In the initial phase of the project
our team gathered data on the problems with the current Mulberry interface.
We conducted user tests and heuristic evaluations. Contacted technical
support departments of institutions running Mulberry and asked them
what were the most commonly reported problems. The data we collected
was sorted and prioritized. Finally, we decided to focus our efforts
on five areas.
Interface Simplification
Attachments
Address Books
Preferences
Mailbox/Folder Management
Below you will find information on our research and
our recommendations. Please note that not all of the changes recommended
will be implemented into new versions of Mulberry. That is up to
Cyrusoft.
Each year a new class of 1300 freshmen arrives at CMU. For most
them it will be their first time using Mulberry. Every college or
university that uses Mulberry faces the same situation. This underscores
the necessity of Mulberry being intuitive and easy to learn. This
was the goal of the interface simplification project.
Making software easy to learn is requires a number
of things. First, one must understand how users learn software,
which is of course through exploration (Carroll and Rosen 1987;
Fischer 1991). Users search an interface for the widget that seems
most apt to produce results they require. Many factors affect a
users ability to find what they are looking for. These include the
number of interface elements presented to them, and the order in
which these are displayed. Also important are the labels on the
interface, their clarity, their mapping to users expectations, and
their similarity to one another. Finally, we must understand our
users, their needs and backgrounds, and their mental models of the
specific domain, in this case email.
Our recommendations are the results of rigorous process
including scores of user tests, heuristic evaluations, cognitive
walkthroughs and surveys of users.
Click here to view our full report(pdf)
A major aspect of electronic mail is the ability to send and receive
files. One of the problems we found during our tests of the Mulberry
interface dealt with attachments.
The Mulberry team performed various tests on how people attached
files and found two major problems associated with the current draft
screen and attachments. First, lack of visual feedback that a file
has been attached. Second, confusing terminology, using the word
parts instead of the more familiar attachments.
Various solutions were prototyped and user tested.
Click on the PDF below to see the results.
Click here to view our full report(pdf)
During our initial round of testing the address book presented significant
difficulties to users. Many could not find their own personal address
books, and many more were unable to add addresses to them. We examined
these problems, created and tested prototypes and recommended some
solutions.
Click here to view our full report(pdf)
The preferences area allows users to customize
the functionality of Mulberry to fit their needs. This area also
allows users to set up email accounts and servers to use with Mulberry.
The team conducted background research to determine what parts of
the preferences area should be fixed and to give some insight into
solutions to these problems. The group examined how current users
use the preferences area through think aloud user testing, analysis
of tech support pages, and analysis of common preferences settings.
A word sort and word definition experiment was also conducted to
help determine the organization of the new preferences area. In
addition, a competitive analysis of how equivalent functionality
is implemented in other email programs was completed.
From this analysis, prototypes of the main preferences
dialog were created using Java and Visual Basic. A startup wizard
and an email account setup wizard were also designed to make setting
up Mulberry easier for users. Both prototypes were created using
Mac and Windows guidelines and were iteratively tested with over
thirty-five users. From this functional specifications and finalized
prototypes were created detailing the new functionality.
Click here to view our full report(pdf)
Before embarking on this sub-project,
the Mulberry team had already performed over 100 user tests. During
those experiments we learned a great a deal about the problems with
folder and mailbox management. In response to user difficulties
in this area some general design principles were developed from
which our solution was developed.
Principle 1: Do not duplicate instances of the
same information.
In the current version of the application if the user has one email
account, the word inbox appears in three places on the left-hand
pane. Even our most advanced computer users had trouble creating
sub-folders, unable to discern which inbox to select.
Principle 2: Separate private and shared mail.
To many users the concept of a shared mailbox is unfamiliar. Furthermore
our subjects were often hostile to the idea that their email could
be anything other than private. Better separation between public
and private mailboxes would make novice users feel more secure.
Principle 3: Only show users the most relevant
information.
Currently the left pane has many purposes. Not only does it show
users all subscribed mailboxes, it is also the place where they
browse through the unsubscribed ones. As well as being confusing
this assaults the user with information not necessarily relevant
to their every-day email reading tasks.
Principle 4: Do not reinvent the wheel.
Most people who use mulberry have experience with other email clients.
In redesigning existing functionality we want to make sure that
we followed as many common conventions as we could. In addition,
we want to use a minimum of non-standard widgets.
Principle 5: Put commands in context.
Our solution makes many functions accessible directly from the graphical
areas they affect. This approach solves two persistent problems,
locating a function, and defining its scope.
To see how we solved these problems check out our
recommendations.
Click here to view our full report(pdf)
|
