The mission of Learning Sites is to take full advantage of the domains of VRML: virtual reality (based in education), interactive networked databases, and the information superhighway. For virtual reality (VR) to have an impact on public education, the content used to drive the lesson plans, the pupils' experiences, and the interaction with computers in the classroom should be based on actualities.

Learning Sites designs and develops educational and research software using interactive, three-dimensional, digital models that are both based on actual archaeological evidence and reconstructed with an accuracy and sensitivity to known details that reflect the highest standards of scholarship. We integrate archaeological data with advanced computer graphics techniques to further education, data analysis, and the preservation of cultural heritage information. We aim to create a globally integrated and interactive network of linked virtual worlds that can be used for teaching, research, archaeological fieldwork, museum exhibitions, and perhaps even tourism.

The dynamic, three-dimensional, digital databases that we develop are designed to promote awareness of past civilizations, understanding of different cultures, and appreciation of different places, peoples, and cultural heritages. Learning Sites is committed to producing serious educational and research tools as we work toward virtual reality-based digital schooling and archaeology for the 21st century.

A Sample Learning Sites VRML-Based Education Package
The selections included here are from the virtual worlds representing the ancient Greek farmhouse at Vari (c.325-275 BCE, from the Hellenistic period) that once belonged to a family of bee-keepers. The complete virtual environment incorporates multimedia interaction and demonstrates the capabilities of hyperlinked worlds. We can demonstrate not only the site both as it was excavated and as it has been reconstructed but also high-resolution models of some of the artifacts that were found in conjunction with the house.

The database, specifically geared for public school education, provides background information about the site’s structures, excavation, dating, and function. Since Learning Sites' databases are independent of the virtual worlds at the program level, the content of the worlds can be tailored to suit the needs of specific constituencies, such as K-12 pupils, graduate students, scholars, or the general public. Eventually, our worlds will link to different scholars’ reconstructions of the site, different sites with related characteristics, and views of same site at different times as the settlement around it changed in size and character. VRML 2.0 is making a difference to those who build 3D environments for the benefit of distance education

VRML
To meet the challenges of positively influencing courses of public education, interactive networked virtual worlds must be available on platforms affordable to schools and museums; the relative stabilization of VRML has given us an enabling technology toward this end. The VRML 2.0 standard is making a significant difference to those who build 3D models within multimedia environments on systems both within the financial range of our target audiences and for the benefit of distance education. Now the full power of 3D environments can be combined with the full power of multimedia to create an unrivaled learning experience.

Interactive virtual environments permit self-guided explorations of 3D information (such as our reconstructions of an archaeological site) with dynamic interaction between the user and the environment. Objects can move, change, and react to the user. Sound can provide ambient context and additional information through narration. Limitations of previous virtual environments prevent wide application for education or research

Previous virtual reality environments emphasized real-time response and self-directed movement, focusing on the immersive qualities of the experience. Limitations of these environments prevent wide application of such worlds for education or research: (1) expensive hardware is required to render the complex, highly textured 3D scenes in real time; (2) text display is rudimentary, usually limited to bit maps or groups of 3D letter objects; and (3) there are no facilities for allowing users to scroll or browse text or pictures within these 3D environments.

VRML has changed all that, making possible the integration of VRML 3D data, standard HTML 2D text, pictures, and video in a World Wide Web page. Advanced features like frames and multiple windows permit simultaneous browsing of 3D and 2D information. 3D environments can now be used for what they do best--allow users to gain a full understanding of a spatial structure through self-directed exploration while retaining all the power and detail provided by the 2D text and graphics of standard Web pages.

Multimedia designers can use VRML and HTML together to build fully immersive worlds or a multiply connected information resource where the 2D data and 3D data combine to make a powerful whole; an example of this might be a 3D building model which serves as the index to construction drawings. The links between 2D and 3D data make it possible to click on a specific detail in the 3D environment to not only hyperlink to a different 3D model but also bring up supporting text and/or pictures in a separate frame in the either same or a different window. "Hot spots" in the text can affect the 3D portion of the screen, and "hot objects" in the 3D environment can access and change the text and graphics in the Web page.

Some Problems with VRML
At its best, VRML permits the integration of 3D data into a multimedia environment. VRML has strengths that complement other multimedia elements, while those other elements can provide details and framework for information that is difficult to implement in 3D. Unfortunately, the situation is not totally ideal. There are VRML browsers for most common platforms (Macintosh, Windows NT, Windows 95, Windows 3.1, SGI Irix, Sun), but the browsers are not all consistent with the final VRML 2.0 specification, nor with each other. The following are some of the problems we encountered:

1. There is no way yet to get standard colors in a VRML world. Different VRML browsers will display the same object with different colors and different shading. Any application that depends on accurate rendering of an object's colors should consider supplying supplemental 2D images. (Web-page designers know this issue too well, as not even 2D images display consistently because of differing color maps on different platforms.)

