Knowledge Structuring:
The Substance of Information Technology

A condensed course (five days, nine two-hour lectures) for student and researchers in

• Computer Science and Engineering
• Information Technology
• Information Sciences (documentation, library management, etc.)
• Communication Sciences (mass media, distance learning, etc.)
• Sociology
• Business Management

This course aims to provide a unifying rationale for the vast array of concepts and techniques of today's information technology: Why are we doing all of this and where do the single pieces fit in the overall picture? This unifying theme is Knowledge Structuring, an adaptive necessity for business organizations, social institutions, and---not the least---human individuals.

Lectures of a more conceptual nature (afternoons) will alternate with lectures of a more technical contents (mornings). Early afternoons will be available for hands-on practice.

Prerequisites: Computer literacy, some exposure to programming. Strong interest in issues of computer and society. Ease to move back and forth between conceptual and pragmatic levels of discourse.

(This is a very condensed version of a one-term course originally developed for Carnegie Mellon University, School of Computer Science, and Boston University, Electrical and Computer Engineering Department, and targeted at Master of Science students in Information Technology.)

1. Introduction: Knowledge and knowledge structuring: a natural sphere of action for the computer and the Internet
2. Semantic tagging: indexing and cataloging; hypertext, markup schemes; XML and all that
3. Fundamental idea #1: Surface structure vs deep representation
4. Information capacity vs computation capacity. Standards in information encoding and knowledge transmission
5. Fundamental idea #2: Scripting: a self outside of our self. The way to talk to a computer
6. On the engineering of intention and rationale; hierarchy, modularity, encapsulation; objects
7. Fundamental idea #3: To model is to understand. Just-in-time models. Expert systems
8. Ownership and privacy issues
9. Fundamental idea #4: Knowledge structuring as adaptation: lessons from evolution. The individual and the organization are different organisms and thrive on different forms of knowledge

Knowledge and knowledge structuring

• Examples of knowledge structuring activities.
• Knowledge is possession of useful models of the world.
• The acquisition of knowledge on the part of a system (an individual or an organization) is inherently a process of modification of the system itself, entailing pain and expense.
  How can we organize knowledge so that the system can achieve large functional enhancements through reasonably small structural modifications? This is the underlying theme of the course.
• In the planning of knowledge artifacts, the nature of the intended recipient is an essential design component; different systems require different information-structuring strategies.

Semantic tagging: indexing and cataloging, hypertext, markup; XML and all that. Applets.

• Citations, keywords, catalog cards. Conventional databases.
• From typesetting to just-in-time Internet publishing.
• Structured markup languages: SGML, XML. One map can say different things to different travelers.
• What happens when passive annotations are replaced by objects capable of action?

Fundamental idea #1: Surface structure vs deep representation

• Both in nature and in engineering, complex objects are fabricated from a recipe ("how to make it"), not from a picture ("what the finished product looks like").
• Limits of the WYSIWYG ("What you see is what you get") interface; literate use of the computer requires even the layman to be, on a humble scale, a prototyping designer, and thus be comfortable with moving up and down between surface structure and deep representation.
• The "split-screen" metaphor as a modern alternative to WYSIWYG: you edit the design plans in one window; the computer shows you in real time the resulting product in another window. Pedagogical and technical advantages of this approach.

Information capacity vs computation capacity. Standards in information encoding and knowledge transmission

• Tradeoffs between storage, communication, and processing. Compression schemes.
• Fungibility of information channel capacity. Fungibility of processing capacity.
• Hierarchical storage and caching. Search engines. Just-in-time computation.
• Language and literacy. The birth of international standards.
• Languages, romanization, and transliteration. ASCII and Unicode.
• UNIX and Windows; the WWW Consortium. Graphic file formats. HTML.
• Most standards go extinct; a good document is able to "jump hosts".
• Human individuals as incompatible operating systems; ways to help make their knowledge interchangeable.
• The "informal records" alternative to semantic tagging: examples, advantages and costs.

Fundamental idea #2: Scripting: a self outside of our self. The way to talk to a computer

• Scripting as the natural way to externalize and share "how to".
• The dual nature of a script: an editable, passive object, and a real-time, active recipe.
• Existing scripting languages. Desiderata for better ones.
• Giving orders, documenting, making dependencies explicit, expressing intention. The UNIX "make" discipline.
• Documentation: after or before the fact?
• Scripting as a tool for cultural transmission.

On the engineering of intention and rationale; hierarchy, modularity, encapsulation; objects

• From lines of comments to lines of code.
• Goals and reasons are vital elements of maintainable design; can we begin to automate their handling?
• What kinds of knowledge-structuring efforts will this require? (Cf. automated typesetting, program compilation, object-oriented discipline, VHDL.)
• The problem with monolithic constructions: they are old as soon as they are born.
• Artificial intelligence vs externally-supported human intelligence.
• Do we have to die 100%? The art of picking (and pickling) one's brain.
• From Simula to Java. Nonlocalized computing.

Fundamental idea #3: To model is to understand. Just-in-time models

• Various kinds of models. Conceptual models; simulation models. Knowledge representation schemes.
• Statistics vs mechanism as a basis for prediction.
• Object databases, enterprise modeling, and decision support systems.
• Various forms of capturing and reusing expertise: expert systems, systems for experts, expertise on demand.
• Custom integration of independently-developed expertise resources.

Ownership and privacy issues

• Accessibility and copyability of information. Property rights; pay-per-use; open source.
• Web pages are public but volatile; can I "memorize" them? And to what extent do I own my memories?
• Knowledge structuring adds value; who owns this value?
• The knowledge that benefits one may be a threat to others.
• Privacy; protection from view vs protection from trampling.
• Cryptography: algorithmic and physical pieces of armory.

Fundamental idea #4: Knowledge structuring as adaptation: lessons from evolution.

• Competition, sociality, and symbiosis. Multiple levels of adaptation.
• Where is the knowledge located? The Chinese room paradox.
• The individual and the organization are different organisms and thrive on different forms of knowledge.
• The support of corporate knowledge. A company's nervous system. Distributed knowledge.
• The individual's many roles. Long-term structuring vs knowledge prototyping.
• What knowledge functions can corporate enterprise outsource to the individual?
• Support and enhancement of the capabilities of the individual. Personal knowledge structuring. "The best way to learn something is to write a book about it".
• External parameters and symmetry breaking; standardization and innovation.
• Knowledge structuring is a valuable and expensive process; it deserves a thought.