ZapThink have recently been drip-feeding a series of ZapFlashes predicting the end of Enterprise Architecture (possibly allowing them to derive some further consulting/education business by creating a little fear, uncertainty and doubt). Here are some of their predictions and my thoughts. Comments welcome.
- The secret to 21st century software innovation. Is, according to ZapThink, in harnessing the power of complex systems (i.e. rather than just enterprise-wide systems). A complex system is one that shows some form of emergent behaviour where the properties of the system as a whole exhibit behaviour that the individual components do not. Most people don’t get this confusing systems that are just complicated with ones that are complex. For me this is spot on. As the enterprise morphs into basically the whole of the internet traditional governance models breakdown. The boundaries of the enterprise can no longer be nicely defined and who is in the enterprise and who is not becomes harder to pin down. People who get this and who can successfully harness the “interprise” (my new word of the month) by using the positive effects of social networking etc will win out over the next few years. Of course this involves huge risks not least of which around how will enterprises keep control of their data and intellectual capital.
- RIP enterprise software. Enterprise software as provided by the large package providers is seen as large and inflexible and not delivering on the benefits promised by SOA. Companies are finding themselves encumbered with expensive and hard to maintain software they can’t “bend” to do what they want. ZapThink’s take is that whilst enterprise software may have failed SOA has also failed to deliver on its promise of providing flexible business processes that can be quickly adopted to new needs. My take is that like everything else in our industry nothing is given chance to bed in before the next wave comes and sweeps everyone along with it. Many clients I see are still grappling with getting an effective process in place for developing traditional systems let alone service-based ones and definitely cannot deal with complex systems properly.
- The beginning of the end for Enterprise Architecture frameworks. Architecture frameworks (especially enterprise ones like Zachman and DoDAF) are counterproductive to developing an effective EA strategy. These frameworks are inflexible, sometimes encouraging what ZapThink refer to as checklist architecture. Checklist architecture focuses on achieving goals laid down by the architecture framework rather than the changing needs of the business. True but having no framework at all leads to even greater chaos in my experience. A framework is a structure which is meant to contain relevant guidance and work products to deliver real-world solutions. Frameworks that become too theoretical are always doomed to fail as they should. I hope this forms the basis of something more relevant to the real world.
So how would I characterise “interprise architecture”?
- Interprise architecture recognises that it is not enterprise architecture that is dead but trying to constrain architecture by the bounds of the enterprise that is no longer achievable. The ‘architecture’ bit still applies, but not the ‘enterprise’ bit.
- Governance models need to recognise that this extended enterprise includes people and other systems that are not controlled by the IT department (people using social networking software who will periodically ‘overlap’ with the enterprises systems).
- The ‘architecture’ bit needs to take into account the fact that systems are complex (in the emergent sense of the word) and it’s not always possible to tie down the requirements in a nice orderly way. Indeed fully defining the requirements may be counter-productive and not allow emergent behaviour. Processes for developing such complex systems need to take this into account.
To be sure this is not a complete list. This idea is not yet fully-formed in my own mind and needs a bit more time to ’emerge’ properly. Watch this space as they say.
Software Architecture Zen 
Peter,
I am agree with your “interprise architecture” view completely. In fact, I have been thinking about a tool for architects to manage this “emergence” effectively – what is generally known as cognitive effectiveness. In the complex and un-ordered problem space, architects need to avoid falling into the trap of making the same decisions which have brought success in the past. Instead, they should exploit the requisite level of diversity in this space to sense weak signals (terrorist threat or market opportunity) through interventions or probes, and monitoring their impact to allow the emergence of a beneficial future state.
I believe I need to start with an ontology that would allow an architect to specify rules of interactions between peers in the eco-system but not bind them in overly rigid structure as EA models do today (which only works in the ordered and predictable space).
What do you think?
Agree absolutely. Architecture is too predicated on 'understanding the requirements' and as we all know this invariably does not work because the requirements are only fully understood when the system is built by which time it may be the wrong system or time has marched on and the business needs have changed. For me two things are needed:
1. A way of capturing what the users do with the system once it is at least partially created and evolving the system to meet those needs.
2. Not rigidly defining the boundaries of the enterprise but allowing external influences to influence the way the system works. For example allowing the wider network to use the system and possibly adapt it to different ways of working. This has been the way all the successful social networking sites such as Facebook and World of Warcraft have evolved.
Thanks for your comments. Feel free to email me if you want to discuss any of these ideas further.
Peter, on complex systems, please see Ultra-Large-Scale Systems: The Software Challenge of the Future at http://www.sei.cmu.edu/uls/.
Love the term “interprise architecture.” We (myself and fellow EA analysts at Gartner) came up with a similar term: “outside-in architecture.” The note defining “outside-in architecture” is behind the paywall (www.gartner.com/DisplayDocument?id=1358516), but other descriptions are not. See http://bit.ly/degLrp and http://bit.ly/cctN3e . Other possible terms might be “industry architecture” and “ecosystem architecture.” I will be watching this space!
Haven't been able to view your “Eight Hybrid Thinking Principles…” document but anything that has “wicked problems” and “design thinking” in the abstract presses my hot buttons so I'll make sure I get hold of a copy. The other link did not work for me.
[…] a previous post I introduced the term “Interprise Architecture” to describe how the internet is breaking down […]