A busy week this has been. The Gartner EA Summit and the Open Group Enterprise Architecture Practitioners conference were both on in London at the same time, little more than a few hundred yards apart. And a lot of other things starting to happen in the enterprise scene as well: more good news on the way.
The highlight, though, was a stream of great conversations on enterprise-architecture.
The first of these was with Nick Gall and Bard Papegaaij from Gartner, and independent-consultant Richard Veryard. As usual, I failed to take notes… apologies. But probably the key theme throughout was the shift away from IT-centrism: Nick with his concept of the panarchy double-cycle applied to architecture, as ‘panarchitecture‘; Bard with a strong emphasis on architecture in government, and on emotional-intelligence and human factors (the latter with some strong parallels to my own themes around ‘enterprise as story‘ and ‘enterprise as language‘); and Richard on expanding out to a broader concept of ‘organisational intelligence‘. There was also quite a bit of discussion on whether the panarchy-cycle of creation, exploitation, collapse and rebuild, could be applied to Gartner’s own ‘hype-cycle‘: Nick was adamant that it couldn’t and shouldn’t, but Richard and I both felt that perhaps it could – particularly if we see the hype-cycle as two iterations of a panarchy-cycle, with the hype-cycle’s collapse into the ‘trough of disillusionment’ representing the second half (‘destruction’) of the first of those panarchy-cycles. A discussion for another time, perhaps?
We followed through on the next day with a stream of what ended up as mostly one-on-one discussions: Richard was presenting at the Open Group conference and could only drop by for a few minutes, whilst Nick and Bard had speaking-slots at Gartner that were almost back-to-back.
With Bard, the conversation started around the work by his late wife Michal, linking the native-American model or metaphor of the Medicine Wheel, and how those concepts can be applied in a business context. In some ways this parallels my own architectural use of the traditional Five Elements model, and also some Jungian-style concepts that I’ve used for many a year, though Michal’s ideas seem to go into even further depth. (We’d planned to meet up at their home in Brisbane earlier this year, and had all been greatly looking forward to it; but we’d had to postpone at the last minute, because she became very ill, and sadly that meeting never took place. A huge loss not just to Bard but – from what I’ve seen so far – a huge loss also to all of us in enterprise-architecture, I suspect. Oh well.) Very interesting, anyway, and I hope at least some of it will surface as a Gartner Note or the like from Bard in the relatively near future.
Another key part of the discussion with Bard was the relationship between agility and stability, somewhat as described in my previous post ‘Agility needs a backbone‘. The hypothetical example that we explored – based on real-world contexts which we’ve both worked – was the classic clash between bureaucrats and politicians in a government department. The blunt fact is that few politicians can see beyond the short-term: they need to deliver quick results of some kind to convince the electorate that they’re doing something of value. That means that they demand agility, to change everything ‘now!’ – which soon leads to a horribly fragmented architecture, with all manner of half-completed projects pulling in all manner of different directions. By contrast, the bureaucrats crave certainty, stability – and they often do hold a true long-term view, albeit often an over-cautious one. Caught between these two opposing forces are the project-managers and process- and IT-system developers, who somehow have to sort out the resultant mess. The way out seems to be an architecture based on some variant of the backbone – which keeps the bureaucrats happy – providing consistency and support for a myriad of smaller agile projects out on the edge – agile enough to keep the politicians happy. The two types of implementations need different emphases: the agile side typically thrown together as ’shadow IT’, whilst the core follows a more ‘traditional’ waterfall-style cautious-change model with much tighter governance. New services feed outward from the core, enabling new agile-style ‘mashups’ – the many GIS-linked ‘citizen services’ being a good example of this. And some of those quick-win services will also slowly migrate into the core. But in terms of dependencies, it’s a kind of spoke-and-hub relationship: in general, services from the core should never be allowed to break anything – especially not without warning – whereas there would often be no guarantees at all for relationships between agile-services out on the edge. This approach would give us a unified form of governance across the whole agile/waterfall spectrum – and a lot more certainty for the developers who’ve too often been caught up as pig-in-the-middle.
Then to the follow-up meeting with Nick Gall. Much of this was a review of what Bard and I had discussed earlier, but there were also two key points that arose from a brief review of my ‘Enterprise Canvas‘ model-type (from my book Mapping the enterprise). One point was a link-up between my understanding of the tension between ‘vision’ and the real-world, that drives the architecture, compared to one of Nick’s own models of architecture as a kind of wasp-waisted ‘double-funnel’ between near-infinite possibilities and near-infinite implementations, with architecture as the ‘waist’ where constraints for key choices are identified and applied. To me, everything in the enterprise is like a cone, extending downward from the single point represented by the vision. But as Nick pointed out, architecture describes a structure that could in principle be used for a very wide range of different purposes – in other words, similar structures that can support different enterprise-visions. The ‘cone’ represented by a layered Enterprise Canvas would thus be one instance of the range of possibilities of purpose represented by the upper-half of Nick’s double-funnel, selected out by that specific vision; a different vision could well lead to an almost identical-seeming implementation below the ‘waist’ of the double-funnel. Hence why reference-architectures and commoditised-services and suchlike do actually work in practice – even though they’re linked to different enterprise-visions.
The other point was an easy way to resolve the age-old argument about architecture versus design. They’re actually part of the same spectrum from vision to realisation, from ‘why’ to ‘how’ and so on. The only difference between them is which way they face: architecture tends to face ‘upward’, towards the big-picture, the vision, or ‘why’, whilst design tends to face ‘downward’, towards the detail, the real-world realisation, the how and who and where and when and with-what. So in practice, almost no-one is ever solely and architect or designer: everyone will do at least some of both. What makes it confusing at times is that the ‘architect’ orientation at a detail-layer – a solution-architect or application-architect, for example – will usually have a narrower scope than someone nominally working in higher-layer business-design or process-design. Once we realise it’s the same spectrum, it makes things a lot easier to explain: the difference between architect and designer is one of orientation – ‘up’, or ‘down’ – on that spectrