National Computational Science Alliance:
Facilities; Personnel Exchanges
Mr. John Shalf, NCSA Staff, created the AtomView software
in collaboration with Drs. Ron Kriz and Diana Farkas to view
largescale atomic structures in virtual environments.
Dr. Ron Kriz worked with NCSA staff to organize a CAVE
programming workshop at NCSA, summer 1997.
Ms. Rachael Brady, NCSA Staff, visited Virginia Tech's CAVE
facility to work with Dr. Deborah Hix on evaluation of the
Crumbs virtual environment application.
List does not include CoPIs on this grant.
1. College of Architecture and Urban Studies: Dr. Robert Schubert,
Professor, Associate Dean for Research Administration. Director
of the VELab, http://www.caus.vt.edu/CAUS/FACILITIES/VELab/.
Author of the 'Cybercore History Project',
http://dev8.arch.vt.edu/cybercoreweb/main/, which used
the CAVE to reconstruct the Cathedral at Cluny, France.
2. College of Architecture and Urban Studies: Dr. Dennis Jones,
Associate Professor, Teaches C++ and other computer graphics
courses in Architecture and advised several independent student
CAVE related projects.
3. College of Human Resources and Education, Department of Interior
Design, Dr. Joan McLain-Kark, http://idfl400.hidm.vt.edu/jmksite/,
and the director of the Interior Design Futures Laboratory,
http://idfl400.hidm.vt.edu. Uses the CAVE extensively in courses
and independent student projects in interior design.
4. College of Arts and Science, Dr. Ed Fox, Professor Department of
Computer Science, Director of Digital Libraries Research Lab,
http://www.dlib.vt.edu/. CAVE project: 'CAVE Electronic and
Thesis and Dissertation (ETD) ('CAVE-ETD'),
5. College of Agriculture and Life Sciences, Dr. Tim Mack,
Department Head, Entomology, and Dr. Alexi Sharov, lectures
and CAVE programmer for the 'CAVE-Insects' project,
6. College of Agriculture and Life Sciences, Mr. Peter Sforza,
http://oak.ppws.vt.edu/~sforza/, author of the Virtual Dandelion
http://oak.ppws.vt.edu/~sforza/cyberweeds.html and Virtual
Nematode project, http://oak.ppws.vt.edu/~sforza/nematode.html.
7. College of Arts and Science, Dr. Crandal Shifflet,
Department of History, PI on the Virtual Jamestown project,
includes CAVE 'walk-thru' of Jamestown.
8. College of Engineering, Dr. Michael Deisenroth, Department of
Industrial Systems Engineering, advised Mr. Jeff Sugar's Masters
Project http://filebox.vt.edu/users/jsugar/page7.html (not a
Masters Thesis) 'Space Station Remote Manipulator System', see
9. College of Arts and Science, Dr. Brian Dennison, co-advised Ms.
Caitlin Kelleher, along with Dr. Deborah Hix on her Senior
Project in the Department of Computer Science 'CAVE Developoment
of Astronomy Visualization', which is part of the 'Virtual
Universe Project, http://www.phys.vt.edu/~dennison/vu/
10. College of Arts and Science, Department of Computer Science,
'Problem Solving Environment Research',
http://vtopus.cs.vt.edu/~pse/, where the CAVE is listed as a PSE
resource for a inter-disciplinary list of professors in Computer
Science and other disciplines at Virginia Tech.
11. Central Virginia's Governor's School (CVGS) at Lynchburg,
Virginia, http://www.cvgs.k12.va.us. Tom Morgan, Director of
CVGS, CoPI on Proposal to 'Extending the Use of Collaborative
Virtual Environments for Instruction to K-12 Schools', project
summary of results posted on the Web at
Activities and findings:
Research and Education Activities:
ACQUISITION OF A CAVE(tm):
BREAKING RESEARCH AND EDUCATIONAL BARRIERS BY
DEVELOPING 3D VISUALIZATION CAVE TECHNOLOGY
ORIGINAL PROJECT SUMMARY OBJECTIVES PROPOSED IN 1996:
The objective of this ARI Research Project is to provide access to and
training in the use of advanced Virtual Environments (VE) such as the
CAVE(tm) that will break barriers in Virginia Tech research and
education-research programs funded by NSF and other organizations. As
a CAVE(tm) partner with the National Center for Supercomputing
Applications (NCSA), Virginia Tech will focus on human-computer
interaction (HCI) development and evaluation with specific
applications in simulation-visualization (SV) of complex
multidimensional biochemical, nanoscale ceramics, and fiber-reinforced
structures. Results of this project will demonstrate how scientists
and materials research engineers can benefit from the use of
well-designed user interfaces in a CAVE(tm) environment. These
activities will become part of Virginia Tech's and southwest
Virginia's future Advanced Communication and Information Technology
Center (ACITC), a building scheduled to be completed by the year 2000
and designed to serve on-campus and related regional off-campus
information technology programs. Both HCI and SV activities have also
been targeted by Virginia Tech in its long term strategic plans to
build the ACITC building. In partnership with NCSA the project goal
is to focus on development and evaluation of VEs that will benefit
scientific applications of interest to both Virginia Tech and NCSA and
within the scope of the ACITC program.
CAVE: is a trademark of the Electronic Visualization Laboratory of
the University of Illinois
1.0 RESEARCH AND EDUCATION ACTIVITIES FROM AUGUST 1996 TO
Since this was an equipment acquisition ARI project, research and
education projects proposed in this ARI depended largely on additional
funding from '..NSF and other organizations'. ARI CoPIs sought
outside funding to explore how CAVE virtual environment technology
would be used in a variety of SV areas with an emphasis on HCI. Only
one CoPI was funded to do ARI research, Dr. Deborah Hix for HCI
evaluation of ARI targeted SV areas. A comprehensive list of CAVE
related research proposals is posted at
In Section 1 we highlight ARI research and other significant CAVE
research and educational projects since September 1996.
