Our Selected Idea

During brainstorming, The End Users had many ideas for improving Student Central. It seems that SC offered a significant number of opportunities for change! Most of the ideas we had involved bringing SC out of the stone-age of the web and building into it many of the same customizable features found on common e-commerce sites.

One idea, however, stood out above the rest--that of the recommendation agent to help students during registration. We felt that this agent was the key to making SC truly valuable to students. The primary purpose of this agent would be to take the grudge-work out of registration--to construct, based on many factors including a student's standing, year in school, major, academic record, advisor's recommendations, and personal preferences, one or more lists of courses for the student to take, and to offer these recommended schedules in discrete alternatives which the student may accept, reject, modify, and prioritize. While we are advocating the agent only perform a limited role in this redesign, it will almost certainly be extended beyond simple registration functions in later assignments.

In reality this agent will be nothing more than a set of preferences stored in a database, some computer scripts, and few clever algorithms, however we felt that it was important for students to be able to put a "face" on this agent. After a bit of deliberation, we settled on the name Tarheel Automated Registration Agent, or TARA, for our agent.

Next, we decided that TARA should be able to act on behalf of the student if he is unable to register at her appointed time. If a student cannot log into SC to register when his registration window opens, he should not be left with only the classes that no one else wanted. Similarly, if a student simply doesn't want to face the headache and frustration that currently accompanies attempts to use a system that thousands of other students are using at the same time, he shouldn't be forced to. We decided that TARA should be able to perform these registrations for students on their behalf. But we also decided that the only fair way to do this is for TARA to act only if the students did NOT log into SC and attempt to register manually on the morning of registration. Furthermore, all students trusting TARA to register them would be registered using a weighted random order. We felt that the weightedness should give a slight advantage to students with higher GPAs and to students who were near the end of the queue in previous semesters, however these weights could be adjusted over time by the system administrators in response to students' and university administrators' desires.

Lastly, we found the idea that SC didn't offer shopping cart functionality, similar to modern e-commerce sites, abhorrent. Therefore, the next part of our idea is to implement multiple "course-carts" within SC. Under normal circumstances, several alternative course-carts would be constructed by TARA immediately after a student speaks with his advisor. The system would evaluate all the possible combinations and permutations of all of the sections of the courses recommended by the advisor and place each permutation into a separate course-cart. For some students this would generate multiple alternative course-carts; for others, only one or two possibilities. Additionally, whenever possible, it would also take the student's personal preferences into account regarding course times, sizes, instructors, or any other factors we might decide to implement. A single course-cart could also contain alternate courses in case a desired section was full at registration time. These course-carts could be modified, created entirely from scratch, and prioritized, by the student prior to registration. The ordering of course-carts (prioritization) by the student is important. In the event that TARA cannot register the student for one of the primary or alternate courses in one course-cart, the system would attempt to register the student for all of the primary or alternate courses in the next highest preferred course-cart, and so on until registration is successful. Of course, as a fall-back, if none of the carts may be achieved in toto, the system will make the most-complete, highest-priority schedule possible.

Scenario: Registration Period

Chatham:
Chatham logs into SC Wednesday at 3am, 3 days before senior registration period opens. He talked to his advisor only yesterday (weeks after he was scheduled to) and his advisor recommended to him that he take 5 courses. Chatham doesn't remember exactly which courses. Nevertheless, when he logs into SC, he finds his advisors' recommendations in SC. In fact, TARA, the new "agent" within SC has already constructed three course-carts for Chatham, two of which have all 5 classes and adhere to his personal preferences--no classes before 10am--and one which has all 5 classes including two 9am classes. He quickly disables the third course-cart and assigns the others equal preference before quickly turning off his PC without a second thought.

On Saturday at 10am, Chatham's registration window opens. He, however, is still asleep, having been up until 4am the previous night partying. Nevertheless, because he does not log into SC interactively at 10am, TARA places him in the registration queue using its weighted random method and he is successfully registered for classes. He is enrolled in 4 of 5 of his first-choice classes and 1 alternate, all in course-cart beta. He doesn't think to even check SC for his actual schedule until the Thursday after registration. After all, he has dates to pickup and parties to attend!

