To the Intelligent Collaboration and Visualization Community, DARPA/ITO recently awarded a number of contracts pursuing advances in collaboration technology, and has also restructured some existing programs, redirecting some projects. As one result of these moves, DARPA has formed an Intelligent Collaboration and Visualization (IC&V) Program of which you are all members. This message is to introduce you to the program, the projects, and the people involved. I, Kevin Mills, am the DARPA Program Manager for this program. I am supported by David Sanders, DynCorp, and by Scott Lane, PRC. David serves as a technical analyst assisting with issues relating to program development and review, project reviews, and special projects. Scott prepares all funding documents and should be your first point of contact for all funding and contract issues. The remainder of the community includes: 1) the set of PIs involved in funded projects, as well as the researchers associated with the projects, 2) other interested personnel from several government agencies, and 3) contractual agents who initiate and monitor the contracts associated with the program. We are making plans to bring the IC&V community together in October 1996, immediately following the general ITO PI meeting. Tentative plans are to meet in Dallas on October 9 & 10. The agenda will consist of project briefs in a plenary session followed by small breakout groups to discuss and identify common research or demonstration themes that run through the program. More information will follow soon about this meeting. In the meantime, I am distributing some information to facilitate early collaboration among the program participants. Below find: 1) a description of the IC&V Program Goals and Objectives, 2) a description of the IC&V Program Structure, and 3) an introduction to some of the current members of the IC&V community. I have met with some of you already and intend to meet with each of you in the near future. I welcome your comments at any time. I look forward to a productive association as we work together to advance the state of collaboration technology. Kevin Mills IC&V Program Goals and Objectives ===================================== The IC&V program aims to develop the generation-after-next collaboration middleware and tools that enable military components and joint staff groups to: - gather appropriate problem solvers together across time and space for rapid response in time critical situations, - bring appropriate information resources together across time and space within the context of a task, and - enhance the effectiveness of collaborating problem solvers. Specific objectives: - enable access via diverse portals, from hand-held through room-sized, - enable interoperability across diverse encoding formats, coordination and consistency protocols, and real-time services, - scale collaborations to groups of 10 active contributors, 100 questioners, and 1000 observers, - reduce by an order of magnitude the time needed to generate collaborative applications, - enable real-time discovery of relevant collaborators and information within task context, - reduce by an order of magnitude the time to establish collaborative sessions across heterogeneous environments, - reduce by an order of magnitude the time to review collaborative sessions, and - improve task-based performance of collaborators by two orders of magnitude. IC&V Program Structure ===================================== To achieve the goals and objectives, the IC&V program is organized into four tasks: Task 1: Develop Collaboration Middleware Develop software, leveraging next-generation networking technology, for collaboration across bandwidths, group size, and computing and display environments to gather appropriate problem solvers together across time and space for rapid response in time-critical situations. Task 2: Develop Tools for Sharing Meaning Develop shareable semantic structures, including techniques for automatically capturing, summarizing, and indexing collaborative sessions, to bring appropriate information resources together across time and space within the context of a task. Task 3: Develop Tools for Sharing Views Develop visualization software that enhances the effectiveness of collaboration by adapting views based on task, by enabling manipulation among groups, by representing various collaboration spaces, and by supporting multimedia annotations. Task 4: Prototype and Evaluate Collaborative Applications Develop, instrument, and evaluate prototype collaborative applications to assess the IC&V technology against the specific program objectives. Current Members of the IC&V Community ===================================== The IC&V community currently consists of 20 projects, including 7 moving from other existing DARPA programs and 13 new starts. The projects, organized according to the four IC&V Tasks, are listed below with the PI name, organization and email. For each project, a brief summary is also given. Develop Collaboration Middleware ===================================== Collaboration Support for the Development and Evolution of Complex Systems University of Illinois Simon Kaplan, s.kaplan@dstc.edu.au Daniel Reed, reed@cs.uiuc.edu (NEW START co-funded with EDCS Program) Proposes to develop Orbit, an extensible, evolving collaboration framework that supports the full range of activities undertaken by current and future team members, integrates the wide variety of tools and artifacts they use to realize their goals, and supports the evolution of the development team and its processes over time. Orbit will embody the notion of locales, virtual settings for interaction, as well as an infrastructure for collaboration support, including shared objects, navigation services, audio/video conferencing, and immersive virtual environments. Orbit will also provide a metalayer for tailoring locales, services, and processes. Using Orbit, the researchers propose to investigate techniques for providing a continuum of collaboration "weights", where weight is a measure of the resources required for a particular service. In addition, the researchers will investigate: 1) the use of trader-based object discovery services to assist in continual locale refurnishing, 2) capture and analysis of trajectory of projects, 3) adaptive user interfaces, and 4) scalability to the Internet. Orbit will reduce the effort needed to provide collaborative support for teams from months to days. ------------------------------------------------- Collaboration Bus: An Infrastructure for Supporting Interoperating Collaborative Systems University of North Carolina Chapel Hill Prasun Dewan, dewan@cs.unc.edu (NEW START) This project will address two main problems: 1) composing new collaborative systems from existing single-user and collaborative systems and 2) achieving interoperability among different collaborative systems. The project proposes to develop a new software abstraction, called the Collaboration Bus, that makes it easy to compose collaborative systems. The Collaboration Bus will be an extensible infrastructure that provides general definitions of collaboration services, default implementations of these services, rules for interconnecting these services, and mechanisms for extending the set of supported services. The set of generic services supported by the bus will include (a) a data model for defining distributed shared workspaces consisting of structured objects; (b) a user interface model that allows shared objects to be manipulated through user interfaces that support text, graphics, audio, video, and virtual reality; (c) a coupling model that enables users in a joint session to share data and user-interface objects and controls their interactions to ensure security and consistency constraints are not violated; (d) migration and replication mechanisms that ensure that the bus does not become a bottleneck; and (e) real-time services, consisting of operating system and network support, for ensuring real-time coupling among the collaborators. ------------------------------------------------- Distributed System for Collaborative Information Processing and Learning (DISCIPLE) Rutgers University James L. Flanagan, jlf@caip.rutgers.edu (NEW START) The DISCIPLE proposals explores several research issues with respect to synchronous collaboration systems, including: - Methodologies for managing distributed objects, particularly with respect to partial failure and concurrency coordination in synchronous groupware environments. - The duality of object downloading (e.g., Sun's emerging Java technology) vs. remote object execution (e.g., CORBA) with respect to resource management, QoS, and partial failure survival. - The role of audio and video communication during collaborative activities. The user interface for controlling the QoS. The grades of service needed for effective collaboration will be determined by rigorous psychophysical experiments. - Knowledge-based planning and learning to maximize the QoS, e.g., response timeliness, and computational accuracy. This is particularly important since the system is intended to be used in heterogeneous environments with different mixtures of resources where it will perform computationally intensive data analysis. A special module will monitor the behavior of the system objects, extract patterns of interaction and response, and create symbolic and visual summaries that can be used in learning. - Supplementing the hand-held input devices with a speech recognizer to facilitate human/machine interaction during collaboration. - Integration of robust statistical techniques in the data analysis modules of the user interface. Robust techniques will produce correct outputs even in the presence of severely contaminated data, provided the contamination does not affect the majority of the datapoints. This will permit considerably more flexible human/machine interaction. ------------------------------------------------- Framework for Integrated Synchronous and Asynchronous Collaboration University of Illinois - Urbana Champaign (NCSA) Daniel LaLiberte, liberte@ncsa.uni (NEW START) This project intends to integrate synchronous and asynchronous technologies into one framework. The generality of the framework will allow applications to be deployed for various military related areas, e.g., group meetings, command and control, crisis management, and training. The framework will be used to