N.B. Back in 1999, when I was a computer science professor and department head at the University of Illinois at Urbana-Champaign (UIUC), I considered all that I knew about technology and research trends in computing, wireless communications, networking, modeling and simulation, and I asked myself, “What could life be like for my typical student in the early 21st century?”
I gazed into my admittedly cloudy crystal ball and wrote a vignette entitled “A 21st Century Student Experience,” which the University of Illinois College of Engineering found interesting (or amusing enough) to publish in its alumni magazine. This essay contains that vignette and an assessment of technological and educational evolution since 1999.
A Team Project …
It’s a beautiful spring day in 2006 as you stroll across campus on Friday afternoon, listening to a custom newscast playing in your earpieces. An interesting story on the World Cup competition catches your attention. A subconscious glance at the video inset in your peripheral vision expands the video for viewing.
You subvocalize “weekend weather” and your heads-up display shows you that this will be a great weekend for the upcoming intramural soccer match. That’s why you really regret procrastinating on your industrial policy term paper. It is due Monday, and your team hasn’t finished all the background research and simulation, much less assembled the multimedia report. Old Professor Reed is gonna fail you for sure.
You’re wearing featherweight glasses with a retinal laser scanner for high resolution color display, two tiny earpieces for digital audio, and a bone conduction microphone in your necklace. A skin conduction network connects all these devices to the lightweight computer in your pocket. About the size of a deck of playing cards, it is your local interface to the information environment, connected via a high-speed wireless network that covers all of Champaign-Urbana. The computer hosts your information space preferences and manages your interactions with other members of the information community. It is also your “bodyguard” – providing a level of security when outside a trusted environment, including encryption and strong authentication. Intelligent agents scan news sources for information that matches your interests, continually adjusting priorities based on monitored psychometric responses.
Your multidisciplinary team has decided to analyze the economic and environmental effects of (now common) electric cars on greater Chicago. Although use of electric cars has dramatically reduced automobile emissions, it has increased electrical demand. Even with effluent scrubbers, there are indications that greater use of high sulfur coal by electrical utilities is leading to more acid rain. Carried by prevailing winds, this seems to be affecting corn harvests on the farms to the east. Despite improvements in battery technology, commuters still complain about recharging far too often. In turn, this has changed commuting patterns, increased the number of “work at home” employees and created an insatiable demand for ever faster communication networks.
You and your classmates are part of a multidisciplinary course sequence that cross educates students from the technical disciplines in sociology and human factors, and humanities and arts students in technology. Your professors team teach the course sequence, using real-world problems to stimulate cross-group interactions and analysis. In your case, the members of your team are drawn from finance, environmental studies, agriculture, sociology, and computer science.
The 21st Century Ambient Laboratory
Nearing the research laboratory, you realize you can’t delay that project meeting any longer; you subvocalize “schedule a team meeting” and walk inside. In the background, your ambient environment context shifts from your wearable computer to a higher performance system that controls the intelligent room. The lights adjust automatically to your preferences, the contents of your heads-up display shift seamlessly to one of the wall screens, and one of your favorite Mozart concertos begins playing softly in the background. Concurrently, your intelligent agent negotiates with the agents of your team members, balancing social commitments and work schedules to set up a meeting at 2 p.m. on Saturday.
As with most classrooms and research laboratories, the intelligent room in your lab has several wall screens. Their principal role is to act as convenient places to chat with remote collaborators, scan your custom newscast, and correlate text and video. Wall screens consist of a large, flat color display panel, a touch sensitive layer, embedded microphone, and speakers, all controlled by a small processor. The wall screen processor uses other services on the network to parameterize the display according to your preferences and needs (e.g., correcting color deficiencies in your vision). In this ambient computing environment, a user’s session moves with him or her, exploiting and sharing nearby resources as needed.
In addition, speech recognition and synthesis, lighting and sound controls, psychometric sensors, smart books, and intelligent whiteboards; all cooperate to manage and adjust conditions based on behavior and context. Many of these devices rely on low-power networks, distributed computation, and dynamic agent negotiation to provide a high-level, perceptual “gestalt.”