1.1 Equipment Acquisition and Construction:
The project was funded, however, the proposed budget was cut from
$1.1M to $850K for equipment acquisition only with university cost
sharing at $797K, for a total budget of $1.65M. Dr. Hix was paid from
university cost sharing funds.
Significant events for the acquisition and operation of the CAVE
facility as proposed are highlighted below. A detailed chronology of
projects events is outlined at
CAVE Event Chronology:
- NSF funds were received on August 30, 1996.
- P.O. for the SGI Onyx Rack CAVE computer & system sent September
- SGI Onyx Rack CAVE computer delivered October 11, 1996
- SGI Onyx Rack CAVE computer operational in Hancock Hall, January
- Pyramid CAVE system shipped and received February 10, 1997
--- unexpected delays in securing CAVE site ---
- Temporary off-campus site for CAVE system operational December
- CAVE system configuration complete, faculty and student access,
--- ACITC ---
- CAVE floor with motion base designed, completed, and approved
July 13, 2000
- CAVE system moved to new ACITC building August 13, 2000
- CAVE computer operational, September 2000
- CAVE floor with motion base construction began January 9, 2001
- CAVE floor with motion base completed February 16, 2001
- CAVE floor modified, CAVE frame and carpet installed March 8, 2001
- CAVE system operational, March 26, 2001 (only took 4.5 years)
Current status of the ACITC CAVE facility can be accessed at
http://www.sv.vt.edu/future/vt-cave/VT/acitc.html. Images of ACITC
CAVE facility are also provided in Figures 1 through 8 in the PDF
1.2 RESEARCH: Target Application Areas:
Additional funding for the large simulations and visualization of
nanoscale ceramic structures came from the NSF Combined Research and
Curriculum Development (CRCD) EEC-9700815 and the Office of Naval
Research (ONR) BAA 98-014. Funding for development of CAVE
Visualization of Biomolecular Structures was realized by Virginia Tech
internal funding, ASPIRES (A Support Program of Innovative Research
Strategies), 'Application of Visualization and Haptic Feedback to
Enhance Molecular Docking'. Additional funding for development for
haptic force feedback for nanotechnology was funded by the National
Institute for Standards and Technology (NIST). No funding was
realized for research in Failure and Reliability in Fiber-Reinforced
Metal and Ceramic Structures, because Dr. Curtin accepted an academic
appointment at Brown University prior to the construction of the CAVE
at Virginia Tech.
1.2.1 Large Scale Simulation & Visualization of Ceramic Structures
Collaboration with NCSA:
John Shalf (NCSA staff) developed the AtomView SGI-performer based
software to view simulation results of large scale ceramic structures
in the CAVE, in collaboration with Drs. Kriz and Farkas, summer 1997
at NCSA. See Access Article:
Web site for Dr. Farkas's Center for Modeling and Simulation in
Material Science (CMSMS) (http://www.cmsms.vt.edu/) demonstrates
research activities related to CAVE technology.
NSF CRCD & PACI:
Additional funding from the NSF CRCD program provided support to
create a distributed interactive visual computing environment for
research collaboration and curriculum development using Java and Web
technologies, see http://www.jwave.vt.edu/crcd/. With NSF PACI
funding (CISE/ASC PACI NSF 96-31) we extended AtomView by first
building the CAVE Collaborative Console (CCC)
(http://www.sv.vt.edu/future/cave/software/ccc/) and then extending
AtomView into this collaborative environment and creating CCC_atom,
(http://www.sv.vt.edu/future/cave/software/cccatom/) with funding from
1.2.2 CAVE Visualization of Biomolecular Structures
Collaboration with NCSA:
Pauline Baker and Randy Heiland (NCSA staff) developed the Immersive
Docking (IDock) program for use in a CAVE. Virginia Tech installed
IDock on the Virginia Tech CAVE and began the development of linking a
'PHANToM' haptic force feedback device with DIVERSE (Device
Independent Virtual Environment Reconfigurable, Scalable and
Virginia Tech ASPIRES and NIST DIVERSE:
Additional internal funding from ASPIRES provided support to develop
a link between the PHANToM haptic feedback device and IDock using VRPN
(Virtual Reality Peripheral Network
http://www.cs.unc.edu/Research/vrpn/), developed at the University of
North Carolina. With additional funding from NIST, the DIVERSE API
provided additional links needed to interface with existing PHANToM
VRPN device drivers. On February 23, 2001 we successfully linked the
PHANToM device running on a Windows workstation to a Linux workstation
running DIVERSE. We can now confidently predict the operation of the
PHANToM device linked to the IDock program running in the CAVE. This
research constitutes a significant portion of Mr. Sanjiv Parikh's
Ph.D. dissertation in Biomechanics in the Department of Engineering
Science and Mechanics. This system will also be used by NIST
researchers for their nano-technology projects that will link the same
PHANToM haptic device to I-Desks and nano-technology hardware for
insitu fabrication of nanostructures.
1.3 RESEARCH: Human Computer Interaction (HCI)
Collaboration with NCSA:
Ms. Rachael Brady, co-creator of the Crumbs CAVE application for 3D
visualization of radiological images, visited Virginia Tech, summer
1998, and worked with Dr. Deborah Hix and Master's candidate Mr. Kent
Swartz on the HCI evaluation of the Crumbs CAVE application user
interface. Mr. Swartz also visited NCSA where he continued to work
with NCSA staff on Crumbs evaluation. Results were presented and
published in the Proceedings of Virtual Worlds and Simulation
Naval Research Laboratories (NRL):
Dr. Hix's pioneering HCI research work with virtual reality
researchers at NRL's Virtual Reality Laboratory includes an Immersive
Work Bench (IWB) in the VE laboratory adjacent to the CAVE. Her
research focuses on user application for battlefield visualization
that runs on the IWB, in the CAVE and on a desktop, as well as another
application for situational awarness of joint forces battlefield.
