The Concept of a Data Factory is an excellent Framework
This is a Discussion Paper intended to establish a common view-point and to
promote exchange of ideas in reaching consensus.
Here is a working definition of a Data Factory for discussion :-
"A set of Human and Automated Components which combine to meet a variety of
requirements for Data and Information."
The approach involves establishing the requirements and determining the best
way to meet the requirements as part of a long-term strategy.
This means identifying a Methodology which contains the ability to evolve without
The Approach is :-
1) Analyse the work involved in 'Business as Usual'.
2) Identify cost/benefit of designing and building Components
3) Assemble Components Toolkit
4) Define standards for Glue,(e.g. XML), and build basic Interface Glue.
Establish the Methodology to repeat Steps 1 to 4.
3. Three types of expertise
1) Writing Interface 'Glue'
2) Creating new Components
3) Assembling existing Components to meet new Requirements.
The long-term Methodology provides a Road Map for the future.
1) Receive new request.
2) Review available Components
3) Define new requirement, if appropriate.
5. The Data Factory Toolkit
The Toolkit must provide for :-
a) Controlling Procedure Behaviour
b) Co-ordinating related Components.
c) Exception handlers
The corresponding Toolkit contents include :-
1) Business Rules
2) Shell Scripts
Here are some useful References :-
||A review of a very useful part of the Data Factory.
||Steps in a general Approach.
The Database Factory
||An excellent book by Stephen Schur.