A Honeynet is an information security system that the Honeynet Project research organization developed to collect network exploit information. The system consists of two computers, one dummy production machine called the Honeypot designed to be attacked, and one logging machine called the Honeywall that collects all network packets entering or leaving the Honeypot. The data captured allows security professionals to learn about how attacks are carried out and what they look like at the lowest level. Unfortunately, data analysis on the current Honeywall system is very tedious and time consuming. The tools available are difficult to set up and use. The user interface is counter-intuitive and the data is hard to understand. To answer these problems, we wanted to create a centralized system that enables users to collect and correlate data from distributed Honeynets and then analyze the data in a powerful yet intuitive Honeynet data analysis interface. This interface should be simple enough for inexperienced users, but powerful enough to support detailed analysis. To gather the requirements for our system, we did detailed interviews and ethnography with three Honeynet users of very different skill and experience levels. Using this data, we created a persona of target user and several wire frames which would be used to justify our design choices. Our final system is a web-based interface using a combination of state of the industry open-source web technologies to tie into a database that stores Honeynet data. After creating an initial system, we presented it to Honeynet users to gather their feedback. This gave us many insights into where our system may need to improve, but also gave us hope that we are on the right track. We hope to incorporate our user feedback into the next iteration of our system.

As part of the MSN/Microsoft's Real People Real Data Program, this project seeks to find, define, and design a personal communication device to enable people to connect with their close friends and family. The mission of Real People Real Data is bringing the voice of the customer to Microsoft and beyond. To accomplish this we use participatory design as one of our methods, which involves real people who are real users in the design process. The sessions are worldwide, and started with Seattle, Washington and Madrid, Spain. From the findings we identify characteristics, features, and desires, which our technology and products must be able to afford.

Student organizations at the University of Washington want to minimize the overhead of administrative tasks such as announcing events, sharing files, and contacting members. Web applications have the incredible potential to facilitate these tasks; they are free, customizable, and compatible with the web accounts given to them by the University of Washington. However most student organizations do not possess the technical experience or time to choose, install, and maintain these applications. To measure the desire of student organizations to utilize web applications, the entire population of around 350 registered student organizations was sent a survey questioning their current method of announcing events, intra-club communication, and file sharing as well as a list of web applications that they currently or desired to use. In-depth interviews were also conducted using availability sampling. The analysis of the nineteen returned surveys and five in-depth interviews yielded better insight into how organizations operated as well as a list of desired web applications; the top four being an online calendar, file sharing, membership management, and a tool to synchronize their several email accounts. A website, named UW Student Grouper, was created that implemented the top four requested web applications. The membership management and email synchronization tool were developed in-house using PHP. The online calendar and file sharing utility were integrated into the website from existing open source PHP projects. UW Student Grouper sits on top of a MySQL database. After the first release of UW Student Grouper, student organizations who previously expressed interest in evaluating the suite of web tools were asked to evaluate the website. One pre-evaluation was done to remove questions in the evaluation that were not understandable before the evaluation was emailed. Two evaluations were received. UW Student Grouper was then modified to reflect the feasible changes expressed in those two evaluations.

The purpose of this project was to create an online language supplement to assist students enrolled in the University of Washington's Technical Japanese Program. Not only would the system allow students to review and practice class concepts, but it would also facilitate the administration of class assignments and the communication between students. Challenges included the correct display of Japanese characters and the tracking of student and assignment information in the backend database. The framework used for the project was created with a combination of technology, specifically Microsoft SQL Server and ASP.NET. One goal of this capstone project was to enable TiP students to extend their opportunity for learning outside the classroom by way of an interactive web experience. Through user assessments that occurred over three different interventions, we built up a basic, first generation prototype that implemented kanji and vocabulary review/quizzes, a community area where students and teachers could share their opinions on various topics, as well as an administrative module to handle content management of the system. This project provided us with great incite for improvement, but because of the size of the scope, we were only able to implement this first generation prototype, not something that could be considered production ready. However, the vast amount of knowledge gained through this experience will aid in our future endeavor toward the completion of this project.

The Interactive Network Incident Response Toolkit (iNIRT) is a tool designed to facilitate the collection and submission of information regarding the security of a particular host within a computer network. The Toolkit is designed so that a network administrator with little security experience can use powerful network analysis tools to collect detailed information about a host and send the information to security professionals for analysis. The original design of the Toolkit, proposed by security researcher David Dittrich of the University of Washington's Information School and Computing & Communications department, called for a bootable CD-ROM that could be distributed by security professionals and used by network administrators, but this project focused on the design of the interface and the interoperation with networking tools rather than the implementation of the Toolkit as part of a bootable CD-ROM. Through a multistage process, the Toolkit was designed to meet the needs of both of these populations. A needs assessment consisting of interviews and a survey assisted in defining the implementation parameters for the Toolkit, which were considered throughout the design & development phase. This phase consisted of prototyping in multiple media and the eventual implementation of a fully functional prototype using Perl/Tk in a Linux environment. Finally, an impact assessment was conducted to evaluate the effectiveness of the interface, and feedback gathered through the impact assessment was applied to the final prototype. The result of the project is a functional prototype for a Toolkit that has the potential to greatly increase the availability of critical network security information to those who can use it to identify problems and increase the availability of information technology resources to everyone within their organization.