demonstrate capture, synopsizing, annotation, editing, and replay of collaborative sessions. The client side, java-based framework, called Habanero, will support generic, web-integrated, real-time collaboration capabilities between small or very large numbers of users. Transparent interoperation with external, legacy applications will be through a Java, CORBA, or OLE API, tightly integrating fine-grained functionality of third party applications. On the server side, a Repository will provide capture and replay services for timed event media streams, including video streams. The Repository will be an object-oriented, extensible HTTP server with CORBA/ORB functionality. Support for synchronous document deposit, annotation, registration, and notification will reside in the Repository, as well as support for capture, annotation, editing and replay of any session activities, including real-time video and full-range of multimedia session interactions. Research will also be performed to: create a video datagram protocol (VDP) that adapts to network and server load and client needs, investigate the security implications of the collaborative environment, and interoperate with KQML to support higher level semantics. ------------------------------------------------- Human-Computer Symbiotes: Cyberspace Entities for Active and Indirect Collaboration Hughes Research Laboratories Michael J. Daily, daily@aic.hrl.hac.com (NEW START) This project proposes to develop methods that will facilitate two critical aspects of collaboration, which HRL calls active collaboration and indirect collaboration. Active collaboration refers to the direct interaction between individual integrated human-computer entities called symbiotes. The term symbiotes highlights the coactive, cooperative relationship between humans and technologies that augment human strengths and compensate for human weaknesses. Indirect collaboration on the other hand exploits the emergence of new information out of the collective access patterns, associations, and interests of individual users. This emergent information enables the system to alert individuals to others of common interests, and to data that either contradicts or supports their current working hypothesis. In collaboration, the activities of symbiotes are used as a valuable source of information to others. ------------------------------------------------- Media Net Cornell Univ Thorston von Eiken , tve@cs.cornell.edu Brian Smith, bsmith@cs.cornell.edu (FROM DTII) MediaNet will be built by adapting previously developed technology components to the new environment and by integrating the components vertically into a single system. Specifically, the project will create: * an extremely high performance, low latency communication technology (Active Messages), * a lightweight programming environment for building reliable, reconfigurable, secure distributed applications (Horus) will be adapted to the U-net communication layer, and * a distributed multi-media server technology (CMT) will be reengineered to use the reliability features of Horus and the low-latency of U-net. (URL: http://www.cs.cornell.edu/Info/Projects/MediaNet) ------------------------------------------------- Middleware Services SRI Nachum Shacham, shacham@csla.csl.sri.com (FROM DTII) The project objective is to develop quality-of-service (QOS)-based services for multi-party applications whose group members operate under heterogeneous communication and processing constraints. Current models for QOS on computer network environments are mostly limited to pairwise applications. Furthermore, these models are often implemented by subsuming much of the QOS negotiation and adaptation in applications themselves. The advent of efficient multicasting technology has given rise to group applications which typically are characterized by different group members having different QOS demands on network and computer resources, and different capacities to adapt. Thus, there is a need for QOS models that support adaptive execution of such group applications on complex and dynamic computer network environments (URL: http://www.csl.sri.com/dga.html) ------------------------------------------------- Multi-Tiered Asynchronous Workspace (MAW) Crystaliz Inc. Sankar Virdhagriswara , sv@mail.crystaliz.com (NEW START) The proposed work will deliver a framework that can be used to develop interactive, multi-tiered, asynchronous workspace collaboration applications that span multiple sites, organizations,and work effectively over unreliable wireless and low speed Wide Area Networks (WAN). The primary innovations of the proposed work are : - the use of run time object migration to address latency, bandwidth, and reliability problems of the underlying wireless infrastructure or WAN - Self descriptive object systems and semantic context exchange to address information exchange across workspaces organized in tiers. Each tier is autonomous with respect to the other tiers and the vocabulary they use is different from each other. - asynchronous distributed object infrastructure to deal with the n to n collaboration problem found in collaborative applications for agile environments. The initial focus of the project, funded