Progrm Official: Nora Sabelli

Program Name: Dissemination and Technology Utilization Section

Award Dates: 5/94 - 1/95

Award Institution: Syracuse University

Award Number: 94-53871

Proposed goals: (a) assess potential for VR in education and the research and development required; (b) plan for a followon proposal.

Methods: A literature search and monitoring of online discussions and electronic articles on VR were already underway and were expanded. Technologies for VR were investigated and evaluated.

Conclusions: Development of the VRML standard and of technologies for integrating VRML-based (initially non-immersive but later immersive VR) applications with the WWW will make it easy to focus on design and content for education instead of on programming. Web server environments need expansion to bridge with VR applications. NPAC could leverage WWW work in progress to become a leader in Web-based VR for education.

The near term value of immersive VR technologies in education is likely to be diminished for three reasons; (a) they are expensive and proprietary; (b) the technology is too immature; and (c) the danger of "simulation sickness" is real, significant and insufficiently understood. An underexplored role for VR in education that has high potential is navigation of conceptual spaces such as information spaces and databases.

Results: (1) A compendium of material on VR applications with many references and extensive information from current online discussions on VR and potential applications; (2) Evaluations of VRML with notes on prototype Web applications that use VRML; (3) NPAC development of server tools that integrate multiple server functions and form an environment for VRML-based applications; (4) NPAC plans for continuing Web server extensions and for tools to integrate VRML applications.

For several years, NPAC has been monitoring developments in VR research and development technology and as of April 1995 report the following:

Until the fall of 1994, VR technology was fragmented and dispersed over a large collection of small start-ups. There was no clear leadership, either from industry or from the government side, hence no natural mechanisms appeared to facilitate formation of standards.

Cost and proprietary issues involved with full immersive VR technology currently appear to make it impractical for education in its present state of development. A further concern for the education community is a potentially dangerous temporary cognitive disconnect called "simulation sickness," caused by immersive VR.

The lack so far of a strong national trend in use of high cost, high bandwidth technology in schools also makes unlikely the near-term development of a large, competitive educational VR market that would drive down prices.

The WWW community is developing a standard for a Virtual Reality Modeling Language (VRML) that will initially support non-immersive VR and will be integrated with the Web via new application environments such as Java-HotJava. VRML defines universal data structures which allow exchange of objects and applications between different networked people and groups. This will allow re-use of products between different groups and equally interestingly, faithful exchange of 3-D objects between collaborating users on the WWW. This suggests a new type of interactive collaborational simulation which could have great educational significance. VRML has excited a great deal of interest in the Internet community and appears to hold great promise for near term applications that can be used in K-12 education. It is public domain, hence applications will probably be numerous and will include many public domain products.

We conclude therefore that the most fruitful current direction for research and development in VR for education will leverage VRML and the new generation browsers such as HotJava. It will focus on exploiting these technologies, developing Web server tools needed to bridge between VRML applications and existing Web servers, and seeking imaginative ways of using these technologies to support learning in the network-connected classroom.

Development of the VRML standard and of technologies for integrating VRML-based, non-immersive VR applications with the Web will make it easy to focus on design and content for education instead of on programming. Web server environments need expansion to bridge with VR applications. NPAC can leverage WWW work in progress (with significant tools in active use) in this arena.

There are two other areas that the team members from Information Sciences have noted as having a high potential near term value and being particularly suitable areas for federal funding, as their value is primarily educational with little obvious commercial appeal. For NSF and other Federal agencies interested in the use of Information Technologies for improving the effectiveness and efficiency of K-12 education, consideration should be given to investing in research that addresses (a) learning processes and (b) distributed multimedia database organization for realistic access to content. These are two issues that have been underfunded in the current rush to develop technology per se. For example, if we can learn which combination of interactive strategies promote sustained activity and interest on the part of K-12 students to meet educational objectives, this knowledge can then be used for guiding VR development, initially as nonimmersive WebVR and later perhaps as full immersive VR, when and if the issues noted earlier are satisfactorily resolved.

VR and Web technologies: Initiated by Tim Berners-Lee, a public consortium was formed during the first WWW Conference in Geneva, May 1994, with the goal of developing an extension of HTML towards VR on the Internet/Web, to be called Virtual Reality Modeling Language (VRML). VRML can be thought of as a 3-D HTML. A proposal based on a subset of SGI Open Inventor format was submitted by Marc Pesce, Enterprise Integration Technologies (EIT), Inc., Toni Parisi, Labyrinth Group, and Gavin Bell, Silicon Graphics, Inc. (SGI), and was accepted in fall '94.

