Zooke is a solution which attempts to
help alleviate the problem of media asset management. By leveraging point
of capture metadata available on mobile devices, as well as
a community of networked users, Zooke provides a platform
for users to annotate images. By distributing the annotation
process among the community the annotation burden is alleviated
and lessened for each individual, thus making it more likely
that each user will participate in the process and collectively
more images will become annotated. Zooke uses game play as
incentive and motivation for users to engage in this annotation process, yet allows the process and annotation goal itself
to remain transparent to the user. The game, instead, presents
with a compelling way for users to creatively explore and
share their own real-life worlds with a community of engaged
individuals. By bringing the user's world into the
game, we create a rich imagery-centric interaction exploring social
actions of individuals and communities and also create a reusable
repository of user-verified annotated images for future use
and distribution.
Note:
Grey italics are works in progress, notes, or things we have not yet
determined.
Tasks can involve one of three types of actions:
Taking an image of something - person, place, object, or action.
task is completed when x% players social group say task is completed. Population is asked if image fulfills task (yes/no question)
reputation is affected by photo completion and level of task
Verifying that another player in the system has taken an appropriate picture (i.e., that it is annotated correctly.)
Ranking a task generated by another player.
Generating a task – creating and sending a tasks to players (for now we assume that players may assign a specific player a task)
Each player will have a
reputation. The reputation will be maintained by playing fairly and
honestly. It is analogous to eBay’s reputation stars. However
the player’s reputation will be determined on a set number of
criteria:
Picture taking
Good/Bad picture. How accurate they are in completing a task. If
they take photos which are always verified by the community as
matching their particular task, then they receive good rating,
however, if their photos do not get verified within the community
their reputation decreases and their picture is rated as bad.
Task verification
Good/Bad. If all other players verify the photo and one player does
not, then that player’s honesty will decrease, but, if that
player verifies all photos correctly, as the community does, then
his/her reputation increases. (We
may want to make this two things??) How do we add
subjectivity? We just know it exists. How about a forum for task
completion?
Ranking task
difficulty. How well they rank tasks compared to other players in
the same level?
Creating
good tasks?
Points: Points are gathered by completing tasks. The more tasks a player completes the more points a player accrues.
Points are assigned to each task, based on the difficulty. They represent the accumulated task completion points.
For example:
Joe has the following reputation of 84% that is determined as:
|
Task
|
Total attempts (cnt)
|
successfully completed (cnt)
|
% of task completed
|
total pts
|
|
|
picture taking
|
10
|
7
|
70
|
17
|
|
|
picture verification
|
10
|
5
|
50
|
-
|
|
|
tasks ranking
|
10
|
10
|
100
|
-
|
|
|
task generation
|
20
|
20
|
100
|
-
|
|
|
total
|
50
|
42
|
84
|
|
|
Points are only assigned to photo challenge completion tasks. In the above axample, the 17 points resulted from the seven picture challence tasks completed with the following distribution:
3 tasks completed at difficulty level 1 resulting in 3 points
0 tasks completed at difficulty level 2 resulting in 0 points
2 tasks completed at difficulty level 3 resulting in 6 points
2 tasks completed at difficulty level 4 resulting in 8 points
0 tasks completed at difficulty level 5 resulting in 0 points
for a total of 17 points
Players are penalized
for having bad reputations. Should their reputation not be
satisfactory, then they may not advance to the next level. They must
first remain within the level they are in for a longer time and
complete more tasks until their reputation has increased enough to
proceed. They keep receiving tasks until
they increase their reputation.
Time:
When
a task is generated there might be a time assigned for completing the
task, but if the task is not completed within that time frame then no
penalties awarded. This is because we do not have a way of
determining when a task is taken by a player. When the time expires
the task may disappear from the task list.
Rating Tasks:
The goal of rating a
task is to estimate how difficult the rest of the community will
determine a task to be and assign a ranking to it. Only players who
have completed level one (newbie) will be able to create new tasks.
Upon creating a new task the player will submit the task. The system
will then distribute the task for rating to members of the same
level prioritizing to their social group. All player ratings will be
averaged and that rank will be assigned to each task. Rating
outliers both above and below the average, obviously have not studied
their community well, and will therefore be penalized by decreasing
their reputation.
Prove It: A player will
have a chance to gain extra points by proving that they deem a task
as very easy when the rest of the community rates it as hard. “Prove
It” will give that player a chance to prove they can perform a
difficult task, which they rated easy, in a prompt manner. A point boost (double points) will
be given to an outlying who rates a task
significantly easier than the rest of the community, yet is still
able to complete the task. Should a
“Prove It” situation occur and the user is unable to
perform the task, then that player will loose more reputation than a
normal player.
Verifying
Each player’s
tasks are verified by the community. In order to verify, player 1
will receive a task that asks them to verify player 2’s task.
By responding yes or no (or selecting the correct answer from a
multiple choice list, or matching test, etc.), player 1 will verify
player 2’s task entry. Should player 1 be an outlying
player of the community for verification of player 2’s
task, then player 1 will have unsuccessfully completed the
verification task.
Only after a task has
been verified, will it be submitted to the database and stored.
Annotation
Annotation is a two
phase process. The first phase of annotation is the completion of a
task by one player. The second phase is verification of that task by
other players. By allowing for a correct verification process to
occur, a validated annotation is ensured and all data stored in the
database is deemed to be annotated correctly.