for one year, is to produce a framework for integrating into CORBA a platform for transmitting, receiving, and launching mobile code, independent of the language. ------------------------------------------------- A Scalable Architecture for Open Interoperable Multimedia Collaboration in Heterogeneous Environments University of California Berkeley Randy H. Katz, randy@cs.Berkeley.edu (NEW START) The project goal is to develop a new system framework that will dramatically advance the state-of-the-art in network-enabled multimedia communications. This will be accomplished by developing a flexible architecture for managing application-level and data type heterogeneity, conference coordination protocols, and new functionality for distance collaboration. The project builds upon two proven technologies developed by the proposal team: the MBone videoconferencing tools and their underlying applications and network infrastructure (vic for video, vat for audio, and wb for shared whiteboards) and the real-time image and video transcoding "proxies" developed for the ARPA Global Mobile Information Systems Program, providing bandwidth adaptive image representations and video streams across bandwidth-constrained wireless communications links. This project plans to dramatically enhance and generalize the MBone video-conferencing application architecture and the proxy concept to support not only heterogeneous media representations for collaboration, but also heterogeneous conferencing protocols, coordination and bandwidth policy specifications and their enforcement. In addition, they propose to develop the technological underpinnings for reliable dissemination of active multimedia objects and new collaboration environment capabilities for session archive and annotation. ------------------------------------------------- Scaling Object Service Architectures to the Internet (OSA) OSC David Wells, wells@objs.com Craig Thompson, thompson@osf.com (FROM DTII) Object Services Architectures (OSAs) are distributed software backplanes with plug-in component object services. Examples of OSA architectures are OMG CORBA and Microsoft DCOM. The objective of this project is to provide objects for the masses by developing technology to make OSAs more scaleable and to merge two exploding technologies, OSAs and the Web. OSCs approach is to develop an Internet Services Architecture (ISA) as an extension to the OMG Object Management Architecture (OMA), which was based partly on earlier DARPA work. This involves identifying technology barriers to OSA scaling, determining how to overcome them, and demonstrating some of the software mechanisms necessary to help bring enhanced interoperability, portability, and reusability to OSA-based Internet applications. (URL: http://www.objs.com/) ------------------------------------------------- Develop Tools for Sharing Meaning ===================================== World Wide Web: Evolving to Information Infrastructure MIT Michael Dertouzos, mld@lcs.mit.edu Al Vezza, av@lcs.mit.edu (FROM DTII) This project pursues the evolution of the World Wide Web protocols for defense, other government and commercial use. The evolution is to ensure scalability, security, extensibility, efficiency, robustness, automateability, and so forth. The activity has three efforts. Because the Web has attracted a great deal of commercial interest, the MIT Laboratory for Computer Science (MIT-LCS) is leveraging this government-sponsored effort by asking industry to cost share development of Web protocols and reference codes. To that end, MIT-LCS has established a World Wide Web Consortium (W3C). The Consortium also provides an excellent technology transfer mechanism of ARPA's research results to the industrial and commercial sector. The second effort is to establish a research activity to solve difficult problems which if left unsolved would hamper the effective evolution of the Web for defense, other government and commercial use. The third effort is aimed at collecting research results available within the computer science community and folding those results into the Consortium development effort where intellectual property constraints allow. (URL: http://www.w3.org/pub/WWW/) ------------------------------------------------- Supporting Organizationally Accepted Practices (SOAP) USC ISI B. Clifford Neuman, bcn@isi.edu (FROM DTII) This project aims to establish a framework for network applications and protocols that support automatic adherence to organizationally accepted practices. Making such practices automatic means that conscious effort is needed to circumvent the practice and allows social mechanisms to be applied when such actions are discovered. The framework will support intellectual property and privacy protection. It will support the selection of information and service providers based on class of content and confidence in the service provider (e.g., accreditation). It will also support prioritization and filtering of service requests according to declared class of service. Meta-data containing copyright clearance information, price, confidence, ratings, and purpose are encoded as attributes to data objects, and fields to pass the declared class of service, ratings,and