During the second VRML conference in Chicago, October '94, a draft specification for VRML 1.0 was presented and the VRML team started developing the first generation of VRML browsers. The Internet community responded to the VRML concept with enthusiasm, evidenced by a rapid rampup to 800 active particupants on the VRML listserver that was started and maintained by WIRED magazine and EIT. By the end of '94 it was clear that the VRML effort would soon have major impact on both VR and Web technologies and communities.

The opportunity these developments presented for rapid development and deployment of VR applications via the WWW caused us to change our original plans for this project implementation. In addition, cost and proprietary issues involved with full immersive VR technology appear to make it impractical for education in its present state of development. Thus the project focus became concentrated on monitoring events in the VRML domain, and on initiating a Web software development effort that would offer a bridge technology between the current, HTML based Web and the coming VRML paradigm.

In the following sections, we continue the overview of VRML and of technologies that can support its implementation, such as Java and HotJava. We also describe the Living Textbook and HyperWorld-WebTools projects at NPAC that have been undertaken in anticipation of integrating VRML and other Web technologies for educational applications -- what we call the development of a prototype Educational Information Infrastructure.

In January '95, Pesce published a VRML parser that reads VRML script and constructs internal memory representations of individual VRML nodes in terms of C++ classes. This software speeds up the VRML browser development process and reduces the task to the platform-specific implementation of the 3D graphics classes, constructed by the parser. Based on this software, several VRML development efforts were under way during the first months of '95.

In March '95, SGI announced the VRML browser called WebSpace for the Silicon Graphics platform. Beta release is expected by the end of April '95 and it will include both a public unsupported version and the commercial fully supported product. At the WWW conference in Darmstadt, Germany in April '95, VRML 1.0 was officially announced by Mark Pesce and early WebSpace demos were presented by SGI. The focus of the VRML team will shift now toward dynamic simulation support for VRML, and the corresponding specification VRML 2.0 is expected by the end of '95. The VRML listserver operation is now moving towards a regular newly formed USENET group called "comp.vr.vrml".

The language: VRML 1.0 is an interpreted language that can specify static 3D scenes in terms of a collection of objects, possibly distributed over the Internet. The core of the language, constructed as a static subset of the SGI Open Inventor format, contains some 40 data structures called nodes that comprise all major constructs for modern 3D graphics and are optimized for OpenGL-conforming rendering platforms. Integration with the Web is provided in terms of special nodes for object inlining and linking. Inlining lets one specify some scene components in terms of URLs rather than VRML scripts, where each URL points to a script that will be dynamically retrieved by a VRML viewer during the scene building and rendering process. Scene rendering is to be performed locally at the client side. The retrieval process is fast as it is based on scripted vector representations of objects (i.e. text) rather than on compressed binary bitmaps as in video streams. Object linking lets one assign URLs to individual objects in the scene so that they become active in the same way as on current imagemaps, i.e. a click on an object activates the retrieval of a Web page pointed by the associated URL.

Driver applications: Most members of the public VRML development forum are from the corporate world. The driving force behind corporate interest in VRML appears to be anticipated Internet commerce applications. These may be in areas such as interactive shopping and product advertisement. The first applications expected in coming months will offer tools for 3D graphics-based Internet navigation, following the existing hyperlink model: individual sites will develop and export their characteristic 3D icons which will be linked to sites homepages or to local VRML objects.

VRML Icon libraries: Icon URLs will be maintained by a digital library or electronic mall service provider. A user of such a service is offered a collection of 3D scenes that are dynamically integrated by downloading and rendering individual icons, and then suitably sequenced, nested, zoomed, panned etc. in response to user clicks on individual icon objects.

Dynamic Simulation: Dynamic simulation support, not provided in VRML 1.0, will be specified in VRML 2.0 and will available by the end of '95. The full Open Inventor model which served as a base for VRML 1.0 specification does provide animation support and hence will likely affect the further development of the VRML protocol.

Java and HotJava: Recently (March 95) released in Alpha by Sun Microsystems, Java is an interpreted C++ subset and HotJava is a multithreaded Java-interpreter-based dynamic browser with support for arbitrary simulation dynamics at the client side. Chunks of Java scripts called "applets" (application inlets") can be inlined as custom HTML tags. A HotJava browser responds to an applet tag by downloading the corresponding script and dynamically executing it at the client side. Already a growing community of porters and applet developers has produced experimental applets that range from games like Tetris, Reversi or Pong, to FFT or sorting algorithms demos.

