Information Technology Reference
In-Depth Information
Very often, a global view of multiple XML schemas is necessary for data ware-
housing in a company. Therefore, we need to integrate their correspondent XML
documents for retrieving information.
1.4.4
Integrated ESs and Database Systems
Chapter 8 describes in detail the EDS technology. It presents a case study that il-
lustrates one EDS scenario, where existing DBs and ESs have been used to build
an EDS application. The consequent lessons are then addressed and some problems
of current techniques for the integration of ESs and DBs are explored. The “ideal”
future for EDSs using object-oriented technology are also discussed.
There are fundamentally different opinions coming from the current ES and DB
communities for EDS. The use of ES functions in DB products is to achieve “deduc-
tive data,” retrieve the semantics of data, and create an intelligent interface, integ-
rity constraints, etc. The use of DB functions in ES products is to represent factual
knowledge in the original knowledge base. These differences mean that current
EDSs have very different working environments.
Different approaches have been taken by various research projects and commer-
cial products to achieve the requirements of an EDS. They can be classified into two
different groups (see Fig. 1.6 ):
• Basedonexistingsystems:Therearefourdifferentarchitecturesinthisarea,i.e.,
enhancing existing database systems, enhancing existing expert systems, master-
slaver coupling of ES-DB, and peer-to-peer coupling of ES-DB. Most current
products can be categorized into one of these four architectures.
• Anewknowledgebase management system: This architecture involves search-
ing for a new model to represent knowledge. One example of this type of system
is Generis (Deductive System Ltd. 1988 ).
Reengineering functions and a high-level synthesis model are two main require-
ments for the future EDS (Huang 1994 ). These two functions cannot be traded-off
against one another. They can combine together to become a very powerful and
sophisticated EDS. Another interesting result is that both ES and DB researchers are
using object-oriented technology. It seems that most people currently believe that
object-oriented technology will become the future for EDS.
Chapter 9 is to normalize relational database and XML database. In general, an
information system is not user friendly if its database is not normalized. Data nor-
malization is to eliminate data irregularity for data update, and also for removing
data redundancy.
In XML documents, elements with same key values can be factorized into a pri-
mary element under root element. The result replaces the repeating data by pointers
and removes the data redundancy between these elements, and makes them easier
to be updated.
Search WWH ::




Custom Search