Seeing the scheduled meeting appear on their calendars, two of your team members hail you. Andrea is driving home to her apartment. Bill is catching a late lunch at a campus pizza place. Audio and video from the visor camera and microphone in Andrea’s car appear on your wall screen and through your speakers, whereas Bill’s audio is localized in your earpieces, providing contextual spatialization.
Andrea reports that she has some preliminary statistical analyses of Chicago area traffic data, projecting commute times and electricity use based on population growth and continued suburban sprawl. The three of you examine this data, shown as a temporally evolving simulation overlaid on the Chicago highway grid. Andrea’s view is a high-level schematic, limited to avoid occluding her vision while driving. Bill and you share a three-dimensional stereoscopic view on your heads-up and wall displays, respectively.
You suggest looking at the effects of shifting air traffic from O’Hare to the new South Suburban airport on commuter traffic; Andrea agrees, promises to pursue it, and records an audio “post it” note on her analysis as a reminder. Bill promises to have similar data on network bandwidth by tomorrow afternoon’s meeting, obtained by analyzing sanitized data from Ameritech’s “network weather service.”
The raw data for traffic and network communication are hosted by state, city, and commercial information archives. Visual data mining, supported by gesture and oral commands, allow users to correlate data across sources (e.g., specifying queries based on geographic regions and traffic routes). Security and authentication services limit access to statistically significant queries that do not compromise privacy.
Data, Simulation, and the Quest for a Good Grade
You’ve put off your part of the project, studying acid rain dispersion, as long as possible. Turning to your wall screen, you begin with Andrea’s data on electricity use and projected growth. You grab simulation modules for atmospheric and weather prediction, as well as ground water percolation and corn growth models, connect them using a “put that there” visual assembly system, and drag data for the initial simulation conditions onto each module. You launch a set of parametric simulations, exploring the effects of varying growth rates, increased work at home, and greater purchases of Canadian hydroelectric power.
As the models start to run, you retrieve satellite imagery of cloud cover, rainfall, and temperature from the Earth Observing System (EOS) archives, targeting the upper Midwest for the preceding five years. You create a two-minute video, false colored with corn yields, which will form the visual baseline for comparison with your simulation.
High-speed networks provide low latency access to remote data archives, and high-performance computing systems host complex, multidisciplinary simulations. The information substrate negotiates resource use from a distributed computational grid, scheduling simulation components and managing data invisibly.
It’s nearing 3 a.m., and you’re desperate for some food and sleep. As you leave the room to walk to your apartment, the room reverts to standby mode and transfers your information context back to your wearable computer. As you tiredly walk home, you send voicemail to Andrea and Bill describing what you’ve done, advising your agent to alert them if the simulation finishes before you awake.
At 2 p.m. the next afternoon, the three of you meet to integrate your data and simulations and begin drafting the report. You “difference” the simulation and the video you created last night to show the effects of acid rain on corn yields. It’s depressing – if no action is taken, the mean reduction in corn production looks as if it will approach 10 percent in five years.
Meanwhile, Bill notes that you’ve been overly pessimistic about the growth in available network bandwidth – those computer scientists are even cleverer than you thought. He drops in more realistic projections and launches new simulations while Andrea starts a hyperlinked report outline.
Because the report will target urban planners, telecommunication network designers, and environmental experts, each will be interested in different levels of detail on particular topics. Andrea lays out a three-dimensional graph of report topics, connecting placeholders for subgraphs of detailed analysis on acid rain, network use, crop yields, commute times, and electricity consumption. With this outline, you each begin creating components.
Bill prefers to type on an old-fashioned keyboard; you’d much rather lie on the floor and dictate to your speech-to-text agent. Andrea is tactile, opting to grab three-dimensional text, audio, video, and simulation icons and place them in the report.
