data / information modelling made difficult?

on this page historical background how we can help you other material information management modelling data modelling cribsheet (pdf) information modelling workshop home page contact us

Information modelling is vital for a broad range of software projects - from simple data storage to large complex data warehouses, and from new application developments to complex system migration and integration projects. Information modelling takes on renewed importance in the world of distributed objects, components and web services.  Information models can be found disguised as class models in UML, or as schemas in XML. The notations may change - but the same principles apply. Information modelling is like chess.  It's fairly easy to learn the basics - but it's hard to become a master. Most books and training courses give you simple examples, which don't prepare you for the difficulties of real projects. Our approach is to confront the difficulties.

How we can help you veryard projects > information management > information modelling  > how we can help you

	Training, Coaching and Mentoring	We can provide formal courses, individual tuition, informal skills transfer, and adhoc or ongoing project support.
	Model Reviews	We can assess the strengths and weaknesses of a model, and provide detailed guidance on improving and extending the model.
	Model Management	We can provide guidance on the coordination of multiple models, and on the selection and implementation of modelling tools and repositories.
	Information Strategy	We can help you develop strategic plans for information management across and beyond the enterprise.
	Free Access to Materials	Our website provides additional materials.

Historical background veryard projects > information management > information modelling  > historical background

Information modelling has been invented many times in the history of computing. The source is usually traced to a 1970 paper on the relational data model by Ted Codd, despite prior work by various people in IBM and elsewhere, including Young and Kent, dating back to the late 1950s.

Various line-and-box diagrams started to appear from the mid-1970s, pioneered separately by Chen and Ian Palmer (who was later to become Technical Director of JMA), showing entities and the relationships between them.  Notation wars broke out - some camps favouring crowsfeet or tridents instead of arrows, some camps favouring rounded corners on the boxes - as if it mattered. Lots of complicated and confusing conventions were invented to show interdependencies between relationships - dotted and half-dotted lines, dotted and half-dotted arcs.  (Can anyone remember what all of these meant?  Does anyone still use them?)

In the 1980s and early 1990s, information modelling was a key component of a number of systems development methods, including Information Engineering and SSADM. A number of modelling tools were developed, either stand-alone or part of an integrated CASE suite.

As fashions shifted away from large mainframe database systems towards distributed client/server systems, and away from traditional 3GLs and 4GLs towards object programming languages, information modelling seemed to be replaced by object modelling.

But was object modelling merely information modelling in a new guise?  Many of the pioneers of object modelling were strongly influenced by information modelling - indeed, you can sometimes find references in the object literature to my own book Pragmatic Data Analysis (published in 1984 and long out-of-print). And some flavours of class modelling seem to be almost identical in substance to entity-relationship modelling - only the names and the diagramming conventions have changed.

I do not want to try and reduce object modelling to information modelling - to claim that we were implicitly doing object modelling all along. There are some important aspects to object modelling that were never included in information modelling, and there are some significant theoretical and practical differences between objects and entities.  However, there are also many common elements.

There are two reasons why information modelling remains important today. One is that many of the ideas of information modelling are valid in the object world, with many of the same patterns and frameworks. The second is that there are many system development tasks that require a proper understanding of information and data structure - from the design of persistent data storage, to the design of interfaces and schemas for information exchange across different platforms.  The same principles and techniques and disciplines apply, whether the information structure is represented as a traditional entity-relationship diagram, a UML class diagram, or a BizTalk schema in XML.

top home page contact us
in asssociation with	This page last updated on April 26th, 2001 Copyright © 2001 Veryard Projects Ltd http://www.veryard.com/infomgt/immd.htm