Previous | Table of Contents | Next |
The previous sub clauses focused on the use of the MOF to support the software development life-cycle and the type management
requirements of CORBA infrastructure services. This sub clause broadens the scope to the more general domain of information
management; that is, the design, implementation, and management of large bodies of more or less structured information.
First, note that some of the ideas outlined above carry over to the information management domain. In some cases, it may be
appropriate to define the information model (that is, the database schema) for the application of interest directly using
the MOF Model. In this case, the technology described previously can be used to automate the production of CORBA-based servers
to store the information and applications to use it. In other situations, the MOF Model can be used to define a metamodel
suitable for defining information models for the domain of interest; for example, a metamodel for describing relational database
schemas. Then a development environment can be designed and implemented using MOF-based technology that supports the generation
of CORBA-based data servers and applications from information models.
In addition, the MOF potentially offers significant benefits for large-scale information systems by allowing such a system
to make meta-information available at run-time. Some illustrative examples follow.
Information discovery: The World-Wide Web contains a vast amount of useful (and useless) information on any topic imaginable.
However, this information is largely inaccessible. In the absence of other solutions, current generation web indexing systems
or search engines must rely on simple word matching. Unless the user frames queries carefully, the number of “hits? returned
by a search engine are overwhelming. Furthermore, it is now apparent that even the largest search engines cannot keep pace
with the Web’s rate of growth.
In the absence of software that can “understand? English text, the approach most likely to succeed is to build databases of
meta-data that describe web pages. If this meta-data is represented using MOF-based technology and an agreed base metamodel
for the meta-data, the framework can support local meta-data extensions through judicious use of MOF-supported reflection.
In addition, because the meta-data framework is defined in the MOF context, it can be accessible to a larger class of generic
tools.