Virtual
reality gets easier
By
Kimberly Patch,
Technology Research News
A common complaint about software is it
is not easy to use. If it is difficult to design an intuitive word processing
program, however, it's even harder to find obvious ways for people to
control virtual environments.
In the dozen or so years since the graphical user interface (GUI) caught
on, we have come to expect standard widgets like buttons and menus, and
consistent ways of interacting with the interface, like dragging and dropping.
So even if it is difficult to find where the programmer has stashed the
"options" menu, it's a matter of looking under all the menu items at the
top of a window rather than wondering where to begin.
Although virtual reality software is not brand-new, the tools that allow
people to develop virtual environments do not provide standard widgets
and interaction techniques; this is largely because virtual reality programs
immerse users in environments that don't all look alike, said James Willans,
a research student at the University of York in England. "Virtual
environments [lack] consistency between applications because they're developed
to simulate real world interfaces or to support highly specialized tasks,"
he said.
This makes it difficult to develop standard ways of getting around them,
he said.
York University researchers have developed software that separates the
process of designing interaction techniques from the process of building
a specific virtual environment, making it easier for developers to design
realistic interaction techniques and try them out on users, according
to Willans.
The Marigold toolset is an attempt to find easier ways to design virtual
environments like flight simulators where it is important to make the
interaction between the user and the environment as realistic as possible.
Making a virtual environment involves using a three-dimensional modeler
to build objects that will reside in the virtual world, then using a virtual
environment toolkit to define the parts of that world that will be interactive.
Although building objects is fairly intuitive, designing how users will
interact with the objects is much more complicated.
No single interaction technique is suitable for all applications, even
for abilities as basic as the way a user is allowed to move around the
environment, said Willans. For instance a user could "move quickly from
one location to another or [be] constrained to moving a certain speed,
or interact in such a manner that they can tell from the position of their
physical body the position and orientation of the virtual world," he said.
This is where the trouble usually starts, according to Willans. Just like
GUI development tools, virtual environment toolkits provide predefined
interaction techniques, but the virtual environment tools do not work
well across all the environments developers may want to create. This often
results in developers choosing standard interaction techniques "without
considering the precise nature of the application they're supporting,"
said Willans. The result is an environment that may look very good, but
won't be used if it is difficult to navigate, he said.
Even when developers try to consider interaction in light of the environment
they're building, the tools don't offer much support, he said. "The abstractions
used in virtual environment toolkits make defining the interaction complex."
The toolkits use simple geometric objects to represent how the environment
responds to a user, said Willans. Visually, "these bear little little
or no correspondence to our real-world understanding," he said.
The researchers attempted to solve these problems by giving developers
a way to visually specify the techniques they're designing.
Marigold uses a flowchart tool developed by researchers at York University
and Rutherford Appleton Laboratories that diagrams interaction techniques.
Marigold then builds a prototype from the diagram so the developer can
try out the interaction techniques in context, according to Willans.
This allows the developer to find better ways for users to interact with
a specific environment. This transition between the design and a prototype
of the design "can be used to evaluate the suitability of the design with
users," said Willans.
The tool automatically checks the design to make certain that it will
support specific requirements, Willans said. "For instance, whether...
a virtual door cannot be opened when locked." It also checks to make sure
users can understand from what they see on the screen how their interactions
will be interpreted, he said.
The work is "in essence trying to create a better specification language,"
said Scott Hudson, an associate professor of human-computer interaction
at Carnegie Mellon University. The work is related to earlier work on
GUI development tools that use specifications, but the researchers have
applied the approach to virtual environments, said Hudson. In doing so
the researchers have made some "significant advances," adapting them to
accommodate the more fluid interactions of virtual environments, he said.
Usability problems are a serious issue because the keyboard and mouse
are often not viable in a virtual environment, said Scot Thrane Refsland,
director of research and Development at the center for Design and Visualization
at the University of California at Berkeley.
"The challenge is to create new ways of interaction that enable the
user to interact effectively. As is the case with many industries, sometimes
the tools aren't developed by the people who use them, creating bad interfaces
that aren't effective," he said. The York University tool addresses
the problem by probing deeper into the issues of modeling and verifying
interaction techniques, he added.
Developing interfaces for virtual environments is a difficult problem
not only due to the lack of viable standardized interaction techniques,
but also because human behavior is so very complicated, said Hudson. Because
of this, "we have little practically applicable theory that can help us
design usable systems from base principles" he said. The usual fall-back
is a cycle of testing with users and redesigning, which makes it especially
important to be able to test interfaces during the design process, he
said.
Willans research colleague is Michael Harrison of the University of York
in England. They published the research in the August 1, 2001 issue of
the International Journal of Human Computer Studies. The research was
funded by the UK Engineering and Physical Sciences Research Council (EPSRC).
Timeline: Now
Funding: Government
TRN Categories: Human-Computer Interaction
Story Type: News
Related Elements: Technical paper, "A Toolset Supported
Approach for Designing Testing Virtual Environment Interaction Techniques,"
International Journal of Human Computer Studies, August 1, 2001; Marigold
Website: http://www-users.cs.york.ac.uk/~jwillans/marigold/
Advertisements:
|
November
7, 2001
Page
One
Hubs key to Net viruses
Electrified
water spins gold into wire
Virtual reality gets easier
Laser emits linked photons
Dye brightens micromachines
News:
Research News Roundup
Research Watch blog
Features:
View from the High Ground Q&A
How It Works
RSS Feeds:
News | Blog
| Books
Ad links:
Buy an ad link
Advertisements:
|
|
|
|