Synergy of new languages: Java and VRML address two complementary aspects of Internet VR: VRML is focused on structure, and Java is focused on function. One possible complete VR simulation environment could be constructed by combining Java and VRML, e.g. by structuring the VRML interpretation and rendering system as a collection of Java classes, dynamically linked to low level 3D rendering libraries. Following the Open Inventor animation model mentioned above is another alternative. Using Safe-Tcl interpreted environment to support VRML simulations, as explored by San Diego Supercomputer Center, is yet another current approach.

Applet Examples: In the distributed Atari Pong demo, two HotJava browsers bounce the ball off the common server. More sophisticated VR simulations will connected many servers and browsers through a suitable dynamic communication protocol. Such high-level communication scripts must be knowledge-based to cope autonomously with complexities of distributed interactive simulations on the Internet, and hence the intelligent agent models such as Telescript by General Magic offer a promising communication paradigm.

Further evolution of the VR technologies on the Web will most likely include and integrate current paradigms such as VRML, Java and Telescript. Several detailed scenarios are possible and a full convergence towards a uniform world-wide standard is unlikely to happen in the near term. It is more probable that the Web will be populated soon with a variety of dynamic simulation models, competing for a broader public acceptance.

For the standpoint of considering VR technologies for education, a crucial aspect of the current developments is that only public, fully open protocols stand any chance of being accepted. Thus, the WWW phenomenon and current VRML efforts have resolved one of the challenges listed in our proposal: how to specify a research agenda and pursue technology assessment of proprietary and expensive VR technologies, secretly guarded by the entertainment industry. In fact, we expect the "conventional" VR industry to be pushed by Web developments to make their systems compatible with VRML, Java and other emerging, interactive Web standards. We hope to witness another phase transition on the Internet, driven by the new standards, that will bring many existing VR and CAD databases to the public forum, as well as facilitate collective development of new VR objects, scenes and worlds on the Internet.

Living Textbook Consortium

A primary forum through which NPAC hopes to reach K-12 audiences is the Living Textbook project, which has catalyzed the formation of the Living Textbook Consortium. The Consortium brings together institutions and individuals with many areas of expertise to create prototype systems that will deliver exciting and educationally sound high bandwidth applications to schools. The technological environment, which we believe represents a realistic norm of perhaps five years hence, includes a scalable digital network supporting both ATM and ISDN technologies, multimedia clients (PC, Macintosh and Unix), and large scale parallel multimedia information, database and simulation servers.

The Living Textbook Consortium is a partnership to develop, use and evaluate education services that build on this leading edge infrastructure. Currently active are NYNEX and NYSERnet supplying the ATM NYNET network, 6 schools (use by teachers and kids), School of Education and teachers (assessment and educational integration), NPAC and Computer Science and Engineering (educational software and technology services integration), School of Information Studies and TextWise Inc. (sophisticated text retrieval), and over 15 content providers covering a range from local culture to top international sources of news in image, text, and video.

Tools that enhance Web server capabilities, both those developed at NPAC and those incorporated from elsewhere, are continually being integrated into the environment of the Living Textbook, just as they are also merged into the general NPAC information system environment (see section below titled "WebTools Based Hyperworld"). The plan of the AAT VR proposal will be to include integration of all relevant educational VR development into the Living Textbook environment as part of the plans for demonstration and dissemination of the technology.

Science for the 21st Century

As computing power and network bandwidth increase, it will become

more and more possible to expose undergraduate and secondary school

students to experiences of "real science" in the form of computer

generated experiments in many fields of scientific study. Web server

gateway applications and Web client APIs will provide the foundation

for integrated systems that can enable bringing meaningful experiments

to any locvation that has adequate bandwidth on the Internet. For example,

students will be empowered to run computational experiments that range

from simple illustrative models to complex simulations of real physical

Following are some examples.

Neural Network simulations are being used in Syracuse Physics 106 Introductory Science for the 21st century for undergraduates. Here it is possible to enable changing the "laws" of the system under study, as these "laws" are update rules and connection strategies which can easily be set by parameter. The experiment with an undergraduate course in neural nets, which is underway, has lectures enhanced with material supplied over the Web; but the real transformation in the teaching and learning experience of the course comes from making it possible for students to set parameters and run sample neural net simulations over the network, using a WWW browser interface.