Paula:
Two weeks before freshman registration opened, Paula meets with her advisor for 35 minutes and discusses which classes she should take. She and her advisor work out a schedule including several options from which Paula must choose. Paula doesn't quite trust computers, despite being technically proficient, so she records her advisor's recommendations on paper even though her advisor tells her that TARA in SC will remember them.

It is two days later before Paula has a free moment to log into SC. True to her advisor's word, SC remembers the recommended classes. It has even pre-constructed four course-carts including combinations of all of the courses for Paula's choice. There are more possible combinations should she want to construct a course-cart manually, but SC has recommended these because of Paula's preferences: She prefers morning classes and prefers classes taught by women and native-english speakers whenever there are choices. Paula clicks on each course-cart to examine the full details of each course and also views the weekly schedule for each course-cart--she prefers to keep her afternoons flexible for study or work. She orders the four carts by preference and logs off.

At 9:30am on the Saturday of freshman registration, Paula wakes and is reviewing her choices in SC. She is pleased with them and plans to take no chances with SC's weighted-random lottery: she plans to interactively register at 10am. Unfortunately, her phone rings and her boss at Elmo's has directed her to report to work ASAP. She will have to trust SC to register her; she logs off and reports to work. Late Sat night, after a long day at work, she checks SC and is relieved to see that it actually worked! She was enrolled in all the classes in her primary course-cart. She signs off relieved and begins her biology homework.

Storyboards

It's time to register for classes. See how Chatham and Paula get the classes they want while living their lives.

Write-up

In thinking about redesigning Student Central in terms of programmability, we had lots of ideas on how to improve the various bits and pieces of the artifact. These ideas were based on, or inspired by our earlier observations, surveys and interviews of current and potential users of Student Central. But early on, we decided that we did not want to merely improve the way things were, nor did we want to go to the other extreme and implement some razzle-dazzle application that would be too complicated for students to use. What we wanted was an application that 1) would actually work, 2) would require the least amount of time and resources from students to use, and most importantly, 3) served students by helping provide them with an enriching and enabling experience, both in and out of the classroom. We wanted an application that “worked” for the student in the sense that it not only functioned properly, but worked diligently, day and night, to serve each student’s own personal needs as a loyal and willing proxy. And we wanted ALL students to benefit from this programmability function.

     We developed our personas, Paula and Chatham, to represent not only the wide range of social, economic and cultural backgrounds of the diverse student body, but also gave them vastly different personalities, focus, drive and academic prowess, each with their own distinct goals of what they wanted out of college. We wanted to design a degree of programmability such that the new system would meet the vastly different needs of Paula and Chatham, but in doing so, make them both better students by increasing their chances of success. Of all our brainstorming ideas, we decided that a new class registration system comprising a “smart” agent combined with shopping-like course-carts would yield the greatest impact.

     Our current registration system was a major complaint among those we surveyed and interviewed, not only in terms of performance, but functionality as well. Searching, choosing and registering for classes was simply a crap-shoot at best, Russian roulette at worst. We wanted to re-do the way students register for classes by allowing a much greater degree of programmability, not only in terms of how the system responds to and accommodates each user’s preferences, but also in unshackling them from the process, freeing them to live their lives, to study, work, sleep, party, socialize, or just meet the demands of their lives. Meet TARA, the Tarheel Automated Registration Agent.

     TARA works for the students. TARA never sleeps. TARA never complains. TARA doesn’t say no, TARA finds a way to say yes. TARA gives students not only what they need, but often what they want. TARA gives students a piece of mind and lets them live their lives, and in the process, become better students. And TARA does this by allowing each student, often working with his/her advisor, to input a whole range of preferences and parameters, which in turn, determine the various choices available to the student, not just in terms of an academic path, but also in scheduling, instructors, and locations. And during registration periods TARA registers each student for a “pack” of classes chosen previously by the student and saved in baskets or “Course carts.”

     Here’s how the new registration system works. Every student has a number of initial parameters set by default in Student Central depending on the student’s class standing, major, if any, and previous work completed. So, for example, when a sophomore Math major searches for a class that will meet some major or perspective requirement, Student Central only displays classes that are applicable and appropriate to the student’s status. Even before any personal customization, the registration system recognizes and accommodates each student by providing them with appropriate choices and search results. Each student may then customize their preferences setting a whole range of parameters from scheduling preferences (hard and soft preferences), to location, instructor, and special needs (handicapped access, impairments, etc.) preferences. These additional parameters allow each student to further customize, or program, the functionality of the registration system based on their needs.

