In connection with the Fourth European Conference on Computer-Supported Collaborative Work an invitation was sent out for a workshop to discuss collaboration in virtual environments.
The physical part of the workshop was held in the West seminar room of the Interaction and Presentation Laboratory of the Department of Numerical Analysis and Computing Science at the Royal Institute of Technology in Stockholm.
Steven Benford was chosen chairperson and started by welcoming everybody. Introductions were made round the table.
SDB presented a number of problems that should be explored further, based on the experience of using MASSIVE, DIVE, dVS, SuperScape and WorldToolKit.
SDB defines the term "virtual environments" as meaning "navigable spaces", thus excluding e g bulletin board systems and the like.
He felt that the main issues that should be explored are
DE delivered a critique of the current approaches to interaction in virtual environments. He proposed using a different set of conventions in opposition to the usual Benediktine ones. Virtual environments should be interpreted more as information spaces than real spaces and this entails encouraging multiple occupancy of space, teleportation, invisibility of participants and subjective displays of the environment. BE then showed a few slides of the BSCW (Basic Support for Collaborative Work) system developed at GMD.
GF described the wOrlds system developed at the University of Queensland and showed a few slides of it. I was apparently partly absent during this presentation since I have very few notes of it. I did get the impression that wOrlds is basically a 2D system. I also have noted a question whether wOrlds is very application dependent or if it can be used as a generic tool.
KMJA (that's me) presented an idea for how to support moving embodiments in virtual spaces by (attempting) keeping track of which object a user is accessing each particular moment. SDB suggested that this really isn't sufficient, as other users may not only want to see which object a user is working with, but what other objects are in her field of view. KMJA tried to regain initiative by suggesting putting beacons on objects which can be flashed should we need to know what objects a user can see.
MR talked about peephole surgery and explained how putting the endoscope view on a video screen enabled other members of the operating team to get a better feel for what was going on and work in better synchrony. Even anæsthesiologists could correlate their ECG/BP/etc graphs with the video image. This video image is an abstract data space, according to MR.
From there he went on to the International Classification of Diseases, a huge WHO database of diseases, causes of deaths and other things. This in itself was a challenge to try to visualise in some way but the ICD does not exist in isolation - in fact much of the information in it is embedded in the minds and relations of groups of experts around the world. Is there a way to go from browsing the raw data to access the users and generators of it?
NS described the OCEAN and DOLPHIN projects at GMD, Darmstadt. Yet again I must have been out of the room part of the time as I only have that the presentation concerned virtual meetings with audio/video and whiteboard support and how 3D presentation could be used to simplify the representation of the various connections between participants.
Videos were shown of conferencing in MASSIVE, the DIVEDIT 3D modeller and some other things which I did not write down.
After everyone had done their presentation SDB called for a collection of the issues we wanted to discuss. I did not remember to write all these down, but judging from the rest of the discussion, subjectivity, teleportation and embodiments were included among these.
At the beginning of the discussion NS asked "Why is something called 'virtual'?" Virtual Reality seems to be defined as 3D simulations of real and abstract objects, but "virtual" could also be defined as people in different places gathered in a meeting system.
Does a virtual environment have to be immersive or not?
GF suggested that "virtual" may simply mean "non-physical".
KMJA wanted to make a distinction between "immersion" and "involvement", in that immersion gives the ability to look and move around in some space, i e a technical distinction, whereas "involvement" is the psychological phenomenon of being (to some degree) unaware of the physical world outside the virtual environment. It seemed the others did not agree.
MR reminded the company of Lynch's expression "as real as it gets". Contemporary technology is as real as it gets and systems have to be built on that basis.
Alfred Kobsa pointed out the importance of the acoustics and the visuals agreeing and having a high refresh rate to give the feeling of immersion.
MR noted that "immersion" is oftenmost interpreted as a technological phenomenon, which may be a mistake. There are also problems with immersion, such as being shut out from the physical environment where still many important things happen.
SDB wanted to make a distinction between places, that can be meeting rooms, and spaces, which are navigable in some sense. David Mitchell noted that London is then a set of places, accessible through the underground. Some discussion ensued on this.
Gunilla Ivefors pointed out that terms are used differently in different contexts and that we should accept that there is not one definition which will in all situations.
Can we translate between languages and views?
SDB defined a "virtual environment" as a set of data, subject to transformations and represented as views. Other users can affect the underlying data and that way affect your view - this necessitates awareness of the other users so that the user can figure out what happened. This in turn requires some kind of representation of the users and at one extreme, this can be in form of an embodiment (if there is a sufficiently common view of the data).
DM noted that some MUDs support multiple representations of a participant. SDB wonders if they count as embodiments? MR initiated a discussion on body language.
NS put in a question on the value added by a MASSIVE type conference as comapred to an ordinary vide/audio conference. The reply seemed to be that the participants can easier keep track of each other and it's easy and seamless to split off into private discussions and then rejoin the main group.
Should teleportation be instant? DE noted that the London underground is not instant but lets users rapidly move between places of interest.
SDB made a number of counterarguments to, or rather suggested restrictions on, teleportation, such as it should not be possible to teleport into somebody's personal space or even that there should be places where it is impossible to teleport into. He also thought the speed of teleportation should be limited. Most others thought it would be better to be able to teleport quickly. SDB counterargues that it disturbs the predictability of the environment.
There was some discussion on the subject of personal and secret spaces. Are these symmetric, ie if one cannot teleport into them, should the owners be allowed to teleport out? SDB also thought that since we will have to expect advertising in virtual spaces, advertisers may demand that users not be able to teleport past their virtual billboards.
A discussion on views and hypertext, overviews and paths ensued. NS noted that "teleporting" really is a bad term. GF generalised it and noted that telephones "teleport your voice".
As a last item Adrian Clark made a few comments about the utility of video workshopping and announced his comments on the experience.