Server and client extensions are being used by NSF funded projects that support remote control of instrumentation over the network. One such project in obserational astronomy allows a student to select an interesting celestial object for observation, send the telescope pointing data from a Web page, select certain parameters for taking a digital picture of the object, and bring the picture over the network into a client tool that does image display and manipulation.

Interactive video will present a multimedia or video artifact where the viewer can interact with the presentation and change between a set of pre calculated video streams. Thus the viewer can change the outcome of a Movie or learn from viewing a particular video stream. At a modest level this capability is present in many current multimedia games such as MYST. However there are now examples like the notorious Nighttrap, where the video plays a central role, or a set of Sherlock Holmes multimedia interactive detective stories, that use this strategy.

Interactive Simulations

Interactive simulations allow the student to change the program as well as the input data. Placed in the context of an appropriate learning environment for the area of science under study, this could allow a student to see the real impact of laws of physics. This is important, as many areas of physics are difficult or impossible to abstract meaningfully to a small realistic model. As a prime example we note the Navier Stokes equations, which are the foundation for all weather and climate studies as well as simulations of airflow around airplanes, oceanic currents, clear air turbulence and many astrophysical phenomena. The use of Cellular automata for CFD is one example of such an attempt, which has limited applicability.

Here we use the full power of HPCC to develop a realistic simulation and do not abstract the latter in order to run with modest resources. Because so many physical phenomena are hard to abstract, this is most general and powerful approach. Note that one can start with this approach and as one gains experience from interacting with and controlling simulations, one builds up an intuition that can enable identifying essential features of the type of simulation that would be needed in an abstraction. A serious problem with full scale interactive simulation is lack of scalabilty. We cannot provide a supercomputer for every school. However, today's supercomputer is tomorrow's PC, so the level at which simulations can be done by a wide audience will steadily improve.

A good example of this is the application at Argonne called Labspace which provides a realistic simulation support to MOO interactive collaboration environments. These are quite popular and use a spatial metaphor to set up a collaboratory environment. It is plausible that realistic simulation will make these ideas far more appealing, as current MOO interfaces are text based and this is obviously limited in their ability to represent a 3D collaborative world. The shared visualization of the 3-D world will probably rely on the emerging VRML standard for its visual display in a Web-based collaborative environment.

Use in a field of biology such as medicine would be challenging in other ways, as the complexity of the system necessitates that the simulation will be partially abstracted, yet the total learning experience is tremendously enhanced by the ability of the student to make certain input choices that will play out to different, physically realistic results. Again, VRML would enable shared 3-D visualization of biological systems.

Original limitations of NPAC: In our original, unfunded AAT proposal submitted in January 1993, we outlined our concept of a HyperWorld, defined as a collection of virtual worlds, linked by suitable portals and offering an experiential and shared VR based interactive educational environment. At that time, we were unable to sharpen and test these concepts through prototyping owing to the lack of VR infrastructure at NPAC.

Leveraging the new paradigm: Many of our original concepts now can be explored in the current HTML based WWW framework, using the support for the dynamic Web server extensions provided by the CGI (Common Gateway Interface) protocol. Hence, anticipating the emergent Web based VR technology developments, we started last fall at NPAC a Web software development effort, aimed at providing a bridge or interpolating platform between current, HTML based WWW and the coming VRML based Web.

We found the problem of "how to implement virtual reality in hypertext" both challenging and potentially leading to practical and useful solutions. The natural starting point for a prototype implementation was to identify a Web server with a "World unit" and to view HyperWorld as suitably linked collection of Web servers.

Such a model lacks two major features of VR: active participation via content authoring, and the spatial metaphor. These features are also of interest for interactive education applications. A natural extension building towards VR thus includes on-line HTML editing support and document tree navigation tools.

Web-based GIS:To explore the spatial metaphor, we have constructed in the context of the Living Textbook consortium, a web-integrated GIS (Geographical Information System). This combines 3D terrain rendering with "web-sites" (multimedia hyperlinked buttons) overlayed on the GIS. This was designed before VRML emerged as a standard and although we can use our prototype for experimentation in educational applications, further development should be based around VRML.

