Chapter 3

This section describes an architecture of information systems for use by the US armed forces in 2025. All the capabilities may not be possible by 2025. However, this paper was written to provide the map to near-maximum expected capability. Any stop short of that destination will still have useful features for air power.

The information operations architecture of 2025 consists of thousands of widely distributed nodes, performing the full range of collection, data fusion, analysis, and command functions, all linked through a robust networking system. It is an open architecture allowing modular upgrades without massive redesign. The architecture collects raw data, organizes it into useable information, analyzes and assimilates it, and imparts it in a form that enhances the military decision-maker's understanding of the battlespace. The architecture also applies modeling, simulation, and forecasting tools to help commanders make sound choices for employing military force.

Figure 3-1 shows one vision of this architecture. (Abbreviations are listed in appendix A.) It is a functional rather than a physical depiction. To understand how this architecture operates, it is helpful to divide it into four functional areas that mirror the Col John R. Boyd OODA loop; that is, observe, orient, decide, and act. This division is for illustrative purposes only. In reality, when dealing with information operations, it is difficult to determine exactly where one function ends and another begins. All the nodes are tied together; they exchange information, share processing and storage capacity, and all work together to solve a common problem-superior battlespace knowledge and wisdom. All four elements of the OODA loop are represented in the architecture and are vital to its proper functioning. However, the focus of this paper is on orient and decide, which can be roughly equated to the knowledge and wisdom components. The observe and act functions are the subjects of other white papers and will be addressed here only briefly.

Within the observe component of the architecture, most data collection occurs. Included are all the traditional elements of sensing commonly found in intelligence, surveillance, and reconnaissance. Also included are sensors for weather and terrain mapping, as well as new collection techniques such as noninvasive magnetic source imaging, magnetic resonance imaging, and aircraft wake turbulence detection. ¹ Sensors process data as far forward as possible, at the point of collection in some cases, to reduce overall observation reporting time. New chip architecture offers the promise of lighter and more efficient hardware, improved power requirements, and reduced failure potential for a host of sensor equipped devices. ²

Many weapon systems, especially airborne weapon systems, are capable of contributing their observations to the overall architecture, as well as being capable of autonomous operations with their sensor suites to reduce their reliance on any vulnerabilities in C² systems.

For Battle Effects Assessment, expendable sensors can deploy with the weapon systems. ³ These sensors could consist of miniature gliding flight vehicles that carry onboard processors, independent navigation capabilities, and various sensing technologies including optical, infrared, radio frequency, and acoustic.

The observe component also includes nodes for the correlation and fusion of sensor data from different sources and nodes for sensor cross-cueing to provide automated sensor-to-sensor tip-offs for collection steerage. Additionally, there are nodes for collection management of preplanned and directed search activities. Finally, the observe functional area is tightly linked, accessible, and highly responsive to the act component.

The elements within the act area include those directly supporting a weapon system in accomplishing its task. Of course, the act component in 2025 may well include air power actions other than "bombs on target." The system must provide navigation, combat identification, and targeting information. Weapon systems have direct links to the observe component. This direct link provides real-time (seconds) sensor-to-shooter and sensor-to-weapon data flow and provides near-real-time (minutes) targeting information to planning cells. These links must be developed in conjunction with the development of the weapon system to ensure full integration rather than an add-on capability. Since specific weapon systems design of 2025 is beyond the scope of this paper, this area of information operations will not be addressed further. ⁴

The orient component of the architecture performs what this paper describes as the knowledge function of information operations. It contains the various nodes for automated data fusion, analysis, storage, and retrieval. It is composed of a mix of old and new technologies in an open architecture that allows incremental upgrades of individual elements as technology continues to advance. The architecture is also networked in a fashion that allows graceful degradation as a result of enemy action or component failure (fig. 5).