Choosing the right campus religious group is a very significant decision for many students. However, the means by which students are currently able to investigate the forty some campus groups are exceedingly poor. Investigation happens through visitation, on a group by group basis, and is usually spurred by advertising. There is no way to browse or search through all of the groups looking for specific types or qualities. The solution presented by this project is a Fellowship Finder website that enables students to browse and search the campus religious groups according to categories and attributes that are relevant to them. In order to develop a design for this website I used a three-phase design model. I conducted staff and student interviews and built an affinity diagram in order to discern the attributes by which students would need to browse and search. I used the categorization scheme that I developed to design the site's database, and then developed an interface by which to browse, search and display the groups. Upon completion of the database and interface, I submitted the system to expert and user evaluation to determine what the next steps of re-designing should be. Over the course of the project, I discovered the most important attributes involved in searching for and representing campus religious groups; from both a group and a student standpoint. I learned about the differences in the nature of groups across religious lines, as well as in the needs of their student members. And finally, I was able to develop the initial design of site that would potentially fill a significant informational niche for many students on campus.

This project used a multi-cycle design model to create a prototype of a graphical user interface (GUI). The objective of this GUI is to permit novice XML users to create XML content to published grammars without knowing any code. This project also examined the relationship between document structure and document content and searched for ways to make the structure of the document more apparent and understandable to the user. Three interventions were used. The first examined three interface styles to determine which was best suited for the GUI. Included were a split screen interface, a multiple single document interface (MSDI), and a single document interface (SDI). Although the test subjects preferred the split screen, other considerations caused the adoption of a hybrid MSDI/split screen solution. The second intervention explored methods for representing document structure, including trees, outlines, and a graphical representation using nested and/or linked geometric shapes. As a result, nested boxes were chosen to represent document structure in the final GUI. The third intervention produced a high fidelity prototype of the GUI based on the results of the first two interventions, including both a structural view of the document and a data entry form, and tested its impact. Based on the reactions of the test subjects, the GUI was a success; however, the subjects hardly used the structural view. In addition, several problems were revealed, including the difficulty of representing large structures geometrically in a small space. Suggestions for improvement included using a linear fish-eye to expand the relevant part of the document, scrolling the document structure view to show the most recent changes, and using a tabbed interface.

Information Problem: The lack of 'real-time' quantitative feedback of individual performance has plagued participants of skateboarding (and other extreme sports/action sports) since its inception. The inability to track statistics and other key data points continue to keep many of these sports out of highly respected institutions like the Olympics, simply because measurement and comparison by the general audience is not possible. For those action sports privileged enough to be a part of prestigious competitions, they too are subject to the precision of a judge's naked eye, versus hard data. Simply put, an information problem exists: a lack of accurate, objective, real-time information for judging and evaluating the performance of extreme athletes. Our team intends to design and implement a system to solve this information problem. This system will consist of a skateboard embedded with the preferred sensor technology and a software package to compute and display the statistics in a thoughtful and meaningful way. Both a Human-Centered dimension and a Technical dimension will be equally represented in the design, an aspect which is crucial to its success. Our project will embed sensors into a skateboard. Using the instrumented skateboard, telemetry will be collected for a few basic skateboarding maneuvers. This telemetry will be used to facilitate the development of the presentation software for use in judging situations.

For this project, our team explored the music-sharing phenomenon that has captured the interest of young adults these days. We questioned how college-aged students can legally share music, learn about new music, and purchase music in a single multimedia application. We decided to create a prototype of such an application using Macromedia Flash, and integrated an instant messaging program as a way of encouraging the social aspect of discovering new music and the interaction amongst friends. It is called the Instant Messenger Music Pool (IMMP). We first took a look at the current working technologies used for sharing music, such as mixed CDs and peer-to-peer networks. We chose to base IMMP upon Microsoft's threedegrees beta program. We then performed a user needs assessment with interviews, and derived our design ideas from the users' feedback. A paper prototype was created using Macromedia Freehand, and the Flash prototype was modeled after that accordingly. Our final step was an impact assessment, where we took IMMP to the original users of the needs assessment, and asked them to evaluate it. From this project, we learned a lot about project development and management, as well as about what type of music-sharing application would appeal most to college-aged users.

This project forms the entirety of my Informatics Capstone Project. It began in December of 2003 and evolved considerably over the course of two months. What began as an attempt to synthesize existing Personal Information Management software features into a single product quickly evolved and shrunk with regard to its scope. My current approach to personal information management does not involve a stand-alone application, exclusive from all others such as Microsoft One Note, but instead involves improving the current auto-complete technology within Microsoft Internet Explorer. The problem this project works to assess and address: Knowledge workers are often faced with a large volume of electronic resources such as documents, email, web addresses, and the like. As this volume increases, the worker's ability to keep track of essential details suffers. Thus, they are less able to effectively complete their tasks within teams or individually.

Many voters take seriously their responsibility to cast an informed vote. However, becoming informed about local political issues and candidates often presents a challenge, despite the flood of information media. Wouldn't it be nice to have a Web site that is searchable, browsable, offered discussion forums for voters to express themselves and could recommend items that the users deemed worthy? The content might include opinion articles, editorials, news articles, cartoons, and the like from local publishers. Such a system might handle local political information in three ways: (1) search for and browse information regarding local issues and candidates, (2) participate in discussion forums, and (3) assign ratings to articles as a way to collaboratively filter information that is of particular value to the people doing the ratings. My implementation focuses on the third part: the system recommends articles to users based on their own prior ratings and other people's ratings of the items being recommended. In assessing user needs, I interviewed voters of different political leanings about how they become informed about local issues and candidates, and how they recommend political information to others; I also analyzed how they rated two sample articles. Based on the needs assessment, I collected recent articles published in local newspapers and asked my users to rate some of them. I used an open-source recommender tool that holds ratings in a database and calculate recommendations based on the ratings. To evaluate the user interface, I created Web pages containing the actual recommendations, and observed my users interacting with the Web pages. Commercially, recommender systems are often used to help people make buying decisions. Deciding which candidate or issue will receive one's vote can be equally wrenching. My project sheds some light on how people might rate and collaboratively recommend political information in a local context.