence such as BlackboardTM; however, the emphasis on generalization in designing such systems tends to limit features and likewise retract from user-friendliness, thereby rendering them less inviting and advantageous. Previously, development of the Cross-Platform CourseSite focused on providing users access to the same data over multiple platforms (web, phone, & pda). Once this aspect of the site was completed, the emphasis has now shifted to the professor?s point of view. We hope to provide the professor with a convenient, customizable, and exceptionally powerful system that offers a truly simple medium through which they can manage their class. Customization and simplicity are perhaps the two universal themes of the CourseSite, and we have determined that customizing does not merely consist of selecting a unique style of button, rather it is giving the professor the freedom of defining their own interests and priorities when accessing certain portions of the site. Simplicity, on the other hand, stresses the overall organization and accessibility to key management options. After having been tested and utilized by Dr. Perez and his Introduction to Object Oriented Design II students over the past semester, the CouseSite has evolved incredibly. We have discovered several ways to enhance the course management experience, including customizable voice-access portals, automatically generated grade-upload spreadsheets, and several other significant features.
- 14. It Came From HCI
David Felton
McCrickard - This project examines the use of claims as a reuse mechanism for human computer interaction design, specifically notification systems design. The work considers how claims can be stored and browsed in an online library. My particular focus is on ways to distinguish good claims from bad ones when searching the library.