Enhanced Web environment: Using Perl as a prototyping environment, we developed a collection of WebTools, structured as a CGI/Perl based extension of the Web server technology, providing the required functionality. The content of a WebTools enhanced Web server is under control of a Manager tool which allows clients, given suitable permission, to create, destroy and manipulate individual nodes (files, directories) of the document tree. The associated tool -- On-Line HTML Editor -- supports interactive content authoring on pages created by the Manager. Each such page comes with the default "HyperWorld Navigation Bar," built out of the HTML substance, and providing the metaphor of spatial hierarchial navigation through the document tree, viewed as a 'space'. The Navigator Bar supports primarily the intra-world navigation confined to a given document tree. We are now also adding support for inter-world navigation in our HyperWorld, through a collection of WebTools-enhanced Web servers, all registered by a MetaServer.

The WebTools project was motived by and started as part of this project, and then was continued based on other NSF educational grants at NPAC. WebTools were recently tested in a computational science course at SU on Web technologies: a WebTools enhanced server was used both for course material authoring and for interactive exploration and extension of the server content and CGI code, conducted by students in a set of class projects. We are also experimenting with other WebTools modules such as full hypermedia e-mail support. In the next stage, will support interactive forums, project management, shared calendar and other applications.

We see that both the national scene(Web and VRML) and our local developments such as the Living Textbook and WebTools, positions ourselves well for the development of a full proposal for the AAT. Components of this proposal could include:

a) Collaboration among NPAC, IST, the School of Education, Industry, Teachers and Kids already developed in the Living Textbook consortium for non VR applications.

b) The NYNET physical infrastructure with ATM, ISDN networking and high performance servers as a microcosm of the future NII emnvironment.

c) VRML and advanced server and client Web technologies on top of which we can build educationally specific applications and services -- the Education Information Infrastructure (EII).

Two complementary development areas to be included in the proposal are the use of VR-based simulations in education and CASE (software engineering) tools for developing VR-based educational software. The educational web content should include VR motivated navigated in information base as well as the direct VR simulations.

We have not yet determined the applications that our proposal should focus on but we will base this on our local activities such as the Living Textbook as well as the national activities as surveyed in resource list produced by Nilan and Small as part of this SGER grant.

Locally we are working with the Living Textbook Consortium, which includes several technologically literate and involved teams of teachers as well as members of the Schools of Education and Library and Information Science. This group is tasked with conceptualizing meaningful and significant applications in science and possibly mathematics that would make use of the Web-VR environment and would also take advantage of the scientific and computational science expertise at NPAC.

We will look through the printed resource of references that the SGER project has produced for possible model examples where we think the Web and VRML can be naturally linked for education. Note that these technologies open up the collaborative VR environment which could be a centerpiece of our proposal. We will also attempt to evaluate other known education efforts in the VRML and Web communities such as that at the Geometry Center in Minnesota where they developed the OOGL (Object Oriented Graphics Language) and are now converting it to VRML.

As outlined above, our exploration of interactive Web technologies includes monitoring the external activities such as VRML or Java, and pursuing the internal development efforts. Our Web technology development efforts are solid and robust, and we are confident of being able to shape our internal Web software engineering so that it is complementary with the external activities. We see the promise of near term results from continuation of our HyperWorld/WebTools effort, augmented by VRML and Java technologies, and focused on authoring tools for educational applications.

Reuse and leveraging commercial developments: We note that SGI will be focused in the near term on Internet commerce applications of the VRML/WebSpace, whereas Sun's main focus with Java is the consumer electronic market (e.g. smart-settops for interactive TV). From these and other efforts, the Web will be populated soon with a broad spectrum of VRML objects and scenes, Java applets, agents and other useful objects and tools. Although script development will be driven by Internet commerce, we can start accumulating scripts and addressing the issues of reuse for building educational VR applications on the Web. Corresponding software industry efforts will be focused on digital media studios and content renting for the game developers, whereas we would provide analogous but public and open object bases, CASE and authoring tools for education and edutainment applications.

We will continue monitoring the interactive Web development activities, especially in the area of VRML, Java and agents integration and continue to participate in the listservs.

Publications.

Note I took first part of their book of resources and made into SCCS tech report

A: VR Papers

M. Nilan, R. Small and W. Tirenin, "Internet Data for VR in Education" NPAC Technical Report SCCS-707

(A) W. Furmanski, "Supercomputing and Virtual Reality", in VR Becomes a Business, Proceedings of the Meckler Conference Virtual Reality '92, San Jose, CA, Sept 23--25, 1992.