DIVERSE is currently being used on the IWB with an eye-tracking
system, see Figure 8.
3DI (3D Interaction) Group, http://vtopus.cs.vt.edu/~bowman/3di:
Dr. Bowman's research also focuses on HCI and usability. Dr. Bowman
was hired in 1999 as a tenure track assistant professor in Computer
Science for his research in usability in VEs. Dr. Bowman has created
a HMD work area as part of the VE laboratory in the ACITC adjacent to
the CAVE. The DIVERSE API has been used to build the JIVE (Just In a
Virtual Environment) for usability evaluation
(http://csgrad.cs.vt.edu/~cwingrav/Jive/home.html). Mr. Wingrave's
Master's thesis, 'Optimizing Virtual Environment Selection Techniques
with Machine Learning' will be published on the Electronic Thesis and
Dissertation (ETD) Web site http://scholar.lib.vt.edu/theses/, Spring
Semester 2001. JIVE is used for development and evaluation of VE
interfaces in this Masters thesis.
Computer-Supported Cooperative Work (CSCW):
Drs. J. Carroll, M. Rosson, and R. Kriz co-advised Mr. Kevin Curry on
his Masters thesis, 'Supporting Collaborative Awareness in
focused on issues in the development of the CAVE Collaborative Console
(CCC) and was later funded by NSF PACI. Dr. Carroll is the director of
the Center for Human Computer Interaction (CHCI)
1.4 RESEARCH: Real-time Large Scale Physics-Based Simulations in
ONR DURIP & NAVCIITI and NIST:
The most significant CAVE related VE research project was funded by
ONR and NIST to develop the DIVERSE API, http://www.diverse.vt.edu.
This project was not originally proposed in the NSF ARI proposal but
was motivated by difficulties in development of collaborative VE
applications with existing CAVE software that was used in the
development of CCC and CCC_atom (previously mentioned in sections
1.2.1 & 1.2.2).
ONR provided additional DURIP (Defense University Research
Infrastructure Program) funds to extend an existing ONR MURI
(Multidisciplinary University Research Initiative) to build a
real-time physics-based simulation of a virtual Navy crane-ship in a
CAVE (PI Dr. Ali Nayfeh). For this project DURIP funds were used to
install a motion base in the floor of the CAVE, see Figures 4 and 5.
This real-time CAVE simulation is used to design future 'smart'
crane-ships where it is necessary to include the human-in-the-loop in
the design of a system where large payloads can be safely transferred
from ship-to-shore or even ship-to-ship under simulated high sea
states. Such a system required a fundamental redesign of CAVE
software that included a 'Device Independent Virtual Environment'
design philosophy which incorporated a variety of different hardware
devices in a remote shared memory architecture. An essay of the
DIVERSE Toolkit (http://thor.sv.vt.edu/DTK/essay.html) provides a more
detailed explanation of the DIVERSE Toolkit design philosophy.
The DIVERSE graphics interface to SGI's Performer is a separate but
complementary component of the DIVERSE API funded by NIST. Performer
was chosen because our group could quickly develop a graphical
interface that would work across a heterogeneous network of computer
hardware: CAVEs, I-Desks, IWBs (Immersive Workbenches), HMDs (Head
Mounted Displays), and Linux Intel-based workstations. Funding from
the Navy ONR NAVCIITI proposal will create a DIVERSE graphics
interface to OpenGL, to be delivered Fall 2001.
The most significant feature of the DIVERSE API is that, unlike most
VE development software, DIVERSE is licensed GNU Lesser General Public
License (LGPL) and GNU General Public License (GPL). The philosophy
behind LGPLand GPL DIVERSE is to facilitate a wider participation of
the VE community to collaborate, co-develop, and become co-owners of a
new device oriented VE API. DIVERSE has already been successfully
used to develop new VE applications both on- and off-campus.
1.5 EDUCATION: Training
Education and training was targeted in the ARI proposal as a
significant factor for scientists and engineers to use and develop
virtual environment applications.
1.5.1: NCSA Training:
Jongran Lee, a Ph.D. student in Interior Design at Virginia Tech
attended the first Visual Supercomputing Institute -I (VSI-I) at NCSA
August 1996. From this experience Ms. Lee developed her Ph.D.
dissertation on 'Comparing the Effectiveness of Computer Simulation on
Computer Monitor vs. Virtual Reality as Communication Tools in
http://scholar.lib.vt.edu/theses/available/etd-030599-172018/ , which
included the design of the ACITIC CAVE research lab.
Dr. Kriz attended VSI-II (September, 1996). Joan McLain-Kark,
Dennis Jones, John Kelso and Ron Kriz attended VSI-III (August 1997).
Dr. Kriz coordinated his Scientific Visual Data Analysis class
ESM4714 with the NCSA CAVE Boot Camp Training (December 3-7, 1997) so
that Virginia Tech graduate students working on AtomView with Dr.
Farkas could further develop AtomView as their class project.
1.5.2: Virginia Tech Training & Education:
Education: ESM4714, Scientific Visual Data Analysis and Multimedia:
An introduction to CAVE programming was introduced into the existing
Scientific Visual Data Analysis Class ESM4714. Since this class was
project oriented, students were encouraged to develop CAVE related
projects. Class projects can be accessed at http://www.sv.vt.edu.
There were four CAVE class projects in 1997, three in 1999, and six in
CAVE Student Led Users Group (SLUGs) Training:
The SLUGs provided CAVE training with Web pages for students
(February 1999) at http://www.caveslugs.vt.edu/presentations/02171999.
CCC started as a CAVE SLUG project.
Training K12 Instructors:
Two introductory classes on 'VRML in the CAVE Collaborative Console'
were offered at Virginia Tech (June 1999) for instructors from the
Virginia Governors School and at the NCSA Access Center (March 2000)
for instructors from the Maryland Virtual High School (MVHS). Class
notes are posted on the Web at http://www.sv.vt.edu/classes/vrml/.
