PLM and The Collapse Of Complex Societies

I had chance to post about the issue of complexity in enterprise systems in general and specifically about the complexity of PLM systems. In my view, the complexity as one of the biggest problems in the development of systems for engineering and manufacturing these days. It comes constantly as a feedback from many customers and professional communities. If you had no chance to read it before, please take a look on one of my previous posts about dependencies between complexity of the systems and user’s adoption (Complexity kills or Three Ways To Improve PLM Adoption).

I came across a very interesting book during this weekend – The Collapse of Complex Societies by Joseph Tainter. It made me think again about the problem of complexity. Some thesis made by Joseph Trainer was the following:

[…societies become more complex as they try to solve problems. Social complexity can include differentiated social and economic roles, reliance on symbolic and abstract communication, and the existence of a class of information producers and analysts who are not involved in primary resource production. Such complexity requires a substantial “energy” subsidy (meaning resources, or other forms of wealth). When a society confronts a “problem,” such as a shortage of or difficulty in gaining access to energy, it tends to create new layers of bureaucracy, infrastructure, or social class to address the challenge. Eventually, this cost grows so great that any new challenges such as invasions and crop failures cannot be solved by the acquisition of more territory. At that point, the empire fragments into smaller units…]

I’d like to make an analogy between societies and enterprise and manufacturing systems. What is my take? Current enterprise systems supporting product development (CAD, CAE, PDM, PLM, ERP etc.) went through the long path of integration from early 2D to 3D, from MRP to MRPII and ERP, from EDM to PDM and PLM and finally rich the point where the almost cannot respond to the changes required and will tend to break again to the smaller pieces. Does it mean these systems will disappear? No, I don’t think so. However, I think, the new organization of systems can come to the enterprise and this new organization will have an ability to scale beyond the current level of possible complexity. Some interesting trends on this way:

SOA, Web and SaaS
The architecture of most of the enterprise systems was built 15-20 years ago with the state of mind of operation in the scope of a local department and/or company. They obviously outgrew themselves. To bring web experience including SOA architecture, Services and other technologies allowing operation on the global scale will be an approach that allows to bring large enterprise systems into smaller manageable pieces.

Global Data Models
As we had chance to discuss in the end of the last week, PLM data, PLM data models and identifications are still in the state that fundamentally assumes the possibility to build “unbreakable” standards and “single data models”. The data identification problem is a very complex issue, especially in case of multiple enterprise systems. We had chance to discuss it on the example of Part Numbering. This problem is real, hard and seems to me not resolvable in current systems.

What is my conclusion today? Current product development systems (EDM, PDM, PLM, ERP) are showing signs of over complexity. Following Tainter’s theory, they are trying to create news layers of bureaucracy and infrastructure. Industry best practices and out-of-the-box solutions are great examples of such new “organizations”. The potential solution will be in restructuring of these systems into smaller functional pieces, including the ability to handle globally scaled data and self-organized components.

Just my thoughts…
Best, Oleg



Share This Post

  • Awadhesh Parihar

    thanks for bringing up this discussion. Haven’t read the book, but Trainer’s theory seems profound, but extending it to comupter systems seems far fetched.

    Firstly I have few questions to ask on Trainer’s theory
    a) does it assume a static society or evolving one? what does it owe its complexity to, in meeting primitive needs or to evolve to higher functions? Am avid science fiction reader, would assume all of us are destined for interstellar travel one day :-)
    b) If it doesnt assume a static society, complexity is associated to ‘state change’ and hence the beaurocratic orders would invariably emerge. Basic human functions itself would change! Some socities are very efficient in crop production and do not involve as much human labour to produce the same, as machines. Human labor has transformed itself in beaurocratic functions or say ‘more evolved’ functions

    So, if enterprise applications have ‘modularized’ or ‘commoditized’ some functions so as to concentrate on higher functions, its just mimicking the world around us. I believe the complex socities are not complex in the level of thinking it propounds but the ‘organization’ of it. So, interplay amongst the various functional consituents of complex society will be simplified at the best..And in that sense, social networking and such concepts could already be seen as disrupting the complexity, but it cant be seen as ‘disintegration’ but even wider integration albiet using modern tools.

    And my hypothesis is applications will continue to be evolving as abstractions, hiding the complexity in them, modularizing and providing clean interfaces, much the same as any modern car still having the accelerator, clutch and brakes interfaces, but providing a very distinctive user experience

  • Douglas


    Another great read about collapse is the book by Jared Diamond with the same title: “Collapse”. It is about the reasons why certain societies collapsed in the past, societies like the Greenland Norse, the Easter Islanders, the Maya’s etc. The main reasons why these societies collapsed was not through over-complexity but because of continuing with out-moded and ill adapted practices under circumstances that changed.

    Ik personally think, and I have stated that frequently in your blog, that most of the basic practices in Product development and even Lifecycle management have been around for some time. It was in the Aerospace/Defense industry that these developments started for reasons of accountibilty and traceability. The society around us has changed in such a way thet these requirements are now applide to many more service and product types, making the process more complex. The process does not have to be more complex than that, but it shouldn’t be less complex either to be effective. To adapt Einstein’s phrase: things should be just as complex as is effective, but no more than that….

    I also feel that the tool builders tend to make things more complex than required from a process perspective and maybe so because they generally lack knowledge of industrial “good practices”



  • olegshilovitsky

    Awadhesh, I still didn’t come to the conclusion about static society vs. evolving one in the context of Trainer’s theory :)… My insight on that is that complex systems may have a tendency to develop a next level of the complexity, instead of becomes more granular. Web and Enterprise Software went opposite ways. These days, enterprise systems came to the situation when “changes of state” you mentioning are impossible. They are trying to build the “next complexity level”, but the only reliable option is to go in the opposite direction and become more granular. Best, Oleg

  • olegshilovitsky

    Douglas, Thank you for pointing out Jared Diamond’s book. I will take a look on that. I agree with you. There are lots of changes happened since first product development systems were established 15-20 years ago. Nevertheless, complex enterprise systems are continuing to build the vision of “a single point of truth”. I think it is a time to reconsider these practices. The cost of change and the complexity of enterprise PLM (and not only PLM) systems are huge. The existing trends that I see trying to change the current trends are 1/open source; 2/cloud; 3/out of the box or industry best practices. I’m expecting to see some interesting changes related to that in the future. Best, Oleg

  • Pingback: PLM Complexity: What Does The Future Hold?()

  • Pingback: PLM Complexity: What Does The Future Hold? « Daily PLM Think Tank Blog()

  • Pingback: COFES 2011 Rountable Discussions()

  • Pingback: COFES 2011 Rountable Discussions « Daily PLM Think Tank Blog()