Think Aloud

The Setup:

We tried to simulate a real setup, we used a beamer to simulate the public display and positioned the kinect under the projection surface. We designated two people of our team to observe the test subject, only those two people were allowed to ask question. The test took place in a meeting room because there we could calibrate the beamer and kinect for our purpose.

The Task:

We developed different stages for the tasks. The goal of the first stage is to start the game, we asked the test subject to stand outside of the view field of the kinect and then enter the "playground". 

The second stage was to win the game, this one was some kind of a trick question because you cannot win the game. You can only earn so much score till you run out of lives. 

We tried to aim for a full circle of the game in about 5 minutes.

Test Subject #1

Sex: Female

Age: 28 

Prior Knowledge: None, never used a kinect or other motion controls

Issues: 

No indication if the user is properly tracked. (3)

Interaction possibilities not clear. (2)

Game rules not clear. (2)

Test Subject #2

Sex: Male

Age: 21

Prior Knowledge: Sparse experience with motion controller but expert in gaming

No indication if the user is properly tracked. (3)

Fazit:

To solve the problem of the uncertain state of tracking we gonna implement a shadow which will follow the horizontal movements of the user, which should indicate the tracking.

The game needs some kind of interactive tutorial to introduce the game rules and interaction spectrum, a scripted start in which the different game elements and their effects are described, with such a tutorial the interaction possibilities are easy to learn.

New Game

We collected feedback from different people and tried to act on it, which had a complete change of our current game as consequence. We will describe our new idea and give a overview of the features and how the game is played.

Our new game idea is inspired by games like Tetris, Bejeweled and Fruit Ninja. We want to create an enviroment for the player in which he has fun but at the same time some tension in regards to his personnel best.

Player sees a wall, on which different elements descend in three columns. His goal is it to collect gems to score points and prevent bombs to reach the floor. If a bomb reaches the floor the player loses one of his initial three lives. The player can increase the number of lives he has through collecting stars, which also descending on the wall. To collect and destroy objects the players has to throw balls at the wall. He has to throw the ball at the right height to hit a target, in addition his position in front of the wall determines at which column the ball is aimed.

A short list of possible game machanics (subject to change):

•  Different ball sizes to in/decrease the chance to hit a target
• Different ball speeds determine the height at which the ball has to be fired
• Fire multiple balls at mulitple columns at once
• Increasing speed of the items over time to increase the difficulty
• Gems which will fall apart into smaller gems, the player needs to hit them multiple times to collect the points

Information which should be visible to the player (subject to change):

• Number of current lives
• Current score
• Current active game effects
• List of the different tiers of gems

   

Heuristic Evaluation

We are going to discuss our game in the context of heuristic evaluations and we will derive necessary changes to our ptototype:

We developed two heuristics of our own for the game:

Average Speed to complete a action

The duration of the execution of an action is very important for the feeling of a game. It can make the difference between an action-packed game and a very boring expierence. Our goal should be to make action feel very fast and impactful.

For the average time for a move we can increase/decrease the roation speed of the cubes, those rotations are the only visual impact of the actions at the moment.

Goalvisualization

How clear is the goal of the game at every point in the game. The player should always be aware or it should be seeable what the goals is.

We show a pattern which visualizes how the cubes should look at the end of the game.

Standard heuristics

We need better feedback of the state of the game. To realise this we have to implement some kind of scroring, we decided to show the player a clock which counts the time the player has needed to complete the pattern.

To improve the error correction we will implement the ability to reverse a rotation action which is currently active. For this the user has to do the reverse swipe action, if he wants to abort a left rotation he has to do a right swipe.

A additional aspect of  error correction is the alert when the player is getting lost and the kinect can track him anymore, then the game should display a message.

Take a look at our first prototype

Here is a sneak peak of our first prototype, we started to implement our first idea before we conducted our uoser tests. Sadly the user tests showed, that our ideas were not ready for the public and needed many adaptions:

This will now take more work, than to implement a new fresh prototype. 

We learnt a important leasson for developing a user friendly interface/interaction-scheme:

Do not think your ideas are any good before someone, who is not your team member, says it is good!

week#3: Who did what?

Alexa: paper prototyping

Nizar: paper prototyping

Stephen: organising the meeting, paper prototyping, pictures

Aouatef: questions and task for user testing

SourceControl

Github Organisation: https://github.com/organizations/MpiPublicDisplays

Documentations Repo:
https://github.com/MpiPublicDisplays/documentation

Source Code Repo:
https://github.com/MpiPublicDisplays/waste-o-mat

Change is imminent

We got great feedback from our testers, we will summarize the important points and derive changes to the prototype:

• User want to what is on the other sides of the cubes, to make better decisions for their turn.
• Buttons for changing the rotation mode are not optimal, the users want to use natural gestures like swipe or rotation of their wrist as command to chnage the cube.
• Selection with the left hand are not optimal: When the user is moving the selection changes, they want to make a selection with a command and disable the selection with a other command.
• The initial goal is not clear to the user, when he sees only the cubes.

What we decided to change for future prototypes:

• Implement some kind of goal, which implicit communicates the kind of interaction. This will make the interaction more intuitive and removes the buttons from the UI.
• Find a way to make the overall goal clear for the user (Interactive Tutorial, Scorecounter etc.)
• Implement swype gestures for rotating the cube.
• Implement a command chain for selecting and deselecting a cube.

The changes for the paper prototype are minimal but significant in their impact, we removed the two buttons from the UI and changed the laser pointer selection to a gesture selection. When the user is pointing on a cube the cube is selected. Now the can use swype gestures to rotate the cubes. When the rotation is complete the user can make another gesture (undefined at this moment) and return to the selection mode.

The selection indicator

The indicator should be aware what are the other sides of the cube and visualize them to support the next decision of the user.