This report is also available as an Acrobat file.

Contents The Design of Virtual Environments with particular reference to VRML

Design into production

The Future of VRML

The current VRML standard is one that concentrates on static objects and scenes which users can move around. As this report is being written the debate and process for the specification of VRML-2 is being undertaken, managed by the VRML Architecture Group (VAG). A number of mainly commercial groups are developing software for interactive social worlds as a test bed, elements of which they hope will contribute to the future standard. The future is in two areas, which are increasingly related.

Dynamic Models

VRML-2, also known as Moving Worlds, provides, as its name suggests, for the introduction of action in the modelled world. Either objects in the world can act and react to each other under program control, or they can respond to the user's actions in some way. For educational purposes, this will in many cases be essential. Byrne (1996) in a controlled experiment found that immersive VR seemed to offer little advantage over a Window-on-the-World (WoW) approach, but interactivity perhaps did. This was specifically in an example where the technology was used to support education. She sought many reasons to explain the data, but the answer may be that immersion may not be required to achieve most of the benefits to education. The results seemed to suggest that it was the interactivity that was important. This corresponds to a constructivist theory of learning, in which the student builds and manipulates problem spaces, rather than being merely an observer.

Social Worlds

One of the most important aspects of the Internet is its ability to facilitate interpersonal communication. IRC and MOOs are examples of such facilities that are very popular. These may prove exemplary for the development of applications for VRML. The history of the Internet is largely one of the victory of chatter over content, but like the telephone the social worlds will be very important. In the HE sector, it will remain the case that much learning is achieved through student-student and staff-student communication: social worlds can facilitate this.

We will look at these two developments, Moving Worlds and Social Spaces, in turn.

Moving Worlds

Moving Worlds extends VRML in various important ways. The VAG (http://vag.vrml.org) proposes (May 1996) that the Moving Worlds specification goes through a series of drafts with a release date of August 4th, timed to coincide with SIGGRAPH 96. However, Silicon Graphics (http://vrml.sgi.com) has already released a beta version of a browser, Cosmo, that implements many of the features of Moving Worlds (VRML 2.0). Clearly, from this response, it is almost certain that a fully compliant browser will be available very soon after the standard has been fixed. For many readers, therefore, it is probably worth jumping over VRML version 1, straight to version 2, the gain in functionality outweighing the risks of working to forthcoming standards.

The features that Moving Worlds currently includes, most of which are likely to remain, are:

International character sets for text can be displayed using UTF-8 encoding.
A set of new nodes has been added to increase the realism' in models that are intended to represent the outdoor world around. It will be possible to create ground and sky backdrops, adding distant mountains and clouds, for instance. The effect of distance can be further enhanced with fog effects. In addition, there will also be the option of defining an irregular terrain rather than being constrained to a flat ground plane.
Sound generating nodes will also enhance the sense of realism. Thus, doors can creak, phones ring and dogs bark. Intel has announced Realistic Sound Experience (http:// www.intel.com/ial/rsx) which will allow spatialised sound to be put into VRML worlds, such as the DirectedSound and pointSound nodes in Moving Worlds.
New sensor nodes will set off certain events when one enters specific areas, or click on certain objects. So, for example, as you approach an object it can be triggered to start some action or make a noise.
Collision detection ensures that objects can act as if solid. That is, the user, or other object will not go through walls and floors.
Script nodes allow for the animation of objects in the world.
Multi-user environments. There are many approaches to creating multi-user worlds, and Moving Worlds aims to provide the functionality required for these, but without dictating which approach is to be used.

Moving Worlds is very extensible, and fully-integrated Java applets can be used to create motion and enable interactivity: there is a small tutorial on what can be achieved at http://www.javaworld.com/javaworld/jw-03-1996/jw-03-javavrml.html. JavaScript allows scripted communication between objects both within and without the virtual world. It is this aspect of the Moving Worlds definition, however, which is still uncertain at the time of writing. Furthermore, Plug-ins within the 3D worlds will facilitate such services as streaming audio and video, and slide presentation.

Social Worlds using VRML

There is a series of worlds being made available, experimenting with increasing functionality, mostly based on the VRML format. They include either scripting and construction facilities, or additional communication, be it textual, audio or even video. Most of these take VRML models and allow them to be used within multi-user shared environments. That is, users from anywhere in the world can connect to these, inhabiting them embodied in an avatar (a 3D character, also usually modelled in VRML). Users can then meet' other users and chat. In VRML-2 the users can also manipulate the world. A list of such shared worlds, or habitats as they are often known, is currently kept by Electric Communities at http://www.communities.com/habitat.html. For the current discussion, some of the most important ones are:

CyberGate

Produced by Black Sun Interactive (http://www2.blacksun.com/about/index. html). This is one of the most popular of the new shared worlds. You can create a VRML model on your own server, which can then be made accessible to others who have a CyberGate browser, and become a shared space.

OnLive

Unlike most others this allows for voice communication between participants (http://www.onlive.com/utopiamkt/index.htm). These projects indicate the kind of features that are now coming available, some of which will make the next version of the standard. In addition to multi-user access, we have interaction and animation. The ability for the world to carry both audio and video is also shown in some, such as InterSpace by NTT.

The importance of these worlds is indicated by the move of some of the largest players in the computer field into the area, such as: Microsoft with V-Chat (http://www.msn.com/v-chat/index.htm) Sony's CyberPassage (http://vs.sony.co.jp/VS-E/works/server/swelcome.html) IBM's Virtual World (http://www.software.ibm.com/software/virworld.html) and Intel's Moondo (http://www.intel.com/iaweb/moondo/index.htm).

There are some popular habitats, such as Worlds Chat and AlphaWorld created by Worlds Inc., that are not true VRML. But they too have recently announced that their next technology, Gamma, will be VRML-2 compliant.

It is our belief that these shared worlds, based on VRML, will be very important in the future. One lesson from history is that the communication aspects of any technology become important and are often overlooked in the early days (the telephone was originally seen as a good way of listening to opera without visiting the theatre). The ability for students to not only visit another VRML to explore it, but also to possibly meet' its authors is very exciting.

Many current uses of VRML-1 are simple at best, due to the constraints described above. In many cases they are existing models created for other reasons, and simply converted to VRML. Proper applications that are designed to fully use VRML are still in short supply, and there is a definite sense that everyone is awaiting VRML-2 to be able to create models that have some real value. Pointers to some of these existing applications can be found at the VRML Repository.

Contents

Graphics Multimedia Virtual Environments Visualisation Contents