As a result of many years of collecting information from a wide variety of sources and methods, the architecture's databases contain information on virtually every potential target set or system vulnerable to combat power, both lethal and nonlethal. This information includes an up-to-date compendium of physical descriptions, multiple view images, floor plans, material lists, subsystem component descriptions, technical specifications and drawings, operations manuals, and relationships with other systems. ⁵

This massive amount of information is too large for humans to maintain and keep current without the help of automation. The architecture automatically recognizes gaps, deficiencies, or outdated information in the databases and, without human intervention, searches the global information net. ⁶ It then retrieves the information directly from the various information libraries around the world, or sends a request for collection of the missing or outdated information. The architecture tracks the progress of the response and follows up as necessary. The architecture also reviews numerous satellite images and alerts human analysts of any changes found at potential target areas making obvious exceptions for weather.

Besides information on potential adversaries, the architecture also integrates information on our own and allied forces as reported from the act component. This friendly information includes maintenance status, crew health and availability, location, and mission status. ⁷

New generations of nonmagnetic media-possibly associated with lasers, optical disks, and other newly emerging technologies-will be used to store data. Client-server and distributed data warehouse models can transfer data from the source to the military users' local storage media. ⁸ The architecture can take advantage of lower-cost technologies as well. If massive communications bandwidths are relatively inexpensive, then users' storage devices do not have to be unlimited since the users have unlimited access to source servers. The users simply download what is required for a given mission. However, if cost favors large local memory, then the system could use it and only rely on communications for updates.

Algorithms specifically designed for synchronization, truth maintenance, and queuing delays are used to efficiently integrate all this data from very large distributed databases. ⁹ Every individual data set is tagged with a location indicator to permit immediate and automatic synchronization and alignment of the data or objects of interest. ¹⁰

Data fusion is crucial to taking the massive amount of data available and turning it into useful information without overloading either the human or the information systems themselves. The fusion process takes place across the entire distributed network of sensors, computing servers, and platforms. The architecture integrates fusion applications across multiple nodes using coordination languages to tie together dissimilar operating systems. To do this it employs many separate tools (target models, search, and filtering algorithms) with very large amounts of common sense knowledge. Key fusion functions include automatic target recognition, multi-target tracking, pattern recognition, and object relationship analysis for dynamic situation assessment. ¹¹

Achieving knowledge-level and wisdom-level fusion requires information access technology (IAT) for searching across very large distributed databases. ¹² One promising approach for IAT is the use of artificial intelligence or intelligent software agents (ISA). ISA are discussed in greater detail in the Key Technologies section.

The next portion of the information architecture is the decide, or wisdom, component. With much of the correlation, fusion, and basic-level analysis accomplished by automation, the human will spend less time on where the tanks are and more time on which tanks would be the most effective to attack. ¹³ This is where modeling, simulation, and decision tools come into play.

The wisdom component includes the modeling, simulations, software agents, forecasting tools, decision aids, planning and execution tools, and the archival methods that enable US armed forces' information and knowledge to be superior over an adversary. Usually, the commander who has explored the most alternatives before combat emerges victorious. The forecasting tools will present a range of possible enemy COAs based on the current situation as defined by the knowledge process and based on historic precedence as recalled from the archives. The wisdom systems also identify potential strengths and weaknesses for each forecasted enemy COA. The campaign planner may try out, through modeling and simulation, various friendly responses to each of the enemy COAs. The system identifies probabilities of success and identifies potential weaknesses in friendly COAs.

