I've Been Thinking
Bernard Robertson-Dunn
I am not a seeker after truth ... just evidence. I will draw my own conclusions. 
  • Architecture
  • Problem Oriented Architecture
  • Problem Solving
  • Modelling
  • Problems and Solutions
  • Systems Thinking
  • Bernard's Parables
  • Projects
  • Complex Systems
  • Humour
  • Bushfire in Canberra 2003
  • Pictures of Surbiton County Grammar

  • Site Map
  • About Me

    Last updated February 2014

    Contact me on brd@drbrd.com
    I've started to  think rather more about modelling and the modelling process. Most of my thinking so far has been done in my blog but here's an excerpt.

    On 11 February 2014, I received an IEEE call for papers for UKSim2014, the 16th International Conference on Mathematical/Analytical Modelling and Computer Simulation. Submissions were due on or by 15 February. Conference website: http://uksim2014.info

    In a mad panic I started to write a paper on nonlinearity, a topic that I think does not receive enough attention from the modelling community. I did a quick analysis of papers presented at the UKSim-AMSS conference in the years 2008 to 2013 and found that that fewer than ten out of over seven hundred papers contained the word nonlinear in their title – hence the allusion to an elephant in the room.

    Nonlinearity – The Elephant in the Model and What the Elephant Can Tell Us

    To cut a long story short, I came to the conclusion that the problem is not that nonlinearity is being ignored, but the whole subject of modelling is not being studied. Many people are using modelling, but IMHO, badly.

    I submitted a paper, more in the hope that it would resonate with the good people at UKSim2014, rather than offering it as a well thought through result of rigorous research. My hope is that it initiates a conversation that leads somewhere mutually beneficial.

    What it has done, though is to stimulate me into looking into modelling, and the subject of nonlinearity, myself. This relates to problem solving because of three things:

    1. Problem Solving involves different forms of modelling – Behavioural, Predictive and Design.
    2. The world is nonlinear and unless you accommodate nonlinearity properly, your predictive models will quite possibly mislead.
    3. Modelling, and Problem Solving need to be treated as processes that require controlling.

    Anyway, for reference, this is the abstract of the paper I submitted.

    Relationships in the real world are generally nonlinear. Relationships between reality and models of reality are also non-linear. Models of reality may or may not be nonlinear. This paper briefly looks at nonlinearity as it impacts modelling. Particular attention is drawn to the failure of predictive modelling, a failure that is at least in part due to a lack of understanding of nonlinearity.

    It is then suggested that the elephant in the room, nonlinearity, is not the underlying issue but problem is a lack of a broad understanding of models and modelling as a discipline in its own right. It is suggested that what is needed is the creation of a science and engineering based discipline that can provide a solid theoretical underpinning of modelling across multiple disciplines.

    See my blog for further information. The site at www.problemsfirst.com is probably where I'll put more of my "models of models" work, so head off over there.

    Before all that I was busy writting about Problem-oriented System Architecture. I started with a chapter in a book:

    "A Problem Oriented Enterprise Architecture Approach Applied to Wicked Problems"
    in Enterprise Architecture for Connected E-Government: Practices and Innovations, editted by Pallab Saha of Singapore University.

    This is a link to a description of the book.

    This is the abstract of my chapter:

    This chapter proposes that a problem oriented approach to Enterprise Architecture can deliver a better outcome than one based upon needs and requirements, especially when dealing with Wicked Problems.

    A distinction is drawn between what an Enterprise Architect does, solve business problems, and what the architect produces, descriptions of end states.

    It also suggests that the approach to modelling and understanding a problem can have significant impacts on the quality, effectiveness and efficiency of potential solutions and the decisions made in identifying optimal solutions and implementation projects.

    Finally, the chapter discusses use of the proposed problem oriented Enterprise Architecture approach to Wicked Problems in the context of e-Government.

    Then I wrote a paper "Beyond the Zachman framework: Problem-oriented System Architecture" for the IBM Journal of Research and Development, details of which are here

    I've created  a new website, www.problemsfirst.com, which will eventually contain lots of information on why large scale IT projects keep failing and what to do about it. Its based on the premise that understanding the problem is what really matters. And as I've said above, it's where my modelling research will end up.