desired restraints on the retention of transaction information are added to transport protocol messages. Endorsements for services are signed certificates that are stored as attributes of the service in directory service entries. Attributes on data objects are preserved when data are cached or otherwise copied. Applications use the meta-data to initiate payments for intellectual property (possibly prompting the user for approval first), and to prioritize or hide from view inappropriate information, services of unknown quality, and objectionable material. Servers receiving requests with limits on collection and retention of summary information act within the constraints imposed by the user or they reject the request. (URL: http://gost.isi.edu/gost-group/arpa/9507-saber/) ------------------------------------------------- Build an Intelligent Information Infrastructure MIT Randall Davis davis@ai.mit.edu (FROM DTII) The information infrastructure should have intelligence embedded into it, allowing it to understand the information it is carrying, enabling it to provide new ways to gather, organize, and transmit knowledge, as well as new ways to run organizations. This project is creating systems that provide and explore a variety of such capabilities. In order to embed intelligence into the infrastructure, they need to identify sources of knowledge that will provide power. They believe that important sources will include knowledge about message content, and about individuals and organizations. Knowledge about message content is being provided by a sequence of successively more powerful natural language understanding technologies, while knowledge about individuals and organizations is captured in knowledge bases they are building. They believe this will prove substantially more powerful than the numerous keyword-based systems currently in use. They have built a variety of systems, including the publication/distribution system used by the White House Office of Media affairs and an on-line surveying system used to determine the size and character of the audience receiving the documents. They have also developed and used the START system to provide an natural-language based information resource at the AI Lab. (URL: http://www.ai.mit.edu/projects/iiip/home-page.html) ------------------------------------------------- C-Space--Collaboration Information Management Carnegie Mellon University William L. Scherlis, scherlis@cmu.edu (NEW START) This project will develop a new class of tools and services to create, organize, and exploit shared information. The project proposes specific technical innovations in the following areas: - Foreground structuring techniques for managing personal task-specific structure information, based on ML and Java data types, hypertext, and object based models. - Incremental Interpretation for foreground structures, enabling structure and meaning to be given in steps, unlike web approaches (always uninterpreted) or knowledge-base approaches (fully interpreted). - Creation and management of shared collaborative spaces and their interconnection with individual foreground information spaces. - Broadly applicable building block tools for common collaborative activities: collaborative note taking, progress reporting and alerting, information sharing (as in collaborative bookmark development). ------------------------------------------------- The Campaign-at-a-Glance: Conceptual Models for Visualization Information Extraction and Simulation University of Massachusetts, Amherst Paul Cohen, cohen@cs.umass.edu (NEW START) This project proposes a hypothesis that semantic models can be built in two layers, where one layer is a model of general semantic concepts from the physical world and the second layer uses the lower layer concepts to construct a domain-specific semantic model. To test this hypthesis, the project proposes to build and demonstrate a Campaign Planner's Assistant that performs visualization, simulation, what-if analysis,information extraction, and language-based discourse from a shared semantic model. The research will develop a novel kind of campaign model for command and control, along with inference methods that operate on this model. ------------------------------------------------- Collaborative Planning As Distributed Programming University of New Mexico David Ackley, ackley@cs.unm.edu (NEW START) As group activity is increasingly mediated by computers, and as computer languages and systems are increasingly needing to handle distributed computations and heterogeneous resources, there is an emerging potential for a synergy between the rules governing a given group's behavior and the rules governing a distributed computation. This project proposes research to unite these two language domains, producing a prototype evolvable rule system for supporting group activity involving both human and computational agents. The goal is to extend the principles of distributed computation upward to meet the typically informal procedures used by working groups, while extending planning language down into a more-directly executable form. Building on an ongoing project, called ccr, this project will build prototype systems that automate several