To:     Rick Habayeh

From:   Ron Kriz

Date:   November 23, 1999
        Modified Nov 29, 1999
        Expanded Dec 12, 1999

Re:     360 DAC Deliverables

-----------------------------------------------------------------------

Data Item No. 0032

SOW 1.6.2

1) Description of features of scalable Virtual Environment (VE) 
   interfaces

A prototype system that would allow participants at remote sites 
to collaborate in immersive environments was created and tested.
The prototype system is called the "Collaborative CAVE Console"
(CCC) which is a software package that can be downloaded and 
installed on any SGI desktop workstation computer running the
CAVE simulator or a SGI supercomputer running an CAVE. An 
installation description is provided off of the CCC Web pages:

CCC download:
   
 http://www.sv.vt.edu/future/cave/software/ccc/disclaimer.html

Installation description:
    
 http://www.sv.vt.edu/future/cave/software/ccc/install/install.html

Since the software runs on both a supercomputer running a CAVE and 
a desktop SGI O2 or Octane workstation, CCC can link users at their 
desktops to users in the CAVE which creates a shared collaborative
environment. The CCC was based on CAVERNsoft and LIMBO that was 
originally used by NRL in their Battlefield Visualization Project.
CAVERNsoft and LIMBO are a product of the Electronic Visualization
Laboratory (EVL) at the University of Illinois at Chicago.  EVL
created the first CAVE and supports the international CAVE Research 
Network (CAVERN) and also created LIMBO, which is 

Summary of the Dragon "Battlefield Visualization" project by EVL:

  http://www.evl.uic.edu/cavern/users/nrl.html

NRL's newest combat VE system is called "Interoperable Virtual 
Reality (VR) System" and they have switched from LIMBO to Panda, 
which is similar to LIMBO, but is based on Bamboo software which 
allows for collaboration across differenct computing systems with 
multi-input and -output VR systems.  Because both LIMBO and Panda 
are written in SGI's Performer OOP (Objective Oriented Program) 
software, extensiblility will allow our collaborative awarness tools 
to be used in Panda.

This system was demonstrated at the NCSA (National Computational
Science Alliance) Access Center at Arlington Virginia on November 17, 
1999 as part of the Supercomputing99 Conference demonstration of 
collaboration over highspeed networks.  We demonstrated CCC on four 
workstations, each running CCC independently were linked using a 
standard T1 network to remote sites at Lynchburg, Shenandoah, and 
Arlington, Virginia.  At the NCSA Access Center the CCC was running 
on a SGI Octane desktop workstation and deskside Power Onyx which 
projected images onto an I-Desk ("one-walled CAVE").  Although voices 
could be linked via the internet with CCC, in this demonstration 
voices were linked using a standard telephone communication 
conference call.  This same system was also demonstrated on September 
22, 1999 at the Hart Congressional Office Building in Washington D.C. 
where two SGI Octane desktop workstations, located at the Hart Office 
Bldg., were linked to the CAVE computer at Virginia Tech. We plan on 
also demonstrating the CCC also at the Inernet2 Studio in Richmond, 
Virginia in the near future.

Scaling the CCC to the "large scale CAVE" referred to in the 
original task description was not possible, because the large scale 
CAVE structure was not built.  The extension of CCC to the General 
Dynamics "Command Post of the Future" was also not realized because 
this project was not funded.

1)Design requirements for high speed distributed collaborative
immersive VEs, and 2) downloadable prototype of collaborative 
immersive tools used between remote sites are posted on the web 
page:

        http://www.sv.vt.edu/future/cave/software/ccc/

A Masters Thesis by Kevin Curry on "Supporting Collaborative 
Awareness in Tele-immersion" for this can be downloaded from 
Virginia Tech's Electronic Thesis Dissertation (ETD) web page.

  http://scholar.lib.vt.edu/theses/available/etd-072099-120203/

Test and evaluation of prototype usefulness with recommendations
for full scale system is summarized in Kevin Curry's Masters Thesis
and on-going.  We are using researchers and educators both on- and 
off-campus to evaluate the usablility of CCC.  As a result several 
versions of CCC have been created, each with improvements requested
by CCC users.  The most significant finding was that users at desk
top workstations prefer a pull down menus interace over a voice
recognition system.  Although more inefficent CCC users prefer
pull down menus because of familarity.  Other features such as 
"shared view", where a remote-site CCC users can attach to another 
CCC users head, was also popular by CCC users, because desktop 
workstation CCC users, who are not in an immersive VE find the CCC
more usable with "shared view".  Users in the CAVE prefer the voice
recognition interface.

----------------------------------------------------------------------