Audio-video network allows colleagues to work together effectively, naturally

Though much attention has been focused on the value of digital media’s potential to revolutionize the home and entertainment industries, equal potential exists for such media to transform the work environment as well.

During a recent visit we made to Cambridge, England, EuroPARC — the research arm for Rank Xerox, a subsidiary of Xerox Corp. — opened its doors to Digital Media to demonstrate how the use of multiple media (including analog audio and video) has vast potential to revolutionize the way people in organizations work together.

By use of what they call “media spaces,” EuroPARC researchers have created onsite a collaborative work environment that obviates the separation of people by offices and even continents, using technologies that are available today.

The social is the technological. What’s unusual about EuroPARC as a technology research facility is its heavy emphasis on human factors and the social aspects of work environments. In fact, EuroPARC is predominantly staffed with social scientists; only one-third of its 30 staff members are technologists. Even EuroPARC director Bob Anderson is a sociologist and ethnographer, not an engineer.

The benefits of such an approach are probably obvious to anyone who’s been subjected to “user-friendly” software designed by software engineers with no grounding in the study of human interaction. “Programmers’ intuitions about the way people want to interact are often wrong,” says Wendy Mackay, the research scientist who oversees EuroPARC’s “media spaces” efforts.

At EuroPARC, researchers look closely at work habits and how they intersect with computing environments. Folding the sociology of work into the technological tools used to accomplish it is called “Computer Supported Collaborative Work,” or CSCW. Supporting such collaboration takes many forms at EuroPARC, but all share the central goal of creating “environmental interfaces to the information world” — and allow colleagues in locales ranging from the next office to the next continent to work together effectively and naturally.

These interfaces are designed to use existing technologies to both work with and segue into new technologies. Projects in this area include computer interfaces for people who are working with paper, as well as video annotation, the tracking of collaborative design projects and the aforementioned “media spaces.”

Supporting collaboration with technology

As we have all heard by now, statistics show that the massive onslaught of computers into the workplace has automated most jobs that used to be done manually, but has not increased worker productivity one whit. Despite the increased use of local area networks, very little software has been designed to increase the level and quality of real communication between workers. Electronic mail is still a theory to most organizations, even those in the computer industry, and “groupware” that facilitates collaboration, such as Lotus Notes, is only starting to gain a foothold in the corporate environment.

Groupware and electronic mail do increase direct, computer-to-computer or person-to-person (or person-to-people) communication. But even if Notes were installed on every network server in the land, and every organization became an intensive user of electronic mail, only part of the problem would be solved.

Ignoring a rich universe. Researchers at EuroPARC have found that the social setting in which computers are used is a critical factor in effective collaboration and communication, and the existing electronic aids to that communication do not address the rich universe that exists outside the computer. This includes not only information sitting on people’s desks in analog form (i.e., paper documents), but also what’s called “the periphery” — the conversations, either overheard or participated in, that take place in the halls or around the water cooler or coffee pot.

John Seely Brown, director of Xerox Corp.’s Palo Alto Research Center, calls this type of communication “serendipitous.” He believes that serendipity must be incorporated into systems design in order to achieve the true potential of computers as communications tools.

An extensive audio-visual network, RAVE, by far the most developed of EuroPARC’s projects, is designed to bring this serendipity into the work environment seamlessly and naturally. It connects the research center staff — based at Ravenscroft House in Cambridge — and selected sites at the Xerox PARC facility in California in fascinating, and surprisingly useful, ways.

THE RAVENSCROFT AUDIO VIDEO ENVIRONMENT (RAVE)

Ravenscroft House has 27 rooms and five open areas, or “commons,” on four floors. It’s not sprawling in the least, but its layout is such that staffers are surprisingly isolated from each other. In a classic example of “this is not a problem, it’s an opportunity,” researchers decided in 1987 to install a complete data, audio and video network to see if they could mitigate the effects of physical isolation. Each room in the building is transformed into an AV “node,” equipped with a video camera, a video monitor, a microphone and speakers. Individual workspaces are equipped with Unix workstations and Macintoshes. Each node is connected with video and audio cables to and from a central switch.

Connections among the nodes are controlled by computer, so individuals can display views from different nodes on their desktop monitors, as well as set up two-way audio-video connections. Two distinct audio links are used for notification of events and for voice communication. This network, dubbed “RAVE” for Ravenscroft Audio Video Environment, creates what EuroPARC calls the “media spaces” that staffers inhabit in conjunction with their physical workspace.

General awareness vs. focused collaboration. Most collaborative work in today’s organizations is centered around two or more people getting together –whether electronically, via groupware or e-mail, or face-to-face in meetings — to solve a problem. EuroPARC calls this “focused collaboration,” and this type of shared work is supported by most of EuroPARC’s work. What’s much more difficult, where people are physically separated, is to create an environment of general awareness — who is around, what sorts of things they are doing, whether they’re busy, and the like.

