Encyclopedia of Computer Science and Technology

Knuth, Donald 267istics or relationships that apply to a given individual orclass. For example, the individual “cat” might have a framethat includes characteristics such as “warm-blooded” and“bears live young.” In turn, these characteristics are alsoassigned to the class “mammal” such that any individualhaving those characteristics belongs to that class. A programcan then follow the linkages and conclude that a cat isa mammal. Linkages can also be diagrammed as a “semanticnetwork” in a structure called a directed graph, with thelines between nodes labeled to show relationships.Knowledge representation systems have different considerationsdepending on their intended purpose. A KR systemin an academic research setting might be intended todemonstrate completeness: that is, it can generate all possibleconclusions from the facts given. However, expert systemsdesigned for practical use usually do not attempt to generateall possible conclusions (which might be computationallyimpracticable) but to generate useful conclusions thatare likely to serve the needs of the knowledge consumer.It is also important to note that epistemology (the theoryof knowledge) plays an important role in understandingand evaluating KR systems. As an example, the assertion“Mary believes she is 600 years old” might be a fact (Maryis observed to hold such a belief), but the contents of thebelief are presumably not factual. The context of this beliefmight also be different if Mary is an adult as opposed tobeing a five-year-old child. Similarly, ontological (state ofbeing) considerations can also complicate the evaluation ofassertions. For example, should a fire be treated as an objectin itself, a process, or an attribute of a burning object?Knowledge representation thus intertwines philosophy andcomputer science.The booming interest in extracting new patterns fromdata (see data mining) and the effort to encode moreknowledge into Web documents (see ontologies and datamodels, semantic Web, and xml) all involve applicationsof knowledge representation.Perhaps the most ambitious knowledge representationproject (and the longest-lasting one) has been Cyc (short forEncyclopedia). Headed by AI researcher Douglas Lenat, theobject of Cyc is to create a massive network representingthe relationships and characteristics of millions of objectsand concepts found in peoples’ daily lives and work. Ideallya wide variety of programs (both specialized and generalpurpose) will be able to use this knowledge base (seeexpert system). Projects such as Cyc and the Web OntologyLanguage (Owl) also offer the possibility of a muchmore intelligent Web search (see search engine) as well assystems that can automatically summarize news stories andother material.Further ReadingBrachman, Ronald, and Hector Levesque. Knowledge Representationand Reasoning. San Francisco: Morgan Kaufmann, 2004.Cycorp. Available online. URL: http://www.cyc.com/. AccessedAugust 12, 2007.Davis, Randall, Howard Shrobe, and Peter Szolovits. “What Is aKnowledge Representation?” AI Magazine 14 (1993): 17–33.Available online. URL: http://groups.csail.mit.edu/medg/ftp/psz/k-rep.html. Accessed August 12, 2007.Lacy, Lee W. Owl: Representing Information Using the Web OntologyLanguage. Victoria, B.C., Canada: Trafford, 2005.Makahfi, Pejman. “Introduction to Knowledge Modeling.” Availableonline. URL: http://www.makhfi.com/KCM_intro.htm.Accessed August 12, 2007.Knuth, Donald(1938– )AmericanComputer ScientistDonald Knuth has contributed to many aspects of computerscience, but his most lasting contribution is his monumentalwork, The Art of Computer Programming, which is still inprogress.Born in Milwaukee on January 10, 1938, Knuth’s initialbackground was in mathematics. He received his master’sdegree at the Case Institute of Technology in 1960and his Ph. D. from the California Institute of Technology(Caltech) in 1963. As a member of the Caltech mathematicsfaculty Knuth became involved with programming andsoftware engineering, serving both as a consultant to theBurroughs Corporation and as editor of the Associationfor Computing Machinery (ACM) publication ProgrammingLanguages. In 1968, Knuth confirmed his change of careerdirection by becoming professor of computer science atStanford University.In 1971, Knuth published the first volume of The Artof Computer Programming and received the ACM GraceMurray Hopper Award. His broad contributions to thefield as well as specific work in the analysis of algorithmsand computer languages garnered him the ACM TuringAward, the most prestigious honor in the field. Knuthalso did important work in areas such as LR (left-to-right,rightmost) parsing, a context-free parsing approach usedin many program language interpreters and compilers(see parsing).However, Knuth then turned away from writing for anextended period. His primary interest became the developmentof a sophisticated software system for computer-generatedtypography. He developed both the TeX documentpreparation system and METAFONT, a system for typefacedesign that was completed during the 1980s. TeX found asolid niche in the preparation of scientific papers, particularlyin the fields of mathematics, physics, and computerscience where it can accommodate specialized symbols andnotation.Knuth did return to The Art of Computer Programmingand by the late 1990s he had completed two more of aprojected seven volumes. With his broad interests and contributionsand “big picture” approach to the evaluation ofprogramming languages, algorithms, and software engineeringmethodologies, Knuth can fairly be described asone of the “Renaissance persons” of the computer sciencefield. His numerous awards include the ACM Turing Award(1974), IEEE Computer Pioneer Award (1982), AmericanMathematical Society’s Steele Prize (1986), and the IEEE’sJohn von Neumann Medal (1995).