A person who gathers knowledge and incorporates it into computer programs such as expert systems and natural-language processing systems.
But now dozens of technology start-ups are commercializing linguistics research, and competing to hire the relatively small pool of specialists on the topic, which isn't even taught at many U.S. universities. Suddenly, linguists have their pick of jobs as lexicographers, 'knowledge engineers' and 'vocabulary-resource managers.'
Daniel Golden, "No Longer Just Eggheads, Linguists Leap to the Net," The Wall Street Journal, May 30, 2000
A good example of how such rule-based systems work is a programme, called Dipmeter Advisor, which interprets geological data from an oil-well measurement device (a dipmeter). A typical rule from the programme goes:
IF there is a red pattern over a fault, the direction of the red pattern is perpendicular to the fault, and the length of the red pattern is greater than 200 feet, THEN the fault is a growth fault.
Other rules start: ''If the fault is a growth fault . . . '', and so on.
Dipmeter Advisor was developed (with the help of artificial-intelligence workers at MIT) by Schlumberger, the French-controlled company that has an international stranglehold on the business of measuring well geology for oil companies. To get at the geology of a well, various measuring devices (including dipmeters) are lowered into it on thin wires and data signals sent back to produce a ''log'' of the well.
Interpreting such logs requires skills normally acquired only by long, painstaking study. Encapsulating the skills in a master computer programme would mean very profitable business for Schlumberger. Dipmeter Advisor is a first step.
The ''knowledge engineers'' who worked on developing the programme drew on the expertise of Schlumberger's top log-interpretation man, Mr Al Gilreath. They sat with him for a year as he carried out his job, asking him to describe how he reached his conclusions and to codify his knowledge whenever possible. They then came up with a set of rules of the type listed above.
"Why can't a computer be more like a man?," The Economist, January 9, 1982
Knowledge engineers are suddenly in demand at dot coms, but the job title itself isn't all that new. It emerged in the early 1980s when the rage to create expert systems hit full force. Expert systems are a subset of artificial intelligence. They're computer programs that make decisions within a narrow field of study. Those decisions are based on a massive database of knowledge and rules gleaned from experts in the field. It's the knowledge engineer's job to codify this data so it can be used by the program.