The idea is to use media spaces to move fluidly between general awareness and focused collaboration in much the same way people do in physical space, and that’s what the RAVE network and software support.

RAVE: THE SOFTWARE

The RAVE software was originally written in Lisp by Paul Dourish, a computer scientist who specialized in artificial intelligence work at both the University of Edinburgh and the Edinburgh Concurrent Supercomputer Project. RAVE (now being ported to the C++ language) was designed to support a broad range of connections and interactions with the network, which EuroPARC calls “degrees of engagement.”

The components of RAVE. Five onscreen buttons allow network users to control various levels of engagement with the network. The lowest level is the “Background” button (see figure), which allows people to select a view from one of the common areas. At EuroPARC, anyone not actively engaged in other functions on the video network usually sets up his or her default video connection to the main commons, which also serves as a group meeting and presentation site.

“Lots of people use it to see whether there’s coffee in the pot,” says Mackay. And sure enough, though the video quality is fairly grainy, a viewer can most certainly see if coffee needs to be made. Perhaps more practically, it’s not necessary to walk to the commons to see if a meeting is about to begin or whether a specific person has arrived. The result is that network users can maintain a general physical awareness not only of their immediate surroundings, but of remote areas that they’re concerned with as well.

“Sweep” is a time-shortened version of Background; it’s a one-second, one-way, video-only connection to all nodes or any nodes the user selects. It is also used to see who’s around and what they’re doing.

“Glance” is a more focused one-way video connection to a single connected node, also brief (about three seconds). Since both Sweep and Glance are brief and one-way, the effect is similar to walking by someone’s office and glancing in. Thus, general information about someone’s presence and activities is gleaned without jeopardizing their privacy. (More on privacy concerns later.) In fact, says Mackay, permission to Glance is granted by the receiver in advance, as part of RAVE’s setup procedure. Anyone receiving a Glance request is first notified by an audio cue.

The “VPhone” (videophone) and “Office Share” buttons are where two-way audio and video connections are exploited more fully. A VPhone call, much like a regular telephone call, must be explicitly accepted by the recipient. Office Share connections are identical to VPhone calls, except that the connections are designed to last longer — hours, days, months — to simulate actually sharing an office space with a colleague.

At EuroPARC, for example, one Office Share connection is set up between Dourish and a colleague, Victoria Belotti. Although they’re located on separate floors, anyone walking by Dourish’s office can stick a head in the door and ask when he’ll be back and vice versa. Since the video image is relatively small and the sound can be controlled, says Mackay, people don’t feel it’s necessary to engage in conversation unless they want or need to.

WHAT ABOUT BIG BROTHER?

RAVE has been designed to address the kinds of privacy concerns that a video and audio network of this type naturally raises.

Godard, a “smart” AV switch. One basic software component of RAVE, called “IIIF” (for “integrated, interactive intermedia facility”), was designed by former EuroPARC researcher Tom Milligan specifically to give users a degree of privacy and control over who may access their video and audio connections. IIIF’s original function was to link AV devices and “plugs” on the network — i.e., video and audio ports — to form point-to-point connections. Each plug and device is “owned” by its user; thus each user can control how others may connect to them. For example, Mackay could select which users on the network she wants to have access to her “video out” plug, which allows them to view the output of the video camera at her workstation.

But RAVE participants found that using IIIF alone was awkward — simply controlling their “plugs” made it difficult to allow a Glance, for example, but not a VPhone. More useful, they said, would be to control access around the network services themselves.

So Dourish added to IIIF a new layer called Godard, which uses IIIF’s underlying control mechanism to organize users’ connections by the function they want to use. Instead of shutting out or including other users on a global basis, RAVE users can now set up a “Glance control panel” or a “VPhone control panel” or an “Office Share control panel.” Each control panel allows them to select in advance who has permission to access their AV devices for each separate function.

For example, with Godard, Mackay can tell the software, “Only establish an ‘Office Share’ link between me and my assistant, but the following 12 people can ‘Glance’ into my office.” She can accept VPhone calls, like telephone calls, at the time of the call.

Notification via audio. Another useful function Godard performs is audio feedback to network users, telling them that a connection is being made and what kind of connection it is.

Different sounds are assigned to the different network functions. When a Glance connection is made to a camera, for example, Godard triggers a sound (the default setting is the sound of a door opening) before the connection is actually made. When the connection is broken, you hear the door close. VPhone requests might be signaled by a knock or a telephone bell, and a Sweep might be indicated by footsteps.