2. Building a VRML world that has complex user interaction requires small snippets of code; however, the VRML spec does not state which language should be used. Browsers currently support Java, JavaScript, VRMLScript, and other languages. Worlds written with a given language will run only on browsers that support that language. Worlds that must run on many browsers will be limited to only those basic actions provided by the VRML specification. VRML developers must design user interaction carefully so that the world will not crash if only partially loaded

3. Unlike previous VR rendering systems, most VRML browsers let the user interact with the world as it loads. This mitigates some of the inherent delays in loading a large world over a network. The situation also requires VRML world developers to design user interaction carefully so that the world will not crash if triggered while the world has only partially loaded.

4. VRML modelers have a rapidly increasing number of tools for creating VRML models and interactive worlds, but the output from these tools varies greatly in compatibility with the final VRML 2.0 specification. Further, important features, such as polygon reduction and support for VRML nodes (level-of-detail and proximity detection, for example) are available on only a few tools. The biggest problem with VRML 2.0 is its newness

The biggest problem with VRML 2.0 is its newness; browsers and coding tools are not completely up to the final specification. All sorts of incompatibilities result, making the developer's job vastly entertaining. The problems should be resolved soon, as the various suppliers publish final 2.0-compliant versions of their product and close current holes in the specification.

Our current worlds are best viewed on a PC running Netscape Navigator 3.0 and the Cosmo VRML Player 1.0 beta2a (developed by SGI). HTML pages provide the curriculum framework. Features, such as frames, anchored viewpoints, and labeled locations, allow us to create multiple self-guided explorations of the site, each emphasizing different aspects of the data, from the daily life of the farmer to archaeological methodologies. We have designed the HTML and VRML source files to allow the user to explore freely, while ensuring that the information provides a coherent learning experience.

The Future of Education
Our primary audiences are schools (especially, but not limited to, grades 5 through 12), museums (especially for exhibitions, publications, and education), and archaeologists themselves (especially for publication, data analysis, and alternative visualizations). Interactive, multi-user virtual worlds, delivered over the Internet or on CDs, place the coming electronic revolution in public education nearly on each building's doorstep. Our three-dimensional, digital, archaeology libraries provide virtual environments for the place, time, or culture being studied. Subjects such as history, geography, archaeology, and astronomy can be taught using virtual worlds based on real-life situations and settings.

For future classrooms, interactive, networked, virtual reality-based education can enhance students' learning in a variety of ways, for example:

1. by offering vicarious firsthand experiences otherwise beyond their reach or their school's ability to provide,
2. by providing interaction with geographically or temporally remote locations, people or objects, and
3. by providing information at levels of detail tailored to individual needs.

Schools will find the computer-linked future classroom financially efficient as well. With access to online teaching materials, each school system need not compile complete sets of instructional materials or full libraries, in the traditional sense; with worldwide connections, a virtual environment can link students voice-to-voice and eye-to-eye with the best instructors and teaching aids in the world. These are tremendous opportunities, and although the promise of educational reform and revolution have in the past been overstated and oversimplified, we believe that judicious use of networked, virtual reality technologies can indeed make a difference.

To ensure that our materials are not created in a vacuum, we:

1. consider the guidelines set by professional archaeological associations for teaching archaeology in public schools;
2. ensure that our worlds meet the stipulations for educational software set by the Virtual Reality and Education Laboratory (East Carolina University); and
3. seek out the advice and guidance of teachers, educators, and administrators so that we can tailor our instructional materials to specific local or state curriculum objectives.

Conclusion
Putting all this into the big picture, we envision that a widely encompassing circle will develop. Digital documentation methods will allow archaeological fieldwork to be entered directly into bi-directional databases as excavation progresses. Field archaeologists will be able to access virtual worlds based on this information during excavation, as well as feeding new data into the system, automatically updating the online teaching aids. Thus, eventually, architects, artists, archaeologists, engineers, computer technicians, software designers, and telecommunications experts will collaborate in constructing multidimensional libraries that will become the foundations of virtual environments. The use of VRML and the Internet as the ultimate vehicles of this data and its interactivity now makes possible a globally integrated, symbiotic system for the enhancement of public education, public awareness, and interdisciplinary research.

Eben Gay has been building virtual worlds since 1982 for museum exhibitions and classrooms. His work on tradeshow displays has brought him from Boston, where he worked for the Computer Museum of Boston, across the United States with SIGGRAPH and as far as Japan, where he worked for TEPIA. He is active in the Boston Computer Society's Virtual Reality Group and designs VRML environments.