Historically, the capstone courses in the CS/CE undergraduate curriculum, together with our partners in team-based honors programs on the College Park campus, spawned a number of successful software projects under the SEAM banner. For information on legacy projects, and a bit how we've come to the point where our leaders won't act to allow us to offer these again, scan down to later in this page.
Select and recent student project experiences in software engineering are reflected in the short descriptions immediately below.
Sorry - many links are temporarily removed as we update our servers.
PrivateMail is a Chrome mail plugin that helps you use - and promote - private email practices by handling foryou much of the overhead of installing encryption tools.
MotorDrive revamped the software used by the Neuromotor Control and Learning Laboratory (NMCL) for their research trials. We created a robust API with far broader reconfigurability than prior tools used in that lab, enabling more experiments in neuromodeling to be conducted more reliably and at lower cost.
We developed a desktop application to support experimental research in the application of artificial intelligence techniques to real-world problems like facilities management. Imitation is an abstract technique in which a machine learns some task by observing how a human expert performs the task first. Prof. Reggia and Garrett Katz at the University of Maryland successfully demonstrated this approach in simple block stacking tasks, so to see how well this can be applied in other domains, we created the tools to allow the Reggia-Katz system to be quickly extended to other knowledge bases without needing to write new programs. This facilitiates experimental study of the ideas. By walking through the step by step knowledge builder, users will be able to generate all files needed to conduct their experiments. The apparatus is now being used by Prof. Jeffrey Herrmann to understand how well the approach works in real-world applications such as facilities planning and layout.
Discover connections between your tables in a database, even when your data have different origins and arrived at different times.
EasyAsPie is an application geared towards medium-level website developers who desire to provide their customers with a better data visualization experience. It is visible to most users as a Firefox plug-in that allows them to navigate through large organized data sets with minimal mouse actions using pie menus. And on the server side, it allows for seamless importing of databases for flexible visualization.
Information you want (news, data collected from your Internet-aware home and more), presented the way you want to hear it and whenever the time is right.
Apparatus built to support research at the University of Maryland, the Geotags web app intelligently combines related topics of discussion across multiple social media sources to provide a navigable map of trends across the United States.
Builders team offers system definition environment which allows users to express rules in a language for definition of expert systems, and then generate run time systems and styling for those tools to be broadly used by others
Bring the Internet-of-things to your home without sacrificing your privacy to the cloud!
OpenAR is an open source, free to use annotated reality library. Our library brings industry standard image matching tools to the masses through our RESTful API. Since our tools are run server side, OpenAR supports almost any platform that can send and receive web requests.
UAV Route Planning System
We developed a web-friendly system for UAV operators to make logical and well-educated decisions when planning the prospective routes of their planes. These tools allow them to determine what acceptable tradeoffs between risk and flight time. MORE
Bandwidth Listening Tool
How can you tell whether you are making the most efficient use of your available bandwidth when connected to the internet? Often on wired connections, and certainly on most wireless connections, consumers pay for bandwidth, yet have little control over how much of that bandwidth is utilized. This is especially true in the case of browsers. Depending on the user’s viewing habits, much of the bandwidth consumed is content not chosen by the user. Video push in ads, background loads, and more all whittle away at the already limited bandwidth, and often without value to the user. Some tools allow blocking of various kinds, but that presumes that the user knows where the costs go in the first place. It is not possible to optimize cost and viewing experiences if there is no way to know the starting point on bandwidth usage.
BLT (bandwidth listening tool) is an extension to Mozilla Firefox browser that tracks and categorizes a user's bandwidth utilization during a browsing session. It works much like ad blockers, except that instead of barring the loading of that content it simply tracks it so the user can see overall how much resource goes to the downloading of content that was not explicitly sought by the user. The tool helps you recognize which web sites are 'lean' and which push a lot of content you never wanted.
BALOO (Breakthroughs Attributed to Large-scale Online Outsourcing) is a SEAM effort to study properties of crowd-sourced systems. While the research questions have to do with measurement of the quality of strategies suggested by an engaged population of users under one or another set of conventions for interaction, the system itself appears as a massive-multiplayer online game in which players propose solutions to real-world problems. By working together, players can determine the best course of action for the researchers who run the game. By playing cards and earning points, players can collaborate and compete at the same time. Remembering that in the Jungle Book Baloo is Mowgli's best friend, some of Baloo's properties were inspired by MMOWGLI, a successful project at NPS, and the first implementation of Baloo was done by the capstone software engineering class here in the fall of 2012.
SEAM's origin lies in a novel approach to software engineering education. In January of 2003 it formalized a practice with which we had been experimenting in the classroom, which was to engage students in 'live' projects as part of their capstone experience. Carefully vetted problems brought to the SEAM Software Cooperative by our industrial partners were solved by the students in the culmination of an undergraduate track which concentrated on industrial practics. This integrated content from the upper-level courses in the CS major, and provided very practical skill sets for students to bring to the workplace after graduation.
