CVE'96, the first workshop dedicated to research invesitigating the design and use of Collaborative Virtual Environments (CVEs), took place on the 19-20th September 1996 at the University of Nottingham, UK.
The intention was that the workshop would present the current state of the art in Collaborative Virtual Environments and foster inter-disciplinary links between researchers in this field. Suggested topics included: techniques for supporting communication and collaboration in virtual environments, metaphors for collaboration, experiences and evaluation of CVEs and CVE-based applications, systems architectures for CVEs, applications of CVEs, collaborative DIS and VRML applications, studies of social action and interaction in CVEs, and user centred development of CVEs.
Over 90 people attended CVE'96 with speakers from 11 different countries, 16 paper presentations and 7 poster presentations. A mailing list dedicated to CVEs is currently hosted at Nottingham.
Mathew Chalmers (UBILAB, Switzerland) presented work on collaborative information visualization. He discussed some of his recent work on BEAD to add legibility features and support multiple users. The talk also highlighted some of the issues and challenges currently facing information visualization such as the problems of artifacts occluding other artifacts and the need to reduce clutter and reveal structure.
"Just because something is on the screen, does not mean its visible"
Chris Hand (De Montfort University, UK) highlighted some problems with the CRC Cards technqiue for evaluating software designs. He then presented CRCMOO a MOO designed to support a group of people performing a CRC Cards design evaluation.
POLITeam is a groupware system which allows German government administrators to cooperate with each other remotely. David England (GMD, Germany) described how DIVE (together with some Tcl scripts) and VRML was used to generate "virtual offices" contain representations of the documents available via POLITeam.
Jon Mandeville (HIT lab, USA) was unable to attend CVE'96, so Mark Billinghurst (HIT lab, USA) presented this paper which focussed on the Greenspace project at the HIT lab. Greenspace 1 was a virtual environment with fours users, two in Seattle and two in Tokyo. The aim was to collaboratively herd a crowd of creatures. Real-time audio communication was provided and video of the users head was textured onto their representation in the VE. However, Greenspace 1 was based on specially written software rather than a general purpose environment. The aim of Greenspace 2 is to produce a general purpose virtual environment that can be used for many applications. The second part of the talk focussed on the design of Greenspace 2 and demonstrated it being used to conduct multi-user architectural design reviews.
Mark Billinghurst (HIT lab, USA) discussed the use of Collaborative Augmented Reality; such systems represent a shift from human-computer interaction (HCI) to human-human interaction mediated by computers. This work takes the view of Ishii & Miyake that "No new piece of technology should block the use of existing tools and methods". Shared space is an augmented reality system that allows multiple users to collaborate on the same task. Mark presented several games using Shared Space which were designed to investigate how effective Shared Space was at supporting collaboration and interaction between two users.
Dieter Schmalstieg (Vienna University of Technology, Austria) presented "Studierstube" an augmented reality for performing science related collaboration. The system is designed to suppor people working together and does not affect normal interaction. Because the system depends on augmenting a real environment it is not possible to collaborate with geographically remote users. Dieter desribed an application using "Studierstube" for scientific visualisation. He also described a novel interaction device, the Personal Interaction Panel (PIP) - a tracked flat panel together with a tracked pen. Virtual objects could be overlaid on the panel. Users can then move the panel to manipulate the object. Another use for PIP is for handling menus, users can use the pen to select menu items "attached" to the panel. Since the pen and panel are physical objects the user gets tactile feedback when the pen touches the panel to make a selection.
Pierre Nugues (ISMRA, France) presented Ulysse a conversational agent embedded within the SICS DIVE virtual environment. Ulysse aids the user in navigating through the virtual environment by recognising spoken commands such as "move me to the red house". When moving the user Ulysse attempts to navigate around obstacles and detect collisions between the user's embodiment and artifacts in the Virtual Environment. Ulysse can also be told to manipulate artifacts in the virtual environment.
Wolfgang Broll discussed his extensions to VRML to support multiple users. His approach uses a combination of and tcp/ip to distribute VRML worlds and multicast to transmit event notifications. This approach divides artifact behaviour into several catagories each of which is supported by different mechnanisms to provide the best compromise between consistency and interactivity.
UnJae Sung (KAIST, Korea) described several VR applications developed at KAIST: Landscape a 3D VRML/wen browser, Sarah an architectural walkthrough with a Tcl/Tk API and VRAT a collaborative VE authoring tool. The main focus of the talk was VRAT and the concurrency contro, model used to enable several users to manipulate the same artifact simultaneously.