A powerful new tool in the wisdom component is genius ghosting (fig. 6). Genius ghosting uses the concepts of historic figures, factors in the current context, provides COAs, then simulates the results to provide probabilities of various outcomes. Academic institutions could provide the historical framework. The knowledge component provides the current context. Models provide the COAs. Simulations provide the probabilities of outcomes. ¹⁴ For instance, the Wisdom Warfare system could apply a principle of Sun Tzu: "The doctrine of war is to follow the enemy situation in order to decide on battle. Therefore at first be shy as a maiden. When the enemy gives you an opening be swift as a hare and he will be unable to withstand you." ¹⁵

The COAs would include a reactive strike rather than a preemptive strike. They would include forces in defensive positions until the time is right to strike. The Wisdom Warfare system then "wargames" those COAs to provide probabilities of outcomes. By comparing the COAs provided by many different "genius ghosts," a commander will have a broader range to choose from. For instance, a commander could ask how Doolittle, Kenney, or Horner might design a particular campaign, then pick the elements that work best. In addition, the commander can avoid the dangers of dogma by selecting an unexpected COA, for instance, Doolittle's raid on Tokyo. The goal of genius ghosting is not to rigorously predict how a particular figure would fight a campaign. Instead, it is to give the commander a wider variety of creative options than he would have without Wisdom Warfare.

The Wisdom Warfare system also has a feedback mechanism allowing for course corrections (continuous updates and suggested corrections) based on pitfall predictors (after analyzing decisions and potential outcomes) and way point and metric analysis (indications of what to look for). The system learns from actual outcomes and advises the warrior.

The distinctive advantage of the 2025 wisdom system is that it is nearly autonomous and produces output just as fast as information is added or subtracted. It can be used during modeling, simulation, acquisition, planning, conflict execution, and conflict termination. In addition, this system applies not just to the strategic and operational levels of military operations but to the tactical as well.

A note of caution is appropriate at this point. There are two areas that may cause concern. First, the architecture design needs to recognize that each decision maker has bias in dealing with information. Second, as the architecture becomes human-like, there may be a tendency for the decision makers to become over-reliant on the architecture. This architecture realizes these two concerns and addresses them through the human system integration (HSI) component.

To make the cycle complete, the system and the decision maker must interact to do something useful with that knowledge and wisdom. Given the proliferation of data and the exponentially expanding capabilities to gather data, a major challenge is to extract only the required data and transform it into a useable format for each specific decision maker when and where it is needed. Links for the information operations architecture maximize the use of the national information infrastructure, both government and commercial.

Using ISAs, the network automatically forwards to each node the essential knowledge that is most relevant for that particular node at any given moment. This requires each node to identify the most essential pieces of knowledge by type, level of detail, and timeliness for it to accomplish its mission. Over time, ISAs help users by learning information desired in a given situation. Each node, of course, retains the ability to pull additional information from the system or each information pushed from a superior node to a subordinate node as required.

The objective of HSI is to make it easier, faster, and more efficient for decision makers to adapt to the environment quickly, gain situational awareness, and apply their wisdom to make the best decisions possible. The architecture incorporates the continued advances in areas like time-critical decision making, ¹⁶ reducing information overload, ¹⁷ and human computer interaction. ¹⁸

To allow quick adaptation to the environment, the human sensory and cognitive capabilities will be improved through a combination of technologies and training. The human senses can be enhanced through technology aids and drugs. "Smart" eyeglasses or contact lenses can present more than just the visible portion of the electromagnetic spectrum. Hearing aids can translate a wider range of sounds. Other aids will improve smell or incorporate scents into various tasks like memory recall or heightened sensitivity to help focus decision makers on the task at hand. ¹⁹ The technology aids also augment other senses to allow recognition of emotions to aid in other decision-making environments such as negotiations. Training is provided to teach the decision makers how to use these enhanced sensory powers. This leads to focusing human cognitive functions so they can make the best use of this information.

With a good understanding of how the human brain works, integration of the human and the system is achieved. It consists of improving the presentation of information to the decision maker given a preference for displays, problem-solving methods, current state of mind, and the situation at hand. The majority of this information will be stored in a personal digital assistant (PDA). The PDA can include training, exercises, and real event data.

Additional tools enhance the human's ability to be trained. ²⁰ The goal is to provide a robust training system that takes advantage of the enhancing technologies described above. Through modeling and simulations, decision makers will be presented with the experiences they need to develop the lessons learned that lead to wisdom. These techniques can be used to speed up the training process-similar to accelerated life-cycle testing of hardware.

