Development Process



Low-Fidelity Prototype


The first time a user starts the Digital Personal Trainer, he or she is required to create a profile (see Figure 1). The profile is used by other parts of the system to tailor workouts to the user’s body type (which is determined by age, weight, and sex). After creating a profile, the main menu (see Figure 2) is screen presented to users when they start the program.



Figure 1: Profile Creation Screen                                            Figure 2: Main Menu



The new phase screen (see Figure 3) allows users to create a workout phase by selecting a goal for the phase (top combo box), selecting the duration of the phase (second box from top), and select zero or more ‘trouble’ spots to be focused on during the phase (second box from bottom). Once all these selections are made, the user clicks the button labeled ‘Create a Phase’ to record the phase in the system. A user who has just created a phase or chooses ‘Continue Phase’ from the main menu (see Figure 4) is presented with the above screen. It lets the user know what muscle groups are scheduled for exercise that day, and lets the user choose the workout type.




                                            Figure 3: Create New Phase                                             Figure 4: Create  Phase Workout



Workouts are displayed by showing the user one exercise at a time. If the equipment for the exercise is available in the gym and the user is ready to perform the exercise then he or she can begin the workout out right away. After the completion of each set, the user should click the button labeled ‘Record Set’ to record the information about the set performed. After clicking ‘Record Set’ the user gets this screen (see Figure 5):



Figure 5:  Record Set Information


The last option under the main menu is to generate reports. These reports give the user information about their workout habits, past performance records, and expected physical capability. A user can view three reports: RM Calculations (Figure 7), Phase Progress, and Performance Highlights.



Figure 6: Reporting Screen                                              Figure 7: Rep Max Calculations



Development of High-Fidelity Prototype



Design Problems and Difficulties of the Primary Design


            The first design is by no means perfect. The key problem with PDA designs involves displaying enormous amounts of information on a very small PDA screen. Weightlifting is a complicated task with a complicated recording and record-viewing process. Compressing information that usually requires several sheets of paper or a desktop system onto a PDA interface is extremely difficult.


One of the things a reader may have noticed is a lack of screens that display summaries (e.g. workout summaries after the user has completed a workout would seem useful, as would a display of all the exercises in the workout before the workout begins). This omission is allowed with the Digital Personal Trainer because:




Nevertheless, there will be times when the users will want to view large amounts of information about their workout history, presenting a problem particularly with the reporting task.



Changes for the High-Fidelity Prototype


Tabbed Menu System


            In our modifications for the High-Fidelity design, the main implementation difference was the incorporation of a tabbing-system located on each screen of the software. The reason behind this change was to increase efficiency of use, along with making the interface more easily understood. Since the average user will be interacting with each feature (which correspond to the tabs in the tabbing menu) frequently, the interface needed a way for the user to see each option no matter where he or she is located at any particular time. This also allows the user to always be aware of where he or she is currently located in the system. Because each feature has some independence from one another, the tabbing-system is appropriate since a particular trail need not be followed. One other advantage of this system is the capability for each feature to have a direct link to a help screen to aid the user more efficiently.


Different View Formats


            In the low-fidelity prototype, users were limited in how they were able to view their exercise. Since our underlying goal is to create a versatile and expedient system, we modified the design to allow the user to view their exercise routine in “Week View”, “Day View”, or “Exercise View”. This way, the user can quickly reference which exercises they are supposed to do on a particular day, or refresh themselves about the format for a particular exercise in a routine. One component that we were able to add in this format was the ability for the user to log comments on a particular exercise in the “Exercise View”. This way, users have a more comprehensive logging system to store results.



Preference-Based vs. Phasing


            One of the more confusing components of the low-fidelity prototype was the “Creation of Phases” for workouts. Less-experienced weightlifters would probably feel unclear as to how to set up phases. To make the system more user-centered and applicable to all audiences, we reformatted this component to allow the user to “Create Workout from Preferences” or “Create Blank Workout”. This way, a user has more choice as to how to set up their program and with better clarity. Additionally, the information for setting up a workout is better organized and reduces the depth the user is required to follow to create the workout.