(A)Faigle, C., Fox, G.C., Furmanski, W., Niemiec, J. and Simoni, D., "Integrating Virtual Environments with High Performance Computing", in Proceedings of the 1st IEEE Virtual Reality Annual International Symposium, VRAIS'93, Sept 18--22, Seattle, WA.

(A)Furmanski, W., "Supercomputing and VR Networking", in Proceedings of the

Meckler Conference Virtual Reality '93, San Jose, CA, May 19--21, 1993.

(A)Furmanski, W., "Integrating Virtual Reality with High Performance Computing using the MOVIE System", paper presented at the 3rd Virtual Reality Systems '93, New York, NY, October 18--21, 1993.

(A)Furmanski, W., "Virtual Reality Technology for Cable TV", Assessment Report for Continental Cablevision, Internal NPAC Report, January 1994

(A)Fox, G.C., Furmanski, W., Hornberger, P., Niemiec, J., Simoni, D., "Implementing Televirtuality", book chapter in Applications of Virtual Reality, 1994, by The British Computer Society, Computer Graphics and Display Group

B: Web Technologies

(B)Fox G.C., Furmanski, W., Hornberger, P., Niemiec, J. and Simoni, D., "Towards Interactive HPCC: High Performance Fortran Interpreter", handout material and demo presented at the Supercomputing '93, Portland OR, Nov 15--19, 1993.

(B)Furmanski, W., "HPSIN: High Performance Software Integration", March '94. http://www.npac.syr.edu/NPAC/PUB/PROJECTS/wojtek/hpsin/home.html Pages in NPAC information server describing the following projects:

HPSIN itself (as an umbrella project)

MOVIE (Multitasking Object-oriented Visual Interactive Environment)

HPFI (High Performance Fortran Interpreter)

HPCGS (High Performance Coarse Grain Distributed Computing Support)

Virtual Reality and Televirtuality -- A technology assessment project.

Pages in this HPSIN space contain also a link to REFERENCES, which offers an on-line version of most recent papers up to the end of '94. Some '95 references are currently linked on the http://kayak.npac.syr.edu:2005 server in /WebTools/Potpourri.html

(B)W. Furmanski and G. C. Fox, "WebTools" -- experiments with interactive Web technologies, currently tested within the CPS600 Course Server at http://kayak.npac.syr.edu:2005. See /WebTools for overview and recent references (currently collected in /WebTools/Potpourri.html).

(B)G. C. Fox and W. Furmanski, "WebWindows: Motivation and Application to Distributed Metacomputing", talk and demo presented at the CRPC Annual Review, Houston, TX, March 22.

C: Applications

W. Furmanski, editor, "Intelligent Machines", A Module for the Course PHY106 "Science for the 21st Century" by Marvin Goldberg and Ed Lipson, Syracuse University, January 1994.

W. Furmanski, and K. Hawick, "Technology Integration Service (TIS) for the dual-use of the Rome Lab / Air Force technologies", June '94 http://king.syr.edu:2001. Includes early version of the VR information space at http://king.syr.edu:2001/vr.

W. Furmanski, "Nearterm Technology Development for Education on the NII", NPAC Technical Report, March 1994.

W. Furmanski, "Next Generation WWW Technologies for Distance Education", in Proceedings of The Virtual University Conference, Wharton School of Management, University of Pennsylvania, Jan '95.

Add

(C) G.C. Fox and K. Mills Mohawk Valley Paper

"High Performance Computing and Communication-Yet another revolution in education"

G.C.Fox, K Hawick, M. Podgorny, K.Mills, "The Electronic InfoMall - HPCN Enabling Industry and Commerce", submitted to HPCN 1995

G.C. Fox, W. Furmanski, K. Hawick, D. Leskiw, "Exploration of the InfoMall Concept - Building on the Electronic InfoMall", Report to Rome Laboratory, May '95

K.Mills, "Information on Demand Technologies and Applications", HPCWire 1514, Nov '94, Special Feature (SCCS-649)

G.C. Fox, W. Furmanski, K. Hawick, D. Leskiw, "Exploration of the InfoMall Concept, NPAC Technical Report SCCS-634, Aug. '94

K. Mills, G.C. Fox, B. Shelley and S. Bossert, "The Living Textbook: a Demonstration of Information on Demand Technologies in Education", Proceedings from Jun 95 NECC (National Education Computing Conference).