Displays are adaptive and flexible to account for each individual's preferences. They provide information through all the senses and include text, graphic, virtual, and holographic methods. They are tailored to optimize each user's learning and absorption capabilities. Additional technologies will be developed to allow human interaction with the displays. These technologies allow the displays to work with the human to adjust to each situation. The displays are scalar to allow zooming to the desired level of detail. ²¹ In this way the commander in chief can see the big picture of the battlespace or zoom to see the situation at the local level.

As mentioned above, the PDA learns the profiles of the items the decision maker believes are important and creates information filters to assist in avoiding information overload. The displays, in conjunction with modeling and simulations, also provide the capability of presenting the ghosting of geniuses as desired. In addition, the display is flexible enough to allow several people to view at the same time and through connections make collective inputs to aid the decision maker. This could be done at the same location or remotely using video teleconferencing for a common view of the battlespace.

Displaying a common picture of the battlespace is critical in ensuring the decision maker's intent is clearly communicated to all levels. Three-dimensional holographic displays are useful, particularly for users working in groups. Another example is "smart" glasses or contact lenses enabling the new concept of "eyes-up display." ²² The systems are completely interoperable and are able to tie into the network wherever users are located. The architecture takes advantage of secure, reliable, high capacity communications systems advanced by the commercial world. Through the combined use of these systems the decision makers are able to communicate their intent to all necessary levels and the advantage of having a common view of the battlespace is realized. Figure 3-4 is an example of this common picture of the battlespace.

This section describes some of the key technologies that apply across the entire architecture, including computational power and software.

The computational power contained in this architecture comes from a mix of old (traditional parallel processors, digital signal processors) and new models. One promising new computational approach is based on deoxyribonucleic acid (DNA) molecules. Computer designs based on DNA promise an extraordinary processing capability that operate at billions of tera-operations per second. ²³ While the operations per second rate is very high, it can take hours to complete an entire DNA reaction. Therefore, DNA computing is best suited for complex problems with many variables, such as long-term surveillance and planning, which do not require response times that are measured in seconds. ²⁴ In addition, pipelined, superscalar, and parallel processors show promise for computing power near six billion operations per second. ²⁵

The use of ISAs is vital to the proper functioning of both the knowledge and wisdom components. These agents are software modules that act independently and have a range of capabilities including directed-action, reasoned-action, and learned-action. ²⁶ Directed-action agents have fixed goals and limited ability to deal with the environment and data encountered. Reasoned-action agents have fixed goals and an ability to sense both environment and data and take a reasoned action. Learned-action agents can do all the above. Additionally, they can accept high-level tasking and are capable of anticipating user needs based on general guidelines. Armed with this information, learned-action agents can issue new goals.

Intelligent software agents demonstrate reasoning and persistence in performing tasks. These agents work with their users to determine information needs, navigate the information world to locate appropriate data sources-and appropriate people-from which to extract relevant information. They also act as intelligent, long-term team members by helping to preserve knowledge about tasks, record the reasons for decisions, and retrieve information relevant to new problems. ²⁷

Neural network software provides another capability. Programmers give the system training data with known conclusions. The system then takes a great amount of information and draws a conclusion. ²⁸ In a future where vast amounts of data are expected, systems that feed on data will be valuable.

Hardware and software must be coupled with advanced automated logic methods. For instance, the statistics of Markov chains can be used to predict the highest probability outcome of COAs. ²⁹ Markov chains could be used to evaluate enemy and friendly COAs.

Another modeling tool is the fuzzy cognitive map (FCM). ³⁰ The FCM draws a causal picture to predict how complex events interact and play. It can even handle imprecise rules like: "Bombing an electrical generator usually decreases generator output." The FCM relies heavily on feedback that allows it to be dynamic until it reaches an equilibrium point where a hidden pattern will emerge. This allows predictions of nonlinear system operations, including social systems. FCMs would also be useful in evaluating enemy and friendly COAs.