Though at first using auditory “icons,” so to speak, might seem a little silly, they actually provide much useful information. Sound cues linked to the different system functions of RAVE don’t require focused attention from the receiver, while they provide intuitive, nonintrusive information about what’s going on.

Godard’s sound cues are reminiscent of the old Sonic Finder (for good reason — EuroPARC’s William Gaver, who designed the Sonic Finder, also designed the audio features of Godard). The Sonic Finder, which never quite made it out of the labs at Apple Computer, gave the same kinds of auditory cues about what functions Macintosh users were performing on the Mac’s desktop.

KEEPING TRACK WITH KHRONIKA

The RAVE system also contains a rich distributed environment for “event notification” and selective awareness of events in the work environment. Though EuroPARC researchers say the Khronika system is a cousin to online calendar systems, it supports a more general notion of events — it triggers reminders to users about everything from when a video connection has been made to meetings about to begin to information about visitors. Users can both browse the database and add to it at will.

Khronika allows a user, for example, to create a type of software agent that watches for all seminar events occurring in the conference room with the string “RAVE” as part of their description. He or she can then instruct the agent to send an audio notification five minutes before the relevant meetings will begin.

ERASING THE BOUNDARIES

Though most of the RAVE network’s video is used in its analog form, one system feature captures low-resolution digital video images for network distribution. Called Portholes, its purpose is to promote collaboration between distributed work groups.

Taking a look around the world. Codeveloped by EuroPARC’s Dourish and Sara Bly at Xerox PARC, the Portholes system links certain individual workstations at the Cambridge and Palo Alto research facilities. It captures, displays and updates a number of digital video images automatically, i.e., without having to initiate the Sweep function. Researchers are finding that these remote connections, despite their low quality, help network users both establish psychological connection with users in remote locations and ease the problems of communicating with them.

The precursor to Portholes is an earlier prototype called Polyscope, which only distributes digitized images locally. But unlike Portholes, when an image is selected in Polyscope, it automatically interfaces to the AV network, allowing Glance or VPhone connections to be made. This feature will be incorporated into Portholes, which today only displays information about the selected image and gives the option of sending e-mail to that person.

WHAT’S THE DOWNSIDE?

Initially, EuroPARC’s research appears to prove that an intelligently designed, decentralized audio-video network could provide immediate and obvious benefits to organizations that need or want to increase their communications bandwidth. In the Seybold organization itself, for example, such a system would ease many of the problems inherent in running a seminars-and-publishing operation from three coasts — eastern and western U.S., as well as the United Kingdom — in four separate offices.

In such a situation, making even a simple electronic mail network useful is often an exercise in futility. The ability to make immediate visual connections in a variety of ways with coworkers in different cities and time zones could make a huge difference in the way business is conducted.

Gating issues. However, there are a few problems with the RAVE approach, none of which has any direct bearing on EuroPARC’s implementation.

One immediate question is that of cost. Mackay estimates that setting up the audio and visual nodes of the RAVE system, including the Xerox PARC connection, cost approximately $1,000, plus video and data lines and codecs.

Cost aside, any organization intending to implement a similar network is likely to encounter heavy internal opposition from workers concerned with invasion of privacy. It will be very difficult for people to become accustomed to the possibility of being under surveillance, even if they are assured that “they control the horizontal, they control the vertical,” in the spirit of the old TV show “Outer Limits.”

However, this is likely to be a familiar scenario in any connected office of the future. Groupware has already raised these concerns; live video and audio will only exacerbate it. Mackay says that the EuroPARC team is already acutely aware of these drawbacks, but that the atmosphere of mutual trust and cooperation, in addition to the extensive amount of user control that RAVE provides, makes it possible for the system to work. (Unfortunately, both trust and cooperation are in short supply in many corporations.)

In addition, she says, it’s important to remember that RAVE is a distributed system, where no one person can throw a big switch and peer into everyone’s offices on a global scale. “Big Brother is about centralized control,” she says. “Sure, all the paranoid stuff could happen. But we’re interested in hooking people together, especially when they’re working in isolation. Rather than being paranoid, our attitude is, ‘The technology exists, it’s here — let’s use it to solve communication breakdowns in groups.’”

Perhaps the most significant gating issue to the adaptation of a RAVE-like system is the fact that it is deep in the bowels of a Xerox research lab. Though EuroPARC is working with British Telecom to install media spaces in two of its physically separated engineering sites, Mackay says she’s aware of no plans to commercialize the RAVE software. (The hardware is industry-standard stuff, readily available on the open market.)

If past experience with the Macintosh (and the many other wonders that have sprung from the loins of Xerox’s labs) is any indication, getting a RAVE system into commercial use will require immense, long-term effort from a very dedicated soul. It would be a great pleasure not to have to wait so long.

Denise Caruso