Great Expectations in Information Technology

Just as Moore's law is not more than a self-fulfilling expectation, Information technology in its entirety has to face great expectations concerning functionality. Driving force behind all these expectations is the example of the human brain. Accordingly, we expect information technology to approach the following ideals

• Flexibility concerns the array of particular situations that are premeditated by the programmer and can therefore be handled by the program.
• Robustness is the ability of a system to handle also unforeseen problems.
• Situation awareness is the ability of a system to notice and factor into its decisions all relevent information in a given situation, especially information about the user's intentions.
• Coherence is the ability of a system to draw all possible consequences from every information item available to it, bringing all data and subsystems in line with that information.
• Autonomy is the ability of a system to put up and enact its own agenda. This necessitates the erection and maintenance of a goal hierarchy.
• Autonomous adaptability is a property that lets a system autonomously alter its internal structure in contact with a given environment in order to be able to act appropriately in that environment.
• Evolvability must be given for a system to improve in small steps instead of by complete redesign.

All these challenges arise for systems by having to face complex or unstructured environments. Some specific expectations concern the following aspects

• Communication systems, where the challenges are ubiquity, transparency, speed, data coherence again, system interoperability and the broad issue of security.
• Coherent data bases. The challenge is to organize a body of knowledge that is doubling in just a few years, to make it coherent and consisten, and to ensure correspondence between reality and data. Part of this is the issue of enterprize software that gives all relevant information at the push of a button. Another aspect are the bioinformatics or neuroinformatics efforts.
• Man-machine interfaces. These have to be improved fundamentally to open the computer world to more and more people without formal training and to adapt the computer world to us humans, instead of requiring us to adapt to the computer world (cite Jaron Lanier).
• Perception. Situation awareness, man-machine interaction and coherence of data with reality cannot be realized without autonomous perception in natural environments.
• Comprehension of natural language. This is a central issue here, and is probably the most difficult to achieve.

Eventually, these expectation and ideals can only be attained by endowing machines with intelligence. They certainly cannot be met without revolutionary productivity increases in software production, which require fundamental rethinking of the algorithmic division of labor.