Uwe Rauschenbach (University of Rostock, Germany) presented some techniques for notifiying users of the Linkworks(tm) groupware system of events occuring as the result of actions of other users. Several different techniques are used to inform users of events depending on the relevance to the event to the user and the user's current activity. A single event may generate multiple (different) notifications to several users and the user can interact with the event to obtain more information.
Annita Fjuk (University of Oslo, Norway) & Elsebeth Korsgaard Sorensen (University of Aalborg, Denmark) gave a talk discussing the complexity of designing distributed collaborative learning. They presented three case studies which used different media, differing numbers of participants and different subject areas and discussed the experience gained from these case studies. They then discussed how drama could be used as a metaphor to aid the design of collaborative learning experiences.
Mirco Bolzoni (University of Padova, Italy) talked about the use of metaphors in collaborative virtual environments. Metaphors have the potential to bridge the gap between abstract and concreate information. He presented the results of a study which asked both users and designers to produce metaphors for interacting in a VE. The users produced 70% of the metaphors which may have been because of their need to interpret what the software was doing. A problem with the use of metaphors is that the attribution of meaning is culturally and socially mediated.
Jerry Bowskill & Doug Traill (BT, UK) presented an overview of some work virtual environments work that BT is involved with and described experiences with BT's VisionDome (currently installed at BT Laboratories, Martlesham Heath).
Tomas Axling argued that most work in a collaborative virtual environment involves creating or changing some artifacts according to some rules. He discussed his work on a general purpose configuration engine called OBELICS. He then gave an example application which used OBELICS to create an agent for configuring journeys which would find the best choice of flight, location, hotel etc depending on constraints imposed by the user. This application used SICS DIVE to provide the virtual environment within which the agent represented its user interface. Tomas also described some other speech controlled agents that he has developed which have various specialist skills and can be commanded by the user to manipulate artifacts in the virtual environment.
Helen Walker (University of Bath, UK) presented the results of some research investigating how the use of CVEs might be shaped by and impact on an organisation. As part of this work Helen conducted 97 interviews with people from 14 sites in 7 countries. The work highlighted the need for CVEs to take into account the way people work and the structure of the user organisation.
Michael Rygol (Division, UK) presented a description of Divsions dVS/dVISE software. He also discussed the experience gained by Division after conducting several multi-site trials using ISDN and the internet. They found that while ISDN gave acceptable performance, the internet was not currently practical for using dVS between sites. Michael also mentioned some of the technical fixes made to dVS to improve performance over WANs and showed some video clips from the trials.
Anna Cicognani's (University of Sydney, Australia) poster dealt with text based virtual communities based on MOOs. She argued that MOO languages can be developed to support design in a MOO environment.
Alistair Duke's (BT, UK) poster presented some work performed as part of the VirtuOsi project. This work produced a system, AP-IA (Availability of Person - Intelligent Agent) which helps users to find out whether a person they wish to communicate with is available or busy and therefore whether they are likely to be able or want to commnicate with you.
Chris Greenhalgh's poster described a persistent user agent that would remain in the MASSIVE virtual environment after the user has left and act as a "virtual proxy" which can record asychronous messages for the user or attempt to contact the user and "bring her back" into the virtual environment to engage in synchronous communication.
Paul Sharkey's (Reading University, UK) poster described work performed at Reading to remove the abrupt state transitions in CVEs due to commnication delays.
A last minute poster was submitted by Helmuth Trefftz (EAFIT University, Colombia). In the poster Helmuth describes the main objectives of the CONEXIONES project (funded by COLCIENCIAS (Colombian institute for the development of Science and Technology), fundacion CORONA (private enterprise), Universidad Pontificia Bolivariana and EAFIT (private universities)). Broadly, the project aims to improve the quality of primary education in Colombia by providing access to school multimedia applications to any educational institution in the area. The project aims to foster international links between schools and, by keeping the cost to institutions low, to provide equitable access to these new technologies. In realising these aims, the current main project involves the development of multimedia applications and of new pedagogical methods and evaluation procedures according to the technologies involved. Currently more than 100 primary schools in the area take part of the project.
Generally, the attendees were very positive about CVE'96 - in fact feedback was so postive that although CVE'96 had initially been planned as a `one off' event there are currently plans to stage CVE'98.
More specific comments were:
And one final interesting suggestion from the feedback questionnaire: "How about a virtual conference?"
"Hmmmm......Ever tried virtual beer?""
We would like to thank a number of people who helped out at the conference. Firstly, Steve Benford and CRG for underwriting the costs of the conference. Also, the Department of Psychology at the University of Nottingham for allowing us to host the event within the department.
Many thanks are due to the programme committee for their hard work in reviewing the submissions for CVE'96. Thanks to Jolanda Tromp for producing the feedback questionaires and facilitating contacts between Psychology and Computer Science.