Chaos theory, a branch of mathematics, provides another modeling tool. Chaos theory deals with the behavior of bounded, nonlinear systems that are sensitive to small perturbations. Chaotic systems often appear to behave randomly but operate within defined bounds. There is reason to believe chaotic behavior occurs in human and organizational decision making and in combat operations. ³¹ Several features of chaos theory should prove useful. First, techniques like "embedding" make short-term forecasting possible and "attractors" describe the boundaries of the long-term behavior of chaotic systems. ³² These would be useful for forecasting enemy COAs, and the outcome of enemy and friendly COAs. Unlike Markov chains and FCMs, chaos attractors can describe the bounds of a number of outcomes rather than just the most likely one. Second, "Lyapunov exponents" help quantify sensitivities to small disturbances. These would be useful in determining what COAs may result in the greatest gains for the smallest additional inputs of military power. Third, calculations of the "information dimension" indicate the minimum number of variables needed to model a system. ³³ The information dimension may indicate that a few variables drive a seemingly random system. Additionally, it makes modeling the system from actual data easier and faster. Overall, chaos theory holds great promise in a wide variety of areas.

Human system integration relies on an integrated use of technologies like: electroencephalograph (EEG), ³⁴ ISAs, information displays, and training programs. EEGs will determine the mental state of the decision maker and tailor displays as appropriate. They will also assist the decision maker in performing computer-related tasks by brain activated control.

The force-multiplying effect of the Wisdom Warfare architecture on the effective employment of US forces presents a potential center of gravity no adversary can ignore. The attack methods expected to be directed against the architecture include the full range of countermeasures designed to disrupt, degrade, deny, and/or destroy, either locally or globally, the information functions provided to US forces.

In an attempt to disrupt the flow of information to decision makers, physical attacks against key nodes using conventional high explosives or electronic signal jamming are expected. These traditional methods of attack are easily countered through hardening (both the electronics and the physical facilities), dispersal, and redundancy. Indeed, the very nature of the architecture, with its multiple nodes and distributed processing, eliminates any "critical node" target or possibility of a single point of failure. Even if individual nodes or decision makers are effectively cut off from the architecture due to enemy action, the immediate effect is felt only at those isolated points and not across the entire architecture. The information flow is automatically rerouted around the disrupted node, allowing a seamless, continual flow of information.

The distributed nature of the architecture coupled with multiple forecasting models also aids its resistance to deception. The numerous observation nodes using a wide variety of sensing phenomenology, correlation tools, and historic databases greatly reduce the probability a battlefield deception effort by an enemy will be successful. By using multiple forecasting models, the Wisdom Warfare architecture is self-defending since the enemy would have to deceive multiple systems simultaneously.

The most dangerous forms of attack are those designed to corrupt, distort, or implant false information into the databases. These types of attacks may occur without any indications the system is under attack. Included in this form of attack are malicious software, computer viruses, chipping (manufacture of computer chips with malicious design flaws), spoofing, video morphing, and surreptitiously gaining local control of the flow of information on the network. ³⁵ Advances in intelligent software, cryptography, and user-recognition techniques offer some degree of protection against these attacks.

The interface software at each node can provide the first level of protection by ensuring the data message that is attempting to gain access to that node is from whom it purports to be. Using message authentication, each node will verify the data message's origin and whether the data has been altered. ³⁶

Intelligent software agents can also be employed to monitor the network for the presence of malicious software and computer viruses. The agents can then attack and eliminate the viruses, or isolate them from the rest of the architecture to prevent their spreading, and notify the human operator for further corrective action.

