Conclusion - Information Systems Reengineering, Integration and Normalization - page 376

certain user's data requirements such as many-to-many cardinality among elements.
Therefore, fragments minimization is a data normalization for XML database de-
sign.

10.4 Future Trends for Information System

Reengineering

The main idea for information system reengineering is to reuse the existing knowl-
edge as opposed to simply the reuse of data. The techniques for knowledge reuse are
extremely important not only because they aid in building an information system,
but also because they help to improve the reliability of the information system.
This topic provides a systematic approach to reuse existing information systems.
Since the existing system may not be perfect and may be partially nonproductive,
it may be necessary to reuse only certain parts of the existing system, but not all.
To reuse knowledge, we must know its structure. Current knowledge repre-
sentation structure has multiple frames. Data modeling from database research
and knowledge representation from artificial intelligence both still have difficul-
ty representing the knowledge completely. A distortion exists between the real
world and the information system. It is extremely difficult to recapture the original
knowledge from the existing information systems. To solve this problem, a heu-
ristic approach has been taken by computer scientists. This approach is to use an
expert system to assist the system developer to recapture the missing knowledge
or semantics.
Another approach for knowledge reuse is to define a standard specification for
the information systems. In spite of the economical success of reengineering ap-
plications, some problems have been detected in using this technology, the largest
problem being the lack of agreed standards for information systems. For example,
there is no standard for the object-oriented technology. Providing standards for
information systems is a way of supporting reengineering, partly because it can
provide portability and transparent communications. Some work on high-level stan-
dards, sometimes referred to as the knowledge level, has been carried out. One
example of knowledge level representation is the language developed in the Knowl-
edge Acquisition and Documentation Structuring (KADS; Tansley and Hayball
1993 ) methodology for analyzing domain knowledge. KADS allows developers to
build libraries of inference models for specific domains (for example, diagnosis).
Computer scientists are now looking at providing a similar approach for sorting the
content of knowledge bases in a reusable way; these reusable knowledge bases are
called Entologies'.
The object-oriented paradigm has been seen as the most common technique for
the conventional software and knowledge base reuse. The object-oriented technol-
ogy is still growing.
Data is a collection of “fact.” Information is the meaning of data. Knowledge is
the application of the information. Knowledge is also a necessity of reengineering.