An essential ingredient in SEAM's secret sauce was the mentoring given to these students along the way. Graduate students with serious industry experience prior to returning for a PhD in this department were brought on board to augment the base of teaching assistants supporting the class, and the more-advanced courses in product assurance or management further expanded the experience base as the practicum in those classes was to serve as first line managers in the capstone course offered that semester. This allowed SEAM to undertake projects which spanned more than one semester, and gave us the capacity to interact with clients on-site. Most important of all, the peer mentoring between groups of students was spectacularly popular and effective.
Over time, the peer engagement model within SEAM was extended in order to engage students in the capstone class (our producers) with students in honors program research projects having substantial technical needs (our consumers.)
The business model which made SEAM so successful has not been allowed by the Computer Science Department for some time, so what we list below are a few of the legacy projects which gained notoriety or saw broad application. Some have withstood the test of time and are still in use.
Today's standard pedestrian map of campus, TerpNav, has its origin in the research project of a Gemstone team which was studying what properties of an information system would tend to make it sustainable. Gemstone Team FASTR, Finding Alternate Specialized Travel Routes, quickly identified maps, especially those suited to the needs of people with a mobility handicap, as a suitable driving problem for their study. The first versions of TerpNav were created in collaboration with students in the capstone software engineering class (the SEAM laboratory of the era.) Team FASTR graduated in 2009. Graduating in 2012, Gemstone Team FLIP, for File Lending in Proximity, used TerpNav as the basis for its research into novel technologies for social networking, which resulted in TerpNav's "flip layers" now in daily use by members of the College Park community. Further research into properties of crowd-sourced, geo-aware social networking systems continues today.
All undergraduate students at the University of Maryland must have on file a valid four-year plan which has been approved by academic advisors. Unfortunately, when handled solely on paper, the management of these plans can become an administrative nightmare for a department. Student success can suffer when plans are allowed to become out of date, since departments have less ability to plan allocation of instructional resources and students draw less attention to advisors when their plans might need to change. In Fall of 2009, the SEAM lab developed MyPlan, a resource for students to interactively develop graduation plans, with immediate feedback provided on course pre-reqs and suggestions given on available options, with all details at an advisor's fingertips during a personal advising session. The college adopted MyPlan and used it with success for several years, until the CS Department elected to no longer host the service. Most students are thus back to paper plans or Word documents until the campus can finally deploy its long-awaited Kuali system.
SEAM teams participated in a broader project called Advanced Health and Disaster Aid Network in collaboration with Johns Hopkins Applied Physics Lab, by prototyping a hand-held digital assistant for nurses to monitor their patients' vital signs from anywhere within a hospital setting, and to get immediate alerts to signs of a possible issue. (2006) Teams also created a version of this prototype for use by first responders in disaster response systems, in order to ensure effective communication and collaboration. (2007)
In collaboration with Eureka Software, a medical software development company based in College Park, SEAM developed a graphical image analysis and search capability for the patient records managment product called Archimedes. This allows doctors to find medical images such as X-rays, CAT scans and MRIs based on either a desired image's properties or by its similarity to some reference image. (2005)
SEAM created an information management tool for Maryland Task Force One (MDTF1) to help them deploy Montgomery County's urban search and rescue fast-response teams. In order to activate as part of FEMA's National Emergency Deployment System, MDTF1 has a great deal of paperwork to process and very little time to do it. They must know team members' availability, their capabilities (since missions may vary widely) and much more in order to organizeeffectively. They made the transition from using a whiteboard to software with the help of SEAM. (2004-2005)
Later, SEAM implemented an inventory tracking and control system for MDTF1, to assist them in tracking gear. While inventory control is normally thought of as an ordinary problem, the issues become magnified when operations might be performed under the pressure of a deployment in an emergency, and are compounded further by stiff FEMA regulations concerning how federally funded equipment must be tracked. SEAM's mobile scanners and database tools did the trick. (2006-2007)
|Directives Management System |
In Spring 2004 the Montgomery Count Police Department asked SEAM to develop a solution to the problem of organizing and keeping track of signoffs on police directives. The directives are rules on how an officer should handle different situations that arise. At the time, a hardcopy had to be printed for each directive for every officer creating a large paper cost. Every officer then had a bulky binder which held these directives. It was very difficult for the higher ranking officers to track who had signed off on a directive and who still needed to receive it. SEAM provided an efficient way to electronically transmit directives to all of the officers as well as keep track of when the officers received them. (2004) Previously, SEAM members assisted MCPD in extension of the county's software for tracking Use of Force reports which were required by the courts. (2003)