Community and
Community Building
A player can either a
member of a social group by either being invited to that social group
or they can be a member of the social group of non-invited players or
public social group. A community is
defined by a social group. The sense of community and community
development is a salient feature of this game. Players must monitor
and watch their community. Therefore, they must have a sense of
involvement with them.
When a player creates a
task, the task will be assigned to players in their social group
first (assignment of tasks will be prioritized to players in their
social group) and then maybe to other players at the same level. The
player can also choose to assign the task to a selected group of
pseudonyms. The task is then assigned to these players as well as
other players in their social group and to other players at the same
level.
Multiple Communities:
In this model communities are established, on any one of the
following:
free agent model
level of play
location and level of play
an invitation system, much like friendster, where someone must invite you to join a community
Community involvement
is the crux of this game. The community and ability to know who is
in your community, by viewing profiles, collaborating on tasks, etc.
will have a very large impact on how the game is played. For example
in one model, you may be able to select players within the community
to send a specific task to, but in another you will never know who is
being sent your task.
The involvement of the
community to the player is great. One model may represent players
who would only like to play part time at their choice. The other, at
time of sign up the player is automatically placed into a community
and their life within the game begins from that time forward. They
are playing at all times, even when they are not. They are always
living in the community.
We are continuing to
explore which model or mixture of models will make the most sense for
this game and how we can effectively leverage the sense of community
to make the game as easy, simple, and
fun as possible.
Beginning of the game:
You can either get
invited by another friend who is already playing or you can just
generally sign up into the system. If you get invited into the game
you are associated with to the player that invited you and indirectly
associated to players they are associated to.
Demotion
Should a player fail to
watch their community, participate in the community by sending tasks,
or otherwise have their reputation drop to 25% full, then they will
receive a warning notice.
Should their reputation continue to fall to below 15%, then
the player will be demoted to a prior level/community.
Promotion
When a player’s
reputation has remained at XX% for quite
sometime, then they will receive a notice of promotion to the next
level. A player may, however, decline the promotion and remain
playing at their current level.
Lurkers
Players that do not
participate in the game and simply look at other players photos will
be tagged as lurkers. Their reputation will drop to zero and they
will have to consult a professional (maybe
receive a task from a high player to rejoin the newbie community),
in order to nullify their lurker status.
Bonus Tasks:
Tasks may also be
assigned by the system or sponsors. These system or sponsor created
tasks will be bonus tasks. The players may not only receive points
in the game/ but also receive real world bonuses, such as t-shirts,
getting entered into contests, etc. Privacy issues and release of
copyright to the sponsor may also have to be addressed with sponsor
related tasks. A system bonus task may give a player a way to gain
double the points.
Multiplayer
tasks?
Location-based
tasks?
PC Component
There will be a
component of the game which allows players to view their own (and
others’) photos from a PC and continue game play as well.
Players may decide to take photos via their camera phone and verify
and rank photos via the PC. They will be able to browse profiles and
photos of members of their community (and
possibly other communities). They will be able to see what
other members/communities have responded to tasks. They will also
be able to view how their own tasks were responded to.
(Users may be allowed to take photos
with other devices, upload, and annotate via the browser.)
We think this will have
a very unique effect of intriguing individuals to create tasks. Not
only will they be able to continue to play as passive/observational
players, but one can imagine if a designer was looking for
inspiration on a certain subject matter and they entered that as a
task. Within days or even hours, they may have 10 or 100s of
responses to their one task. We believe that people, even without
looking for inspiration, will just like to see other individual’s
photos of something they tell them to photograph.
Added Value
At the completion of
the game (or on running) a verified correctly annotated photo
database will exist. The system will put all photos which did not
get x% of verification into an unaccepted category and not add them to the
database. This unaccepted category of photos could be later studied for game improvement. The rich value of this information would be of value to
both sponsors and researchers. Privacy issues will have to be
addressed, as to who owns these photos and who has access to them.
Before signing up players may have to approve a privacy
waiver.
What will we will
create this term?
In the short term, the
next few weeks, we will be mocking up (in paper or other low-fi
mechanism) an initial screen concept of the phone interface and maybe
the PC browser as well. We will also research previously made games
of this sort, in order to explore ideas of community and strategy, as
well as interaction mechanisms. Additionally we will be interviewing
researchers who have previously explored mobile communities and
social gaming to gain further knowledge in this arena.
We will also research the annotation and metadata aspects of the game. We will look at the current annotation process both in Flamenco and
MMM (Mobile Media Metadata) project to try to understand what are some of the lessons learned from those applications. We are proposing a distributed
annotation approach as part of the game architecture. The system architecture will need to take this into account.
Preliminary Schedule and Work Distribution/ Roles for this term.
|
Week
|
Milestones/ Goals
|
Work Distribution/ Roles
|
|
Nov. 24-30
|
Finish
game rules
Personas
Goals
Task
Analysis
|
Finish
game rules - All
Personas
- Erick
Goals
- Nathan
Task
Analysis - Anita
|
|
Dec. 1-5
|
Scenarios
Flow
Diagram
Competitive
Analysis
Initial
Designs
|
|
|
Dec. 8-12
|
Initial
Low-fi Prototype
Presentation
Report
|
|
|
Dec. 15-20
|
Presentation
Due Dec. 15
Report
Due Dec. 17
|
|
**Usability
Testing and Iterations to ensue throughout the break
|