Email Component


            After researching many other current products on the market, we came to the conclusion that an interactive component (with the outside world) was something that would further distinguish the Digital Personal Trainer. As a result, we added an email feature to the results menu. With this capability, the user can easily forward their workout results and statistical analyses to a physician, trainer, or other fitness specialist. This is quite advantageous to many users who are closely tracking their workouts and need further assistance with a personalized program. Once again, this feature adds to our versatility and efficiency design goal.




Usability Testing Process





For the usability test, the design team gathered a good representative sample of the expected users of the system. A total of 8 users were subjected to the test. The profile of test subjects ranged from a 39 year-old male ex-Navy SEAL with 2 decades of weightlifting experience to a 21 year-old female with under a year of experience. The other participants had between 2 and 15 years of experience as weightlifters. All of the subjects indicated that they typically recorded the results of their workouts on paper[1]. The following is a listing of information about the test subjects:



Subject 1:

   Age: 39

   Gender: Male

   Occupation: Heavy Equipment Mechanic, Petty Officer U.S. Navy SEAL (ret.)

   Weightlifting Experience: 20+ years


Subject 2:

    Age: 21

    Gender: Male

    Occupation: Student, researcher at Physics optics lab at UMCP

    Weightlifting Experience: 5 years


Subject 3:

    Age: 21

    Gender: Female

    Occupation: Student/Entrepreneur

    Weightlifting Experience: < 1 year


Subject 4:

    Age: 32

    Gender: Male

    Occupation: Computer Technician

    Weightlifting Experience: 7 years


Subject 5:

    Age: 22

    Gender: Female

    Occupation: Student

    Weightlifting Experience: 2 years


Subject 6:

    Age: 39

    Gender: Male

    Occupation: Professional Powerlifter/United States Marine Corps (ret.)

    Weightlifting Experience: 15+ years


Subject 7:

   Age: 19

   Gender: Male

   Occupation: Student

   Weightlifting Experience: 2 years


Subject 8:

    Age: 34

    Gender: Male

    Occupation: Sales Director

    Weightlifting Experience: 8 years





Some of the tasks listed in previous deliverables for this project turned out to be somewhat beyond the expertise of the design team. For example, to generate a workout based on past workout performance in addition to user preferences involves a rather complex constraint satisfaction problem (i.e. Artificial Intelligence) that no one in the team is familiar with. This limited the range of testing tasks a bit, but not to a critical extent. The basic tasks tested during the usability tests were as follows:



The usability test itself gives a more detailed description of the tasks performed during the test. (To see a copy of the Usability Test, see Appendix A.)




The pretest, usability test, and posttest were all modified slightly after a brief pilot usability test conducted in late March indicated ambiguity and confusion in some of the questions. Also affecting changes in the tests were comments made by Dr. Shneiderman on our usability test deliverable, as well as the fact that some facets of the project listed in the usability task list could not be implemented[2], or had not yet been implemented.



The Usability Test


There was one problem with the administration of the usability test: we did not have a

 Pocket PC available on which to download the prototype. So instead of having the ideal situation where we could perform the tasks in a gym, we had to settle for a PC-based ‘emulation’ of the user’s domain. Subjects were asked to perform some of the tasks in the usability tests as if they were in a gym, and report if they believed any problems would arise from trying to perform some of the tasks in the gym as opposed to in a comfortable room on a laptop.


All of the tests were administered either in an apartment in South Campus Commons or in the home of one of the test subjects. To simulate the noise of the gym, subjects completed the usability test with a television playing rather loudly in another room close by. Subjects were seated at a couch with the laptop on a coffee table. After completing the pre-test, a subject would be given very brief verbal instructions on how to use the Digital Personal Trainer (no further instruction would be allowed after that point). At this point, the administrator started the PC-based DPT and the user was asked to complete the tasks in the usability test. Once completed, the administrator would ask the subject for general comments about his or her experience in completing the tasks before administering the posttest. Subjects were tested one at a time, and in isolation of one another.



