This issue of DG-Online focuses on homeland security research now underway by grantees of the Digital Government program. This is hardly a surprising choice given recent events, but it is definitely not just a case of "Me too". Defending against terrorist attacks like those of September 11, and managing the recovery from them, is a function clearly reserved for agencies of government. Information technology already plays a large role in such efforts and can be extended to provide new and better counterterrorism tools.

Authorities typically find themselves struggling not just with the consequences of the catastrophic events, but with the problem of obtaining timely information about unfolding events. Background information about the environment, available personnel and other resources is often unavailable, and planning and coordinating a response in such uncertain and rapidly changing circumstances is nearly impossible. Furthermore, trying to act in this ill-defined environment can leave the authorities and society exposed to additional attacks.

While it may appear on the surface paradoxical to attempt to "prepare for the unexpected", the current state of information technology makes it possible to create a general infrastructure and develop general capabilities that can be adapted instantly to any threat. Society cannot afford to prepare for every eventuality, but it is possible to create a foundation upon which a response can be constructed quickly. What is needed is the following:

• Encyclopedic knowledge and training A first priority should be to amass a general digital collection of all information about geography, environment, resources, establishments, buildings, computational facilities and personnel everywhere in the US, together with software systems that have the ability to search out answers to pertinent questions. The goal of research should be finding general tools that can analyze very large stores of data, identify patterns, reason, extract and summarize in natural language and in graphical and other forms, translate text, incorporate incoming data into the existing store, and answer questions. Given any task, relevant information and instruction in how to use it must be extracted and provided automatically.
  
• Unlimited Grid-supported computational ability A network of all available computational and data storage facilities supporting utilization of any resource should be assembled. Computational Grid technology already is being developed that can support the sharing of computational and data storage resources distributed over the Internet. What are needed are secure and automated methods of finding the resources required to execute computational tasks, as well as methods of splitting up computations and data, delivering them to different locations on the network and recombining the results. The process should be transparent, and functional within priority constraints and network and resource limitations.
  
• "Eyes and hands" everywhere Instantly deployable, self-configuring networks of sensors and effectors-from simple actuators to autonomous robots - must be created. Devices capable of detecting motion, heat, light/images, sound, pressure, the presence of metal, and much more, already exist. There is research on developing micro-scale "machines" and reconfigurable robots. We also need to be able to quickly deploy such sensors and effectors so that they can autonomously network among themselves and communicate with controllers outside the crisis zone. The units would be able to gather data and function autonomously, as well as convey first-hand information to emergency managers who could issue additional commands remotely.
  
• Virtual Organizations: Collaborative autonomous intelligent agents Organizations that unite geographically dispersed humans, software and hardware systems into a flexible, resilient, dynamic and coordinated effort must be established. Many modern computer systems are more than just tools- they understand group goals and their own roles in achieving them, and thus can become active participants. Research is needed on agents that perform jobs such as information discovery, task management and coordination, as well as support functions that do not require the human hand. Showing how human/agent organizations can best function in a collaborative fashion is another worthy goal.
  
• Security Data, computers and networks must be rendered secure from eavesdropping and sabotage. A concerted effort to develop new methods to protect computer systems from intrusion and sabotage and recover from such disruptions-possibly utilizing extended virtual organizations-will ensure that emergency response teams can operate in safety, free from surveillance and malicious interference. New security initiatives must cover wired and wireless networking and communication, for both speech and data.
  
• A legal framework for emergency collaboration Ways must be found to resolve the jurisdictional and legal problems that hamper or even prevent information sharing and collaboration among different governmental and other organizations, while at the same time maintaining essential privacy protections.
  

Although new initiatives in these areas might be crisis-driven, they could be brought to bear on a wide range of problems endemic to digital government. For example, the permit process for building a new home, which now involves multiple computer systems and human-staffed agencies, could be reduced to an automated and transparent system handled by proxy software agents with no human involvement except when absolutely necessary.

No one foresaw the explosion of media, art, electronic commerce and education that resulted from the creation of the communication infrastructure of the Internet. The organizational infrastructure I'm proposing may be aimed at protecting our current way of life. But if fully realized, it may yet end up transforming it. Our special role in the war on terrorism is also an obligation. We must work as never before to bring these promising technologies to fruition.