Sunday, October 16, 2005

Zachman Lecture

I attended a two day seminar on the Zachman framework for Enterprise Architecture last week. I have read about Zachman before, and have even bought a book on the subject (which as an aside must be the work book ever written). However the seminar was presented by John Zachman himself so I thought it would be a good opportunity to get the message from the horse's mouth.

I was not disappointed; I have always thought of the Zachman framework as just being a means of classifying the various artefacts created during the development and maintenance of an enterprise's architecture. However John's compelling vision is that the Zachman framework is a schema for defining the primitive elements of enterprise architecture. He describes the framework as being the basis for enterprise physics, and draws an analogy with the periodic table. This analogy also allows him to justify the fact that not all of the models that comprise the cells in the framework have been articulated yet. Continuing the analogy, he argues that any architectural artefact needed will either be one of the primitive models in the framework, or a composition of these primitive models. By separating out the independent variables in the enterprise (represented by the columns, defined by interrogatives) the enterprise can support the flexibility needed in the information age.

One of the points made repeatedly during the two days was that in the absence of architecture models there are only three ways to support change:
  1. Change by trial and error
  2. Reverse engineer the models
  3. Scrap the legacy solution and start again
Another interesting observation concerned the use of COTS; here the advice was to change the organization and/or business processes to fit the COTS package, and not vice-versa.

John's presentational style was very interesting; it was not a seminar in the sense of dialogue and discussion. It was more a high-powered intensive lecture, with a huge quantity of facts, knowledge and anecdotes delivered at breakneck speed.

All in all, it was an excellent lecture to attend. I left with a feeling that I need to change the way that I think about architecture, which is all I could really ask for.

3 comments:

lcne said...
This comment has been removed by a blog administrator.
lcne said...

Hi Paul,
I liked your blog on achman. Since I am working on a project - along with Rune Raunow (systematic.dk) - where enterprise architecture will a key issue, I would like to know your opinion on a matter, that have intrigued many since Zachman published his work in 1987.

Even though Zachman stated in the lecture, that this is a schema. Do you regard Zachman as a enterprise architecture framework, the foundation/basis of a framework or primarily a classification?

With regards,
lcne

Paul Mukherjee said...

This is an interesting question. Zachman himself thinks of his framework as a means of classifying the primitives that are needed to describe an enterprise architecture (like the periodic table of elements). However it does not provide any guidance in how to populate the cells in this classification. So in that since it is purely a system of classification.

If you want a method that provides guidance in how to populate the framework, then you need to consider a framework such as TOGAF.

My own feeling is that enterprise architecture is still an immature discipline, and no one framework provides all of the answers. Any specific enterprise needs to look at the different frameworks available and customize one to their own needs.