     The actual registration for classes is performed via TARA. Before registration period begins, the student searches for and selects classes, perhaps working with her advisor, then drops them into a “course-cart,” which may contain the max number of courses that a student can attempt. Given that some classes are more popular than others with limited seats, and classes inevitably fill up, a student may have multiple course-carts of different class combinations, depending on various scenarios and contingency plans. Cart A is first choice, B second choice, C third, etc., up to five carts. The cart or carts represent the student’s course selections for an upcoming semester. The actual registration can be done in one of two ways—automatic and interactive.

     In automatic mode, TARA takes care of registration without any need for the student to be present, in person, online, or on the phone. Students who select this option delegate their carts to TARA, which in turn places the students’ carts in a registration queue using an algorithm that gives some weight to class standing and grade, but also has a degree of randomness—a weighted random queue. Students are registered for their chosen classes based on availability, starting with their first carts, then second, etc. and their place in the queue. TARA notifies the student via the student’s chosen notification method once the registration process is complete.

     In interactive mode, TARA assists the student in assembling her course-carts, and then during the registration period, the student logs on to “check out.” This option is for those who want to avoid the weighted-random queuing system and want full control of their registration. Interactive registrants may get slightly better placement in the queues, and are able to change their minds until the last minute, swap carts, etc. while those choosing automagic registration relinquish control. TARA has a number of safety mechanisms to ensure that everyone gets successfully registered for classes. For example, if a student has not elected the automatic method and does not queue up interactively by a certain time during the registration period, TARA places their carts in the automatic queue.

     This degree of programmability accomplishes several things for the student population. First, the automated “set it and forget it” registration system frees the students from the drudgery of the present system and allows them to better use that time to study, rest, work, socialize, or do whatever it is they want to do. It gives them control over their time. As we can see in the case of our personas, both Paula and Chatham value their time, and both can better spend their time than to be stuck staring at time-out errors while trying to register for classes that are already full.

     But most importantly, the degree of customization gives students much greater control over their academic careers by presenting them with choices appropriate to their goals, tastes, ambitions and other stated preferences. For example, had this system been in effect, Chatham probably would have graduated by now because his academic path and course choices would have been much clearer. Instead, he ended up taking a bunch of courses that were not applicable to his various majors. Even without personal customization, TARA can give each student a full map to follow, with primary and alternate routes.

     With customization, TARA can not only provide a full road map, but also provide interesting but valuable detours along the way, based on stated preferences and other factors. For example, TARA can easily and effortlessly provide Paula, a busy biology major, with courses outside of her major which she would find both interesting and valuable. So, instead of just randomly and unknowingly picking courses because they happen to fill a slot or meet some perspective, TARA can provide Paula with choices that would not only fit into her schedule and meet her perspectives, but would interest her as well. TARA simply helps students be better students by helping them set a course they can follow, and by letting them spend more time being students instead of being “class-registerers.”

     TARA will greatly improve the class registration process and may even contribute to the success of students by helping them make appropriate and timely decisions concerning their academic careers. But it also has some potential limitations. For example, some may view the registration process as a rite of passage, a natural part of college. Some may view the torturous hours spent trying to register a time-honored tradition. We disagree. But for those who want to experience “old school” registration, certain features could be built in to TARA that would allow students who check the “old school” box to experience the pangs of this ancient rite of passage by arbitrarily dropping their sessions, losing their carts, etc. :-) We think most students will opt out of this feature, and wisely so.

     Another drawback may be that by making an academic track so customized, based on personal preferences, that serendipity is reduced, and the chances of the student taking a class outside of his/her comfort zone may detract from rather than add to the academic experience. This is a legitimate concern but the agent can certainly recommend random courses that the student is qualified to take, and the advisor will continue to play a key role in working with the student in choosing a path.

     The bottom line is that TARA can not give all students what they want, because there are a limited number of seats and classes, but it can certainly help give students what they need, in terms of helping them make accurate and timely choices concerning their academic careers, and free them from the drudgery of the current registration process.