Education: CS5984, Designing Virtual Environments:
Dr. Doug Bowman created the first course at Virginia Tech that
specifically addresses the design of virtual environments with a focus
on HCI and usability evaluation
Training: Advanced CAVE Workshop:
Dr. Lance Arsenault and Mr. John Kelso participated in the Advanced
CAVE Workshop Series at Old Dominion University, September 17-19 where
DIVERSE was presented and demonstrated on SGI and Linux workstations
and a CAVE.
Training: Advanced Visual Computing Workshop:
The ACITC CAVE and VE laboratory area will be fully functional
summer, 2001. DIVERSE will be installed on 25 Linux workstations. A
workshop on advanced visual computing is scheduled to be taught in
August 2001. Topics will include: VRML & 3D-Models to CAVE, HCI User
Interface Design, DIVERSE & VTK APIs, and Clustering.
Education: Scientific Modeling and Visualization Classroom (SMVC):
Three proposals to build ACITC facilities were submitted in 1996: 1)
NSF ARI Acquisition of a CAVE, 2) Co-sponsored Sun Microsystems, Inc.
and Visual Numerics, Inc, 'Breaking Barriers in Education and Research
with Distributed Visual Computing' (objective was to build the SMVC),
and 3) NSF CRCD 'Computer Simulation of Material Behavior -- From
Atomistic to the Continuum Level'. All three were funded. The SMVC
was used for CRCD classes and a number of other classes and training
workshops prior to relocation in the new ACITC.
1.6 EDUCATION: Software Development
The CAVE Collaborative Console (CCC):
CCC, previously mentioned in the research section of this report,
was also developed to be used for education and distance learning by
two Virginia Governors Schools: 'Extending the Use of Collaborative
Virtual Environments for Instruction in K-12 Schools'
CCC was built with CAVERsoft, Limbo, and the CAVElibs, CCC is a
Performer-based software application that only runs on SGI computers.
Hence for this project it was necessary to purchase SGI computers.
SGI donated one SGI O2 computer and matching funds for a second SGI O2
computer came from the Institute for Connecting Science Research to
the Classroom (http://www/icsrc.org). The awareness tools in CCC were
developed by Mr. Curry as part of his Master thesis. Awareness tools,
such as 'Shared-View' and manipulation of simple molecular structures
on desktop CAVE-simulators, were designed in collaboration with the
Virginia Governors School instructors. Hence CCC was a
The CCC was also developed in collaboration with EVL's Jason Leigh,
Limbo author. CCC added awareness tools to Limbo to enhance
guidelines set by PACI Team-C encouraged collaboration between PACI
team members with CAVE technology, Virginia Tech chose Limbo to
develop both CCC and CCC_atom for both research and education. PACI
Team-C focused on 'Enabling Technology (ET) Data and Collaboration'.
2.0 MAJOR FINDINGS AND RESULTS FROM AUGUST 1996 TO FEBRUARY 2001:
2.1 CAVE Acquisition, Construction, and Operation:
- Partnership with NCSA was essential to the success of this
Once the CAVE system was shipped and received it took about a year
before a temporary site was available that provided necessary access
to on- and off-campus partnerships originally proposed. A temporary
site was necessary because the ACITC building was not scheduled for
completion until 2000. Once this site was acquired, the construction,
hiring, and an operational CAVE were completed quickly, as outlined in
Section 1.1, 'CAVE Chronology of Events'. CAVE systems are not
commercial-off-the-shelf systems that can be setup and maintained with
traditional technical computing support staff. The necessary overhead
knowledge needed for the successful construction and operation of a
CAVE came from our NCSA partners.
In partnership with NCSA, Dr. Kriz was invited to spend the summer of
1997 at NCSA, prior to the construction of the CAVE at Virginia Tech.
From this visit Dr. Kriz acquired the knowledge and experience needed
for hiring CAVE personnel and getting the CAVE operational. CAVE
construction was contracted by Pyramid Systems Inc. In retrospect,
the time spent and knowledge and experience obtained at NCSA that
summer was invaluable. Acknowledgment goes to Dr. Tom Defanti and Mr.
Bill Sherman for assisting Virginia Tech with the computer
configuration, and to John Shalf and Tom Coffin for answering endless
questions and who organized a special CAVE training workshop at the
end of the summer for Virginia Tech faculty. Special acknowledgment
goes to Dr. Larry Smarr, Director of NCSA, whose original letter of
support for this proposal and whose philosophy of access and
partnership provided the foundation for a successful project.
- Technical support staff was critical to successful CAVE
Although we hired qualified CAVE operation personnel (see the next
section), adequate part time system administration personnel was not
fully realized. This continues to be a primary challenge with CAVE
2.2 CAVE access:
- A user friendly environment was essential to developing CAVE
Many CAVE users have commented on how well the CAVE was setup up for
access. This was largely due to hiring a well qualified CAVE Research
Associate, Mr. John Kelso. He created simple to use procedures to
improve CAVE access and remote on-site support at satellite CAVE labs
to promote collaboration. This included providing customized
interfaces that allowed CAVE users access to essential CAVE functions
such as resetting tracker daemons, killing existing CAVE programs and
freeing CAVE processes -- all of which would have otherwise required
system root access. Mr. Kelso also setup a video recording system for
CAVE users to record their CAVE sessions on video tapes. Mr. Kelso
had also helped create the 'GLUE' project at the University of
Maryland, where hundreds of UNIX workstations were 'glued' together
into a shared network environment. This experience facilitated work
with our on- and off-campus remote-site VE labs.
These user friendly interfaces not only made the CAVE more usable but
also enhanced development of CAVE applications. In addition, Mr.
