What’s wrong with IFW?

The Information FrameWork – Architectural Framework

I created Information FrameWork (IFW) as an architecture framework – but IBM seem to have lost sight of this over the years, and now the phrase is used as the collective marketing term for a set of industry models.

This is from a comment I made to a discussion on LinkedIn: What is wrong for IBM IFW?

As an architectural framework, in 1990 IFW was a radical departure from the Zachman Framework. Inheriting the innovations that Westpac’s CS90 pioneered, it was the first architectural framework designed to explain a holistic, building block approach to architecting an enterprise. Its population by a set of industry reference models was way ahead of its time – the basic concepts of the data model, inherited from Westpac, have remained constant for over 20 years, and the function model was effectively the first business capability model.

Here are four fundamental things that could be better:

  1. Make a clear distinction between IFW as an architecture framework and as a marketing label for a set of industry models. This would not mean going back to the original 1990s architectural framework, but taking the fundamental architectural principles and ideas and updating them for contemporary needs. Why is this important? Because without a clear statement of the underlying architectural concepts its very difficult to get the full potential from the models….
  2. Revive the distinction between the organization, business and technical views. This was a key aspect of the original framework. Why is it useful? Because it makes it much easier to integrate and modularize the models….
  3. Modularize the model materials into genuine Architecture Building Blocks. The models have always been packaged into a number of large chunks because they fetch a premium price in this format. Architecturally the intent was always that they were modular components that combined to form a complete enterprise architecture. Why do this? Well it means that you only have the building blocks that you need. It also makes it easier to separate the general from the specific – to separate general data concepts such as Involved Party or Product from industry specific ones…
  4. Update the architecture and models to keep them relevant! It is always easier (in some respects) to fix things at a point in time and never change them, but this gradually erodes their value and relevance.