Friday, September 28, 2012

Beyond Multiview

#entarch #bizarch #UnicomEA #SSM #VSM #VPEC-T At yesterday's EA Forum, a rich picture of the oil industry was presented by Mesbah Khan, who then posed a question about the possible use of such a picture, for enterprise architecture and beyond.

The rich picture is a core technique of the Soft Systems Methodology (SSM). Over the years, there have been many attempts to produce hybrid methodologies combining SSM with more structured systems approaches. One of the earliest such attempts was Multiview, produced by Trevor Wood-Harper and others in the mid 1980s. In Multiview, a version of SSM is used to analyse issues, and these are then aligned with the analysis of tasks using a more conventional IS modelling approach.

Among other things, Mesbah's rich picture model features stakeholders and their concerns (shown as bubbles), and identifies the conflicts between different stakeholder concerns (shown as crossed swords). This is clearly equivalent to Multiview's analysis of issues.

Nowadays, the terms "stakeholder" and "concern" are mandated by ISO 42010, and incorporated into several enterprise architecture frameworks. (However, stakeholder concerns are often presented as a relatively homogeneous and consistent set of goals and objectives, e.g. using the Business Motivation Model or other schemata, and there is often not enough attention given to the conflicts between stakeholder concerns.)

As I see it, each of these conflicts calls for a series of capabilities or activities to govern the conflict - to allocate resources, balance priorities, and contain the risks. (However, these capabilities and activities are often absent or marginalized in the structured models produced for IT purposes.) They also have important implications for organizational design and trust.

One way of looking at these management and control capabilities is to use a cybernetic view of the enterprise, such as Stafford Beer's VSM. Mesbah indicated that his rich picture can be mapped onto VSM, but he did not have time to present this mapping at yesterday's forum.

Meanwhile, the trust issues may require a different style of analysis, such as VPEC-T, which helps to highlight those issues that are typically "lost in translation" when we go from a "rich" model to more abstract and homogenized "structured" models. See my note Does Rigour Matter?


ivo said...


I agree about VSM. But I don't understand how is trust related to getting from " a "rich" model to more abstract and homogenized "structured" models". Could you please explain.

"Trust" is used in a sense closed to "Belief", as well as to "Fidelity" (and both are roots in to the etymology of trust). Trusting a model could be seen from the perspective of relating experience, assumptions and beliefs as in the cycle of inference (here's my take on that). The other perspective, "fidelity", is either related to the source, and so again beliefs, or to some intrinsic characteristic of the model. The latter has something to do with the perception of "rigour", which was addressed in this and the previous post. But that is also a matter of belief. For example it is believed that a formal language like BPMN is rigorous because, among others, it's metamodel is described in UML which also is believed to be rigorous and it's wide adoption reinforces that belief. In fact it is far from rigorous which can be easily proven using the set theory but that's not the point here. The trust in UML has been fueled by the trust in the OO paradigm. Not by the fact that something is provable. So we are back to Beliefs again.


Richard Veryard said...

Ivo, thanks for your comment. I wasn't talking about trust in the model, although I agree that is important and interesting, but trust in the relationships and interactions between the actors in the model. The conflicts in concerns, motivations and obligations, as identified in Mesbah's rich picture model, almost always raise trust issues. In the Oil Industry case, there is a fairly elaborate superstructure of contracts and subcontracts, whose existence is necessitated by the fact that the parties don't fully trust one another - and shouldn't be expected to. So an analysis of the trust issues helps to explain, improve and perhaps re-specify what we might call the contract architecture. I alluded to VPEC-T as a method that could be used to explore these issues further, but as far as I know this method has not been applied to the Oil Industry case as yet.

Ivo said...

Unfortunately I'm not aware and cannot comment on Mesbah's model. But I think that trust can be represented in both rich pictures and in more simulation-capable models like stock-and-flow diagrams and agent-based models. As I’ve used these and they helped in understanding WIGO – to use your term – and what to do about it, including variable such as “trust”, I developed a trust in the capability of these approaches. The book “Lost in translation”, although a valuable resource, somehow failed to build the trust that VPEC-T can help in dealing with trust issues. I’ve either missed something or the book itself is not enough. In any case I’m willing to learn. Can you please give some examples of applying VPEC-T? Thank you!