Kelso worked with faculty to create the HCI labs in the Department of
Computer Science. This experience contributed to HCI goals set in
this ARI proposal. A summary of Mr. Kelso's first year's job
performance provides details of essential requirements for CAVE
operation at http://bleen.sv.vt.edu/~kelso/CAVEwork.html. Mr. Kelso's
Masters thesis in computer graphics from George Washington University
(advisor Professor James Foley) provided the necessary experience to
be co-developer and co-author of the CAVE DIVERSE API with Dr. Lance
Arsenault (discussed in section 2.4).
- Students played an important role in developing CAVE applications.
A direct result of the open access and user friendly environment was
the formation of the CAVE Student Led Users Group (CAVE SLUGs), see
http://www.caveslugs.vt.edu/. From this student organization several
significant results were realized, including the CAVE Collaborative
Console and an Introduction to CAVE Programming. Another example of
and undergraduate SLUG participation was by Ms. Caitlin Kelleher who
created a virtual astronomy workshop in the CAVE, simulating the
behavior of a black hole for training physics students, see
http://www.phys.vt.edu/~dennison/vu/. Her undergraduate honor's
thesis on this work was entitled 'AVE Development of Astronomy
Visualization'. Ms Kelleher was also awarded Honorable Mention in the
National Computer Research Association honors for Outstanding
Undergraduate Research, in 1999.
- Remote VE lab sites played a supportive role in developing the
In the NSF White Paper that preceded the NSF ARI CAVE Acquisition
proposal, participating faculty shared a common goal: to create a
connection between the CAVE, remote sites, and existing desktop
software unique to their disciplines. For example, Architecture
created the Research and Demonstration Facility (RDF) Virtual
Environment Lab (VELab)
(http://www.arch.vt.edu/CAUS/FACILITIES/VELab/). The VELab has a
one-wall projection system with head tracking. The Department of
Interior Design created the Interior Design Futures Laboratory (IDFL)
at http://www.chre.vt.edu/NearEnvironments/IDweb/idfl.html. The IDFL
has an SGI Octane (MXI) running the CAVE-Simulator. Other labs such
as the Center for Modeling Simulation in Material Science (CMSMS)
acknowledge the CAVE facility as a tool associated with their research
(http://www.cmsms.vt.edu/). Entomology developed VRML files on a
variety of insects used on educational web sites that could also be
viewed in the CAVE. These are shown on their 'Cyber-Insects in the
CAVE' Web site http://everest.ento.vt.edu/~sharov/3d/cave.html. Other
CAVE related projects are listed under 'On-Campus Partnerships' on our
CAVE Web page, http://www.sv.vt.edu/future/vt-cave/VT/#on-campus and
in the section on 'Other Collaborators' in this report. Very few of
these applications represent much more then simple VE walk-thrus.
This reflects a beginners knowledge and skill level of CAVE
programming, significant nevertheless, considering the diversity and
number of disciplines and level of interest.
If the university continues to support the CAVE facility, these
existing CAVE applications will mature in complexity, but their
success will be judged by the content within that discipline, not if
the project is SIGGRAPH worthy. It is interesting that both the IDFL
and RDF-VELab refer to their respective labs as CAVE remote site or
CAVE satellite Labs. Considering the diversity of disciplines and
that the majority of VE development is accomplished at these remote
site labs, perhaps it would be more accurate to view the CAVE facility
as a shared remote site in the same sense that a regional hospital is
a shared resource when needed. We observed that the total immersive
environment of a CAVE is not often needed, but it is invaluable when
the researcher/educator deems it necessary. Even if it is only a
simple 'walk-thru'. Hence, the CAVE facility depends on the
collective need for immersive virtual environments as it relates to
existing use of 3D graphics within each discipline and especially at
the desktop. Simple 'walk-thrus' are not to be played down. Like
'beauty', the immersive VE experience is 'in the eye of the
2.3 CAVE training / education:
- There never seemed to be enough training on how to use the CAVE.
Although ESM4714 and CS5984 provided an educational framework for
students to learn about virtual environments, the fact that both
classes emphasized class projects was significant in developing CAVE
applications within the discipline of the student. CAVE related class
projects (97/99/00) for ESM4714 are listed at
http://www.sv.vt.edu/classes/ESM4714/ESM4714.html. CAVE related class
projects for CS5984 are listed at
http://vtopus.cs.vt.edu/%7Ebowman/cs5984/ student_projects.html. In
both of these courses we encouraged students to work with professors
on existing research or education projects. For example, CCC started
as a class project in CS5734 Computer-Supported Cooperative Work which
later became an NCSA PACI funded project. These classes focused more
on educational goals, not training. What was needed were classes
devoted to training students on how to program in OpenGL or Performer.
For training we recommended that students and faculty take focused
classes provided by off-campus commercial vendors such as SGI or SUN
Microsystems. Since the mission of a university is to educate and not
train, this remains a fundamental problem inherent at most
The only training realized was when we participated in the Advanced
CAVE Workshop at Old Dominion University, October 2000. For beginners
we taught a two day workshop on 'VRML and the CAVE Collaborative
Console' to the Virginia Governors School, June 1999 and again to the
Maryland Virtual High School, March 2000. In both cases these classes
were extremely successful based on the comments from the high school
instructors even though the limit of their use of CCC was a simple
CAVE 'walk-thru'. For both training sessions high school instructors
strongly encouraged us to organize another training session for their
students. For the Virginia Governors School this training provided a
'teach-the-teachers' component for the proposal, 'Extending the Use of
Collaborative Virtual Environments for Instruction to K-12 Schools'.
Results of this project have been posted at
In retrospect more could have been accomplished had we organized more
training sessions on a variety of software tools such as VTK,
Performer, and OpenGL that would have provided training beyond the
beginners level. Now that ACITC classrooms are available for
instruction and training, some of the ARI CoPIs hope to exploit these
resources for education and training. For some faculty, however, the
problem remains to train ourselves first.