It’s now 3 AM, pizza boxes are scattered on the floor, and that oldies rap group “2 Live Crew” is jamming on speakers. However, the report is coming together. You’ve shown that the best balance of economic growth, environmental protection, and crop yield will require a 20 percent acceleration in network deployment, closure of one coal-fired electric plant, and purchase of an additional 5 percent of greater Chicago’s electrical power from electrical power brokers.
The draft report is a combination of computer narrated text, interactive video, multivariate data displays, and three-dimensional, immersive visualizations, all annotated with multimedia notes from the authors. Sill, it seems to be lacking a bit of pizzazz. You decide the report really needs a human interest angle. You query the media archive for news reports on corn yields, commute times, and work at home trends. The query returns some samples, and you select those that match most closely, launching a new query that emphasizes video.
Jackpot! Here’s an Indiana farmer discussing low corn yields and wondering if he’ll be able to pay his daughter’s college tuition next spring. In addition, there’s a Chicago commuter with a dead battery and a crying three-year-old stranded on a downtown expressway. You juxtapose the two, segueing through an image of an electrical generating station.
You look at each other and think, “This has gotta be an ‘A;’ even doddering old Professor Reed isn’t that hard nosed.”
Predictions: How Did I Do?
Ah, how cloudy was my crystal ball? Looking back, what did I foresee correctly, and where was I clueless or maybe just a little too optimistic? What were the transformative, ubiquitous technologies? Which were exhibits in the museum of vaporware? Where did reticence and traditionalism unduly constrain organizational change?
Ubiquitous
I got some things right.
- Broadband wired and wireless communications. Mobile phones are everywhere and broadband wired and wireless access is common. Sadly, bit rates and access still lag in the U.S., particularly in rural areas, which continues to lag other parts of the world.
- Large, flat panel displays. LCD/LED displays are now mainstream consumer items, and flat panel television sales have made tube displays just a memory.
- Web service integration. In addition to mature web services for business and standard APIs from Google, Amazon, Microsoft, and others, mashups and everything-as-a-service (XaaS) where among the hottest investment targets in Silicon Valley, at least until the AI boom.
- Cloud services. Cloud services are now ubiquitous, enabling consumer and business-to-business data sharing. I underestimated their dramatic growth and the rise of generative AI. Ironically, I spent years at Microsoft working on cloud services and supporting public-private cloud research services. (See Microsoft/NSF Cloud Research Partnership)
Vaporware
I was optimistic (and wrong) about some other technologies, but not completely so.
- High-resolution, consumer heads up displays. I had hoped we’d see more innovation in the portable I/O space. Alas, laser retinal displays remain but military prototypes. Google Glass and Microsoft Hololens failed in the consumer marketplace, and the future of Ray-ban Meta glasses is now being tested in the market. (See Through a Google Glass, Darkly)
- Gesture interfaces. We are still largely confined to the 30 year old WIMP (windows, icons, mouse and pointer) metaphor. Where are the new interfaces that combine gesture, voice and situational awareness? How do we escape the limitations of the QWERTY keyboard, itself designed in a mechanical era? (See Don’t Be a Wimp)
- Mobile device integration. The cell phone is as close as we have come, but its secondary features (camera and sensors) and user interface are not yet up to the task. However, generative AI is helping, and mobile device control of other devices (music and home automation) is now common.
Inertia
I underestimated how hard it is to change higher education, and I hoped too much for software interoperability.
- Multidisciplinary education. Sadly, it remains the exception rather than the rule. The archaic organizational silos of our educational institutions perpetuate educational stovepipes, rather than the consilience of knowledge. More worrisome, educational reform is not keeping pace with the rate of societal and economic change; the societal backlash reflects a belief that our universities are insufficiently nimble and responsive to shifting workforce and economic needs.
- Rich semantic interoperability. Although web services exist, complex composition by users remains out of reach. At a minimum, one must be proficient in a scripting language. Solving this problem would open access to a broad audience. After all, the web browser didn’t really allow one to do anything that could not be done with anonymous ftp, but it did make it easy. Perhaps agentic AI will help.
Please leave a comment …