aspects of the function and management of groups on computer networks--including, most significantly, mechanisms for changing the rules themselves. The goal is an evolvable metasystem that allows the group members themselves to control many aspects of the computation and communications systems that hitherto have been fixed by a priori software design. The advantages possible with such an approach include: - improved use of resources, as groups tailor the system to their specific needs; - better access control and data security, by recognizing 'groups' as first-class objects within the system; - improved group cohesion, identity, and productivity; and - the emergence of reusable new, effective group interaction and management styles from the bottom-up, user-driven evolution of the group rules system --in effect, better user-created 'groupware'. ------------------------------------------------- Develop Tools for Sharing Views ===================================== Visualizations and Briefings that Acquire Portray and Communicate Experience Carnegie Mellon University Steven Roth, Steven.Roth@cs.cmu.edu (NEW START) This project will develop an integrated environment for acquiring and maintaining situational awareness in information-intensive domains. Not only will they create innovative component technologies for supporting information visualization, exploration, analysis, explanation, and briefing, but they will also coordinate these technologies within a single, seamless user interface environment. The resulting environment will reduce the data gathering, analysis and briefing cycle from weeks to hours. Situational awareness arises from a wide array of experiences involving the gathering and communication of information: analysis and simulation system results, memos, phone conversations, comments between collaborators discussing a map, doctrinal manuals and other text documents (e.g., OPLAN's), digitized images, briefing materials and meeting minutes (and the notes people write on them), email messages, and WWW documents. The approach will go far beyond current technologies, which focus on collecting structured database facts about crises. They will provide external supports for capturing, portraying, enhancing and communicating the vast accumulation of information-based experiences that must come together to coalesce in situational awareness. ------------------------------------------------- Distributed Clients Open Group Research Institute (formerly OSF) Paul Dale, pauld@osf.org Murray Mazer, mazer@osf.org (FROM DTII) The Open Software Foundation Research Institute (OSF RI) proposes to design, build, demonstrate, and evaluate a fundamentally new class of access software, the distributed client. Such clients will support information browsing, search, retrieval, publication, and archiving in the world-wide web. Distributed clients will deliver more advanced functionality, better performance, improved reliability, and greater security to users than today's monolithic clients such as Mosaic and Netscape, while making better use of network resources, including storage, transmission bandwidth, and processing power. (URL: http://www.osf.org/www/dist_client/) ------------------------------------------------- Virtual Environments for Direct Software University of Illinois reed@cs.uiuc.edu (NEW START) Proposes to develop virtual environments that allow software developers to directly manipulate software components and their behavior while immersed in scalable, hierarchical representations of software structure and real-time performance data. The output of this project will be Virtue, a collection of libraries, all with standard APIs, embodying: 1) hierarchical software representations, 2) software component controls, 3) manipulation tools, 4) annotation tools, and 5) external interfaces to performance sensors and octuators. ------------------------------------------------- Prototype and Evaluate Collaborative Applications ===================================== Virtual Work Rooms Carnegie Mellon University Stephen E. Cross, scross@cs.cmu.edu (NEW START) Extend current research efforts to create and demonstrate a Virtual Work Room (VWR). A work room is a place where all the objects, people, facts, analysis, and deliberations related to a specific project come together. The dedication of physical space and time makes actual work rooms impractical for all but the most important projects, but computer mediation makes it possible to overcome some of the spatial and temporal contraints to actual work rooms. On the network, one can create as many VWRs as needed. All concerned can visit at any time and from any real place. This work will build upon MITRE's Collaboration Virtual Workspace (CVW) environment. The main thrust of the proposed research is to develop and apply methods for evaluating collaboration technology. ------------------------------------------------- Evaluation Approaches and Metrics Additional involvement in evaluation approaches and metrics during the first year can be expected from at least three other sources: the National Institute for Standards and Technology, the National Exploitation Laboratory, and the Mitre QuickTurn project (which uses CVW).