- Regardless of access to CAVE training, faculty who are not in
computer science typically choose not to learn how to program
Although CAVE programming workshops were offered at NCSA and Virginia
Tech, in the last 3.5 years of CAVE operation at Virginia Tech only
two faculty learned how to build CAVE applications that were more then
a simple CAVE 'walk-thru'. Typically faculty outside computer science
defer to their graduate student who can build sophisticated CAVE
applications. But of course their experience leaves when they
graduate. This pattern is largely due to the fact that senior
faculty, who are not in computer science, do not have the background
and perhaps more important the time to learn how to program in virtual
environments. With all of the demands on how faculty manage their
time, a major cultural change must first take place that will reward
faculty, outside computer science to learn how to program in virtual
environments. Perhaps this is more easily done with new faculty who
already have a programming background and then this problem will be
solved by attrition.
The CAVE was also used, to a limited degree, for two classes ESM4984
& ESM5984 which were created for NSF 'Combined Research and Curriculum
Development (CRCD): Computer Simulation of Material Behavior - From
the Atomistic to the Continuum Level', EED-9700815,
http://www.jwave.vt.edu/crcd. In these two classes students learned
the theory of material behavior at the nano-scale, micro-scale, and
macro- (continuum) scale and used computer simulation modelsto study
the theory of material behavior at the different scales. Students
used Java Web-based interactive modules to access the computer
simulations and viewed simulation results with Web browser plugins or
in the CAVE if full immersion was prefered
(http://www.jwave.vt.edu/crcd/farkas/modules/examples). Results of
this combined research and education project were published by Kriz et
al. and listed in the 'Publications and Products' section of this
2.4 CAVE Research: Targeted SV Application Areas:
2.4.1 AtomView and CCC_atom (Drs. Diana Farkas, CoPI and
Ron Kriz, PI):
- AtomView was effectively used in the CAVE, but Windows tools
were used more at the desktop, because they were cheaper and
AtomView was used extensively by Dr. Farkas to interpret
supercomputer simulations results. Dr. Farkas used AtomView in the
CAVE when she needed to study very complex 3D structures, but most of
the time 3D tools on PC Windows computers were more accessible.
Although AtomView ran on Dr. Farkas's SGI Octane workstation, the
larger simulation models ran too slowly because her workstation lacked
sufficient texture memory. These particular results required a more
expensive ($30K) SGI Octane MXI model. To view these results she
reserved time in the CAVE where she could view results of the larger
models. In the CAVE Dr. Farkas eliminated FCC atoms of the grain so
that she could view only the atoms along the grain boundaries. The
remaining atoms observed along the grain boundaries experienced a
stacking fault as the result of a partial dislocation emission. With
AtomView this stacking fault can be observed in an animation where an
atom color changes from gray to red
(http://www.cmsms.vt.edu/~diana/helena.html). This example and others
were published by Farkas et al. in six journal articles listed in the
'Publications and Products' section of this report. The CAVE was used
routinely by Dr. Farkas in her research on design of ceramic
nanostructures as indicated by her numerous publications.
On two occasions Dr. Farkas arranged for a visiting scientist, Dr.
Helena van Swingenhoven from the Paul Scherrer Institute in
Switzerland, to view these results in the CAVE at Virginia Tech. To
enhance the use of AtomView for collaboration, Dr. Kriz combined
AtomView with CCC and created CCC_atom that would allow Dr. Farkas and
Dr. Swingenhoven to collaborate with their SGI desktop computers.
However since this required the use of high end SGI Workstations, less
expensive PC Windows computers and tools were preferred.
In conclusion, while the immersive environment of AtomView in the
CAVE provided insight for the interpretation and analysis of
supercomputing simulation results, visualization software that runs on
less expensive PC Windows desktop computers was also used for
interpretation of simulation results. Now that the CAVE facility is
more accessible by faculty on-campus we anticipate increased use,
however the use of Windows PC computer tools will continue to be an
important visual tool. Unfortunately visualization tools used on PC
Windows computers do not connect with applications that run on the SGI
CAVE computer. Future funding is now available to rewrite the DIVERSE
API (Application Programming Interface) that will run on SGI and
future Windows operating system -- linking these two resources.
2.4.2 Biochemistry / NIST Nanoscale VE Technology (Drs. David Bevan
NSF-ARI CoPI and ASPIRES PI / Ron Kriz, NIST PI):
Activities on 'CAVE Visualization of Biomolecular Structures' were
already outlined in section 1.2.2. Results are only summarized,
because this research is in-progress. Two funding sources, VT-ASPIRES
and NIST, funded us to link the PHANToM force feedback device with the
Immersive-Docking (IDock) VE application. We now have a working
prototype that uses DIVERSE & VRPN to link the PHANToM device running
on a Windows Intel PC to two simulated spheres. Using DIVERSE, one
sphere is located in shared memory on a Linux Intel PC and the second
sphere is located in shared memory on an SGI deskside Power Onyx
running the IDesk. These results will be presented at the Virginia
Tech Workshop on Bioinformatics and Computational Biology, March 19,
This fall 2001 we expect to replace the spheres with actual
biochemistry molecules in IDock running on an SGI Power Onyx Rack in
the CAVE. This will complete our proposed research funding for
VT-ASPIRES. For NIST, our second year funding will be applied to
linking the PHANToM device to remote site nanotechnology devices used
to fabricate nanostructures and to a visual simulation of these
nanostructures running on an IDesk.
2.5 Future development of CVEs for research and education:
- Collaborative VE tools are rarely used because they run only
Collaborative VE tools such as Performer-based CCC are rarely used
becausePerformer runs only on IRIX and Linux and the underlying
graphics are proprietary and expensive. Until collaborative tools run
on desktop Intel computers running the Windows operating system, only
a small group of researchers will use tools such as CCC or CCC_atom.
The one exception was the K12 project with the Virginia Governors
School, but this only happened because SGI and ICRSC provided the
necessary SGI computers. Future development of GNU/Linux offers a
possible solution. Recently the GNU/Linux operating system, like
Windows, runs on Intel computers but is licensed LGPL (GNU 'Free'
software). With some effort in reprogramming, CCC and CCC_atom could
run on less expensive Intel computers.
