Information Systems Reengineering, Integration and Normalization - Information Systems Reengineering, Integration and Normalization - page 18

be the remedial action? How can the existing system join the system analysis and
design phase? How do we test this type of system during the development lifecycle?

1.4 Approaches to Reengineering

Reengineering information systems involves reusing the existing outdated database
systems and ESs by upgrading or integrating them to new technology systems to
meet the new users' requirements. Database upgrading, in a logical sense, is to up-
grade an old database technology, i.e., one using a hierarchical or network model, to
a new database technology, i.e., a relational, object-oriented, or XML model. Reus-
ing an ES can be accomplished by integrating it with a database system.

1.4.1 Database Reengineering

Database reengineering consists of three parts: schema translation, data conversion,
and program translation. It can be described as follows:
In schema translation, there are two approaches:

• Directtranslation—Onecandirectlytranslateanonrelationalschema to a rela-
tional schema. However, such translations may result in the loss of information
because of their primitive mode of operation that cannot recover or identify all
the original nonrelational schema's semantics. Certain advanced semantics are
lost once they are mapped from a conceptual schema (e.g., ER model) to a logi-
cal schema (e.g., Hierarchical or Network schema). Thus, users' input is needed
to recover the lost semantics.
• Indirecttranslation—Indirecttranslationscanbeaccomplishedbymappinga
logical hierarchical or network schema into a conceptual ER model schema in
reverse engineering. The translated conceptual schema must have all the original
logical schema's semantics. User input can be used to recapture the semantics
of the conceptual schema. A knowledge base can be used to support the process
of recovering such semantics. Then the conceptual schema can be automatically
mapped to a relational schema. Similarly, in order to translate a relational sche-
ma to an object-oriented schema, we can map the relational schema first into the
ER model, then into a Unified Modeling Language (UML; Booch et al. 1999 ),
a conceptual model for object-oriented model, and finally translate the UML
model onto the object-oriented model of the target database. Similarly, we can
map relational to XML model through DTD graph and XSD graph.

Chapter 3 describes in detail the methods for schema translation.
In data conversion, there are three approaches:

• Physicalconversion—Thephysicaldataofthenonrelationaldatabaseisdirectly
converted to the physical data of the relational database. This can be done using