Test Analysis and Resulting Problems


For the most part, users completed the tasks with relative ease given the instructions they were provided with prior to the test. They all reported that setting up the user profile and phase goals was the easiest part of the system, as the program seemed to guide them from one action to another. Most of the participants, however, stated that without the instructions, other aspects of the trainer would have been a bit more difficult to use. The problems that occurred multiple times with multiple subjects, along with the overall importance and difficulty rankings based on the post-test are elicited in detail below:



Problem 1: Main Menu Tabbing system is not ‘action oriented’


Description – After creating the user profile upon starting the system, the user is left with a main screen that has a series of clickable tabs to take them to different parts of the program. The intent of this was to allow users extremely quick access to every part of the system from anywhere in the program flow. Most of the users, however, said that they would like to have buttons on the main menu in addition to the tabs, such as “Create Workout, View Progress, etc.


Importance – 3/5

Difficulty – 2/5



Problem 2: Difficulty ending the program


Description: It was overlooked by our design team that the DPT could only be exited from the main menu. Most of the users were looking to end the program once they had finished recording the results of the final exercise in their workouts, but found themselves having to look for the main menu to exit


Importance – 3/5

Difficulty – 1/5



Problem 3: Difficult backward navigation in “View System” used in creating and maintaining workouts


Description: Workouts are created in the high-fidelity prototype by choosing to either make a new blank workout or a workout from preferences in the ‘Overall View.’ This selection takes the user to the ‘Week View’ where they can view the general information for each day’s workout. Clicking on one of the days takes the user to the ‘Day View’ where they can view all the exercises for a given day’s workout. Clicking on one of the exercises takes the user to the ‘Exercise View,’ where the user can record data about sets performed for that exercise. This system is intuitive and easy to use going from the ‘Overall View’ down to the ‘Exercise View’ (i.e. forward navigation), but not vice-versa. In particular, users were looking for a quick and easy way to go from recording their data in the ‘Exercise View’ back to the ‘Day View’ (in effect, they were looking for something like a back button that you would find in an internet browser.) Instead, they had to use the combo box at the top of the screen to go back. The bottom line is that backward navigation was not made as easy as forward navigation.


Importance – 4/5

Difficulty – 3/5



Problem 4: Awkward transition between creating a workout and editing the workout


Description: In the Make/View Workouts section, the user clicks one of two buttons to create a workout. Once the button is clicked, the name of the workout appears highlighted in a list of other saved workouts below. The user can then press the ‘View’ button to view the details of the workout in the ‘Week View.’ The test subjects, however, anticipated a different behavior, and this led to a bit of confusion. In the posttest discussions, all eight of the subjects indicated that they were expecting the details of the workout to show up on the screen as soon as the ‘Create Workout’ button was pressed. It did not prevent them from completing the assigned task, however it is an important problem since it caused a hiccup in the test for all eight subjects.


Importance – 3/5

Difficulty – 2/5



Miscellaneous Problems and Issues


Smaller problems came up (as expected) because there are parts of the DPT that have not been implemented or tuned yet. Included among these unimplemented parts is the ability to choose exercises after creating a custom workout, change exercises in a pre-existing or preference-based workout, and edit the user profile. So excepting the unimplemented sections of the DPT, the following is a listing of the smaller problems that were encountered by a minority of the test subjects, but must be addressed nonetheless:



None of these problems poses a serious technical challenge, and they can be remedied rather easily. However it is important that all of these problems be fixed, because with those problems in place the end product will appear unpolished.


[1] This information was not gathered in the pretest or posttest. Rather, it was gleaned from informal conversations prior to the administration of the test

[2] See the first paragraph of the previous section entitled Tasks for details

Web Accessibility