Preventing computer viruses or malicious software from entering the architecture is a high priority. Cryptographic technology provides very high levels of security against unauthorized, surreptitious access to the information network. Encryption techniques can develop keys that may take eons to break (even using the computational power available in 2025), ensuring secure data at individual nodes and throughout the net. ³⁷

Unauthorized access can also be partially controlled by breakthroughs in biometric identification technologies. These technologies use physiological traits such as voice, fingerprint, eye, or face recognition to provide a continuous identity check of all operators who are using the system's HSI devices to retrieve information from, or input information into, the architecture. If these techniques fail, the system can disconnect any node believed to be compromised or captured.

Finally, unbreakable codes and biometric identification technologies offer no protection against the threat of compromised personnel. Renewed efforts are required to ensure national security policies monitor those individuals who are authorized access to the network and identify potential lapses in architecture integrity. Because technology is constantly evolving, countermeasures and concomitant countercountermeasures will similarly be changing. The operators and maintainers of the wisdom architecture must remain vigilant and continue to make changes to the security structure to stay ahead of advances and changes by an adversary.

In 2025, the system described in this chapter can be used by anyone: the commander in chief, unit commander, supervisor, or technician. Somewhere in the workplace, in a vehicle, or on the person will be a link to the sensors, transmitters, receivers, storage devices, and transformation systems that will provide, in push or pull fashion, all the synthesized information needed to accomplish the mission or task. Information will be presented in a variety of forms selected by the user. Key technologies like advanced processing, intelligent software agents, neural network software, automated logic methods, improved modeling techniques, and improved human system integration will make this system a reality. Certainly, there are countermeasures to such a system and one of the challenges in 2025 will be to protect the architecture both with physical and software security measures.

Notes

1

Spacecast 2020, Surveillance and Reconnaissance Volume (Maxwell AFB, Ala.: Air University, 1994), 3.

2

"Chip Architecture Removes Signal Processing Bottleneck," Signal, February 1996, 58.

3

2025 Concept, No. 900404, "Built-in Battle Damage Assessment," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996); 2025 Concept, No. 900578, "Bulls Eye," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996); Department of the Army, Force XXI, 15 January 1995, 15-17.

4

For additional information see white papers on counterair, counterinformation, strategic and C² attack, close air support, surveillance and reconnaissance (S&R) real-time integration, S & R information operations, and space S & R fusion.

5

2025 Concept, No. 900374, "Living World-wide Intelligence Base, 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996).

6

2025 Concept, No. 900446, "Automated Enemy Analysis Software," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996).

7

USAF Scientific Advisory Board, New World Vistas: Air and Space Power for the 21^st Century, summary volume (Washington, D.C.: USAF Scientific Advisory Board, 15 December 1995), 24.

8

Andrew C. Braunberg, "Data Warehouses Migrate Toward World Wide Web," Signal, February 1996, 35.

9

New World Vistas, (unpublished draft, the information applications volume), 13.

10

Ibid., 57.

11

Ibid., 8.

12

Ibid., 45.

13

Majors Kevin N. Dunleavy and Lester C. Ferguson, "Command and Control and the Doctrinal Basis of the Theater Air Control System," in Lt Col Albert, ed., Concepts in Airpower for the Campaign Planner (Maxwell AFB, Ala.: Air University Press, 1993), 135.

14

Force XXI, 16-17.

15

Sun Tzu, 140.

16

Eric Horvitz and Matthew Barry, Proceeding of the Eleventh Conference on Uncertainty in Artificial Intelligence, Montreal, August 1995. This paper describes methods for managing the complexity of information displayed to people responsible for making high-stakes, time-critical decisions. The area of focus is time-critical applications at NASA's Mission Control Center during Space Shuttle flights.

17

Pattie Maes, Massachusetts Institute of Technology Media Laboratory, "Agents that Reduce Work and Information Overload," Internet address: http://pattie.www.media.mit.edu/ people/pattie/CACM-94/CACM-94.pl.html, 1 February 1996. This paper describes a new style human-computer interaction, where the computer becomes an intelligent, active and personalized collaborator using interface agents that employ artificial intelligence and learn from the user as well as other agents.

