The SuperManual interactive electronic technical manual
Peter W. Foltz and Thomas K. Landauer
Abstract
Maintenance of increasingly complex technological systems is a critical
and difficult challenge for defense,
government and private sector organizations. The quantity and technical
complexity of systems is constantly
changing and expanding. Increasingly, personnel trained or experienced
on one system must quickly switch to
working on another. The overall goal of the SuperManual project is to
design and prototype better ways to
dynamically organize, present, and customize information for particular
tasks and individual maintainer knowledge
and levels of expertise. The SuperManual is a personally adaptive
electronic maintenance manual (IETM) that is
designed to permit personnel to rapidly locate and learn material from
maintenance manuals.
SuperManual leverages off of two technologies. One is the powerful
capabilities of Latent Semantic Analysis
(LSA), an automatic machine-learning technology that accurately
simulates human judgment of the similarity of
meaning of two texts. Among other things, LSA is used to improve search
based on natural language queries and
semantics (not just the key-words) of target text, and for determining
the optimum sequence of texts to provide just
the right material for supporting a given task for a given individual.
The other major technological basis is the
usability tested and application-proven functionality and features of
the SuperBook hypertext browser. Developed at
Bellcore in the early 90s, SuperBook increased speed and accuracy on
information-dependent tasks by large factors.
The basic SuperBook has been updated with other and newer techniques
for presentation of text and graphics.
Because SuperManuals are produced virtually automatically from existing
instructional and operations-support
texts, they can be constructed much more quickly for new systems, and
at much lower costs than current electronic
manuals.
Full Paper (PDF)
|
