Biomedical Engineering Reference
In-Depth Information
on the computer display. Good graphical interface design is an art that's difficult to master. For
example, even subtle differences in the relative size of objects displayed on the screen can
profoundly affect how they are perceived. The graphical user interface typically makes use of mental
models, which are the metaphors that give a graphical interface meaning. The desktop metaphor,
with its desktop, trashcan, documents, and file folders, exemplifies how a metaphor can be used to
provide a large number of users who have diverse backgrounds with a conceptual model of how and
where information in a computer operating system fits together. Graphical interface designers have
to make assumptions about the previous experiences of the typical user for an interface to work. The
level of graphic complexity most appropriate for a graphical interface balances the need to focus a
user's attention on a 3D model or other data, with hardware limitations, and the resources necessary
to create a graphical interface.
The logical interface level is about rules, guidelines, and standards of interface behavior, such as how
an interface should display the image of a molecule. A well-designed logical interface layer, like a
properly executed graphical interface layer, allows users to focus on the problem at hand, such as
identifying the location of a specific gene on a physical map, rather than on the mechanics of
operating the interface. Logical interface design relies heavily on the concept of information design,
which deals with the organization, presentation, clarity, and complexity of information. Information
design focuses on communications and on developing a framework for expressing information, not
aesthetics. The primary metric for assessing the degree to which an interface supports a logical
model is commonly referred to as cognitive ergonomics.
Intelligent interfaces rely on a variety of pattern-recognition techniques to adapt to the user's
behavior. Ideally, an intelligent interface learns user preferences by monitoring the user's responses
to certain situations, tailoring the experience to the user's current interests, and never demands that
users explicitly state their preferences. The inner workings of an intelligent interface may be very
complex and rely on an elaborate knowledge base coupled to an expert system or statistical analysis
program. Intelligent interfaces share many properties in common with intelligent agents, which are
independent programs capable of completing complex assignments without intervention, as opposed
to tools that must be directly manipulated by a user. By monitoring a user's activity on a Web site or
within an application, an intelligent interface may learn, for example, user preferences for the color of
a protein's acute region or the responsiveness of the protein rotation to mouse or joystick movement.
Alternative Metaphors
Visualization, whether as part of the user interface or as a means of presenting structure or sequence
data, is largely about creating and supporting metaphors, which transform the data into a form that
means something to the user. The pie chart is a useful metaphor only to the extent that users
understand the difference between the slices of the pie and how that translates to relative quantities.
The pie chart works for most of us because we intuitively understand the metaphor. However, the pie
chart, like the Windows desktop metaphor, isn't very data-dense and doesn't lend itself to
communicating advanced biotechnology concepts, such as tertiary protein structures. As a result,
visual metaphors for user interfaces intended to present molecular biology data are necessarily more
sophisticated than ordinary business graphics, especially when the challenge is to present large
volumes of complex data.
One of the challenges of creating a suitable metaphor for bioinformatics work is the variety of
potential users of the applications and their level of expertise. For example, bioinformatics
researchers, high-school and college students studying molecular biology, research fellows, clinicians,
and even the marketing departments of international pharmaceutical companies may use a given
suite of applications. Devising a reasonable interface metaphor is therefore a compromise between
information density, ease of use, and power—the ability to quickly and easily manipulate data
communicated through the interface.
Bioinformatics is pushing the metaphor component of visualization technology to new levels. For
example, even though the desktop, folder, and trashcan user interface—introduced by the Xerox
Star, popularized by the Apple Lisa and Macintosh, and fully exploited and commercialized by
Microsoft—is the dominant metaphor on desktop computers, it fails to reflect the needs of
bioinformatics. Many researchers in bioinformatics contend that a new user interface is in order, one