Since cost is a factor for researchers like Dr. Farkas, it was not
surprising that educators are even more sensitive to costly software
APIs. CoPIs writing the K12 CCC proposal were reluctant to use VE
software that might break the budget. There are now a variety of
GNU-GPL/LGPL licensed VE software APIs. One excellent example is the
Visualization ToolKit (VTK) that runs on the CAVE-Lib desktop
simulator and DIVERSE. DIVERSE, like VTK and GNU/Linux, is licensed
LGPL which is not commercially supported software. We support the
CAVE-Libs because some organizations require that software be
commercially supported, but with GPL/LGPL we now have a choice.
Although GPL/LGPL is 'Free', this software is not commercially
supported, hence, programmers accept responsibility. If it doesn't
work, they fix it, and redistribute the improvement 'freely'. Given
enough time this 'free' model can produce some impressive software
like the GNU/Linux operating system that is very reliable even though
it is not commercially supported software.
Using GNU-GPL/LGPL software is not for the faint-of-heart if the
software is relatively new. GNU/Linux, however, is a mature product
that started over 10 years ago. New software like DIVERSE is just
starting out so we recommend DIVERSE only for experienced VE
programmers. With GNU/Linux and LGPL licensing of DIVERSE the VE
community can chose to become co-developers and hence co-owners of the
DIVERSE VE API. Presently we are being funded by ONR to extend the
current DIVERSE graphics interface to Performer (DgiPf) to OpenGL
(DgiGL) which then has the potential to run on Linux and the Windows
operating systems. When realized this fall 2001, this new VE API will
add new life to collaborative VE tools running on Intel desktop
- DIVERSE --- 'The Glue', but only for experienced VE programmers.
DIVERSE can run on Intel computers because SGI recently ported
Performer to GNU/Linux. Until DgiGL is completed this fall 2001,
'freely' available collaborative tools like CCC and CCC_atom will be
available much later. Because 'freely' implies not commercially
supported, the CCC user community must be experienced VE programmers.
Give this user community up to 10 years and perhaps CCC and CCC_atom
will be mature and ready to be used by the scientific and academic
community who are not experienced programmers. With this concept as
motivation, our group, the Center for Virtual Environments and
Visualization (CVEV), is committed to building the DIVERSE graphics
interface to OpenGL (DgiGL) which is funded and scheduled for beta
release this fall 2001. The Performer based DIVERSE 1.01 API is
currently available at http://www.diverse.vt.edu/ and we welcome
co-development and co-ownership.
With the current Performer based DIVERSE API we have been funded by
Lockheed Martin Astronautics Group to move the collaborative awareness
tools from CCC to DIVERSE. At that time we will re-explore future
funding for rewriting CCC_atom so that researchers like Dr. Farkas and
educators like the Virginia Governors School can avoid the high cost
presently associated with both hardware and software.
2.6 HCI Results (Drs. Jack Carroll and Deborah Hix):
HCI results on 'Usability Evaluation Techniques: A Novel Method for
Assessing the Usability of an Immersive Medical VE' have been
published in the conference proceedings of Virtual Worlds and
The creation of the CCC was also an HCI project that was published
as Mr. Kevin Curry's Masters thesis on the Virginia Tech Electronic
Thesis and Dissertation ETD) Web pages.
Training and Development:
Since one of the major objective was training, discussion on training
is embedded through out the 'Project Activities/Findings' section of
Outreach Activities is also included in the 'Project
Activities/Findings' section of this report.
Other Specific Products:
Brief Software Desciption:
1. Device Independent Virtual Environment: Reconfigurable, Scalable,
and Extensible (DIVERSE): A general SGI Performer based VE API
that was used develop VE applications in the CAVE, IDesk, IWB,
desktop IRIX & Linux workstations and laptops.
2. Just In a Virtual Environment (JIVE): JIVE is a VE application
built with the DIVERSE API that was developed for research in HCI
interface evaluation and usability.
3. Atomview: An SGI Performer based VE application used by material
science research to visualize and interpret supercomputer
simulations of nano-structure physics based models.
4. CAVE Collaborative Console (CCC): A general collaborative VE that
links desktop workstations, IDesk, IWBs, and CAVEs into a shared
working ('design') environment. CCC is an SGI Performer based
application, that was built on top of CAVERNsoft and Limbo in
collaboration with the NSF PACI project.
5. CAVE Collaborative Console Atomview: A Performer based VE
application that specifically linked CCC and Atomview to enhance
collaboration between material scientists.
Each software item listed above has a Web site where the software can
be downloaded and installed with tutorials and examples. These Web
sites are respectively:
1. Device Independent Virtual Environment: Reconfigurable, Scalable
and Extensible (DIVERSE): http://www.diverse.vt.edu
2. Just In a Virtual Environment (JIVE):
3. Atomview: http://www.sv.vt.edu/future/cave/software/atomview/
4. CAVE Collaborative Console (CCC):
5. CAVE Collaborative Console Atomview (CCC_atom):
This Web site provides links to all know CAVE related activities
both on- and off-campus
Contributions within Discipline:
- Virtual Environments (VEs) are inherently multi-disciplinary.
At Virginia Tech the acquisition of a CAVE has prompted interest from
many disciplines, because independent of the discipline, we all share
a common need which is to explore complex three-dimensional (3D)
structures. CAVE technology allows each discipline to explore 3D
structures in full immersion which can, and often does, lead to
insight in new structure property relationships. How each discipline
uses VEs to explore these structures is unique. How a CAVE virtual
environment can contribute to the base of knowledge, theory, and
research and pedagogical methods within that discipline is also
unique. Comments below will apply to all disciplines in general, not
how VEs benefit each discipline uniquely.