18

Allen Sears and Robert Neches, Advanced Research Projects Agency, Information Technology Office, "Human Computer Interaction," Internet address: http://www.ito.arpa.mil/ ResearchAreas/HCI.html, 10 April 1996. This program will support effective and efficient communication between human users and computer-based systems. A key focus is on interactive agents that focus the attention of the user and the software components on critical issues for specific tasks.

19

Richard Axel, "Mammals Can Recognize Thousands of Odors, Some of Which Prompt Powerful Response," Scientific American 273, no. 4 (October 1995): 154-159.

20

Advanced Research Projects Agency, "Computer Aided Education and Training," Internet address: http://www.ito.arpa.mil/ ResearchAreas/CAETI.html, 31 January 1996.

21

2025 Concept, No. 900667, "Real-time War Status Board," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996).

22

2025 Concept, No. 900385, "3-D Holographic Battlefield Display," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996); 2025 Concept, No. 900417, "Battlespace Awareness Holosphere," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996); 2025 Concept, No. 900206, "Commander's Universal [order of] Battle Display," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996); 2025 Concept, No. 900161, "Holographic C² Sandbox," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996); 2025 Concept, No. 900115, "Don't Blink," 2025 Concepts Database (Maxwell AFB, Ala.: Air War College/2025, 1996).

23

New World Vistas, (unpublished draft, the information applications volume), 14.

24

Ibid., 16; Kristin Leutwyler, "Calculating with DNA," Scientific American 273, no. 3 (September 1995):18.

25

David A. Patterson, "Microprocessors in 2020," Scientific American 273, no. 3 (September 1995): 48.

26

Pattie Maes, "Intelligent Software," Scientific American 273, no. 3 (September 1995): 66.

27

New World Vistas, (unpublished draft, the information technology volume), 38-41.

28

E. B. Baatz, "Making Brain Waves," CIO, 15 January 1996, 24.

29

Lt Col Robert J. Wood, "Information Engineering: The Foundation of Information Warfare," research report (Maxwell AFB, Ala.: Air War College, April 1995), 39; John G. Kemeny and J. Laurie Snell, Finite Markov Chains (Princeton, N. J.: Van Nostrand, 1960), 24-25, 182-184.

30

Bart Kosko, Fuzzy Thinking (New York: Hyperion, 1993), 222-235.

31

Maj Glenn E. James, United States Air Force Phillips Laboratory, Edwards AFB, Calif., interviewed during visit to Air Command and Staff College, Maxwell AFB, Ala., 8 March 1996; J. A. Dewar et al., "Non-Monotonicity, Chaos, and Combat Models," RAND Library Collection, Santa Monica, Calif.: RAND, 1991.

32

Maj Glenn E. James, "Chaos Theory: The Essentials for Military Applications," in Theater Air Campaign Studies, (Maxwell AFB, Ala.: Air Command and Staff College, 1996): 38.

33

Ibid., 45.

34

New World Vistas, (unpublished draft, the human systems and biotechnology volume), Appendix M. Using EEGs to determine the state of the operator/user to determine workload and cognitive effort; Airman Magazine interview with Dr Grant McMillan, Air Force Armstrong Laboratory, Internet address: http://www.dtic.mil/airforcelink/pa/airman/0296/look.htm, 10 February 1996. The article describes Dr McMillan's use of EEGs to allow pilots to command a flight simulator to roll to the left or right by thinking about it.

35

Daniel Magsig, "Information Warfare in the Information Age," Internet address: http://www.seas.gwu.edu/student/dmagsig/infowar.html, 7 February 1996, 7.

36

"Public Networks Require Tailored Security Action, "Signal, March 1996, 23-26.

37

New World Vistas, (unpublished draft, the information technology volume), 92; Gates, 109-110.