- Researchers/educators typically do not realize the full potential
of a CAVE application because they lack the necessary programming
The full potential of a CAVE VE is only realized when an expert in
that discipline creates content so that, within that context,
exploration and possible insight can occur. Building meaningful
content is a necessary, but not sufficient, condition for exploration
and insight in the CAVE. Within each discipline collecting and
organizing the content, although important, is not where the problem
is encountered. There are two major issues that can prevent the CAVE
application from being successful:
1) Most discipline experts cannot program in virtual environments,
2) Discipline experts who can program (they are rare) typically
cannot design usable interfaces.
The inability to program is perhaps the most common and limiting,
because if one cannot program and build a CAVE interface the second
point is academic. Typically discipline experts are professors,
however professors chose not to learn how to program applications in
the CAVE, even if training is offered (see 'Major Findings and
Results). Instead they defer to their students who eventually leave
and take their programming knowledge with them.
Creating teams is a solution, where professors with the programming
skills work with discipline experts who build meaningful content,
along with HCI and usability researchers. This would be 'the dream
team'. But these teams are rare and although collaboration is
discussed positively, it is not rewarded by academia in general. At
Virginia Tech we attempted to build such a team and targeted specific
application areas. Our three original target areas were: 1) ceramic
nanostructures, 2) biochemical structures, and 3) ceramic composites.
We realized limited success with 1) and 2) and are committed to build
on this success.
With funding from ONR MURI/DURIP/NAVCIITI another SV application was
identified: physics based simulation of a Navy crane-ship. Dr. Lance
Arsenault, Ph.D. in physics and experience in VE prototype programming
with Caterpillar at NCSA, worked with Professor Ali Nayfeh, the
discipline expert in the Department of Engineering Science and
Mechanics. They teamed with Mr. John Kelso, whose experience is in
graphics, HCI ,and usability in CAVE VEs. This particular project is
in progress and is scheduled to be completed this spring 2001 now that
the motion platform is built into the CAVE floor, see Figures 4 and 5
in prjact.pdf file.
- Shared VEs from desktop to CAVE and across heterogeneous
operating systems offer a new potential for distance learning
and collaborative research.
For all disciplines we believe the highest potential for success is
shared collaborative VEs that link participants across heterogeneous
operating systems (Windows, IRIX, Linux, MacOS-X) that scale from the
laptop, desktop, Head Mounted Displays (HMDs), to IDesks, IWBs, and
CAVEs. As discussed in the section on 'Major Findings and Results' we
have a plan to begin the creation of this 'diverse' environment.
Since our equipment acquistion project objectives were to apply CAVE
technology to all disciplines the topic of 'Other Disciplines' has
already been explained in the previous section on 'Contribution within
On two separate occasions we provided exposure to science and
technology for pre-college teachers, young people, from: 1) the
Central Virginia Governor's School at Lynchburg and the Central
Shenandoah Valley Regional Governor's School at Troutville, and 2) the
Maryland Virtual High School (MVHS) in Silver Spring, Maryland. In
both cases we organized a two day training workshop for the high
school instructors with the intent to 'teach the teachers' how to
develop CAVE content using VRML desktop 'free' software and then how
to load their content into the CCC. With the two Virginia Governor's
schools we were funded by Silicon Graphics Inc and the Institute for
Connecting Science Research to the Classroom (ICSRC) to buy the
necessary equipment to develop K12 curriculum based on CCC. Results of
this project, 'Extending the Use of Collaborative Virtual Environments
for Instruction to K-12 Schools', have been summarized at
http://www.cvgs.k12.va.us/ccc/. For the MVHS the same VRML and the
Collaborative CAVE Console training session was incorporated into the
Academic Global Excellence Summer 2000 program. In both cases the
'teach the teachers' training was extended for a select group of
students from the two high schools. Training and results are posted
on the Web at
The CAVE acquisition is part of the University Visualization and
Animation Group (UVAG) of the Advanced Communication and Information
Technology Center (ACITC) at Virginia Tech. The ACITC is now located
in a new building at the center of the Virginia Tech Campus,
http://www.acitc.vt.edu. see Figure 1 in the prjact.pdf file. The
ACITC is a joint state and private funded building that is a regional
resource for research and education that is part of our reach mission
as a Virginia land grant university. The UVAG is one of several
information technology activities: Digital Libraries, Human Computer
Interaction, High Performance computing, Institute for Distance
Learning, etc. are all located in the ACITC, see
http://www.sv.vt.edu/future/future.html#ACIT-UnivVizAnimGrp and also
When the CAVE first opened in February 1997 at the Virginia Tech
Corporate Research Center (CRC), a new privately owned company,
Virtual Prototyping and Simulation Technology Inc., not affiliated
with Virginia Tech, established offices next to the UVAG offices in
the same building. The CRC site, as previously mentioned, was a
temporary site until completion of the ACITC building.
During the three years at the CRC, the UVAG worked with VPST on a
variety of projects. UVAG and VPST collaborated on a proposal funded
by Virginia Center for Innovative Technology on 'Infrastructure
Development and Planning Project to Explore the Benefits of a
Collaborative Virtual Environment in Virginia Universities and
VPST served as the university industrial affiliates program and
coordinated industrial access to the CAVE UVAG facility. VPST was
closely affiliated with a private company in Sweden called Prosolvia,
which went out of business in 1999. VPST also went out of business
shortly after Prosolvia went out of business.
Prior to working with VPST, Dr. Ron Kriz continues to work with other
computer software companies, i.e. Visual Numerics Inc. (formerly IMSL)
based in Houston, Texas. Visual Numerics continues to support our Java
web technology projects and is a co-sponsor with Sun Microsystems of
the Scientific Modeling and Visualization Classroom which is part of
our UVAG/CAVE facility. Presently our industrial affiliates program
exists by working with the university development officers, who are
identifying collaborations with Virginia based industries interested
in collaborative virtual environments. The DIVERSE API funded by NIST
(U.S. Department of Commerce) has the potential to be an e-commerce
tool with the GNU-GPL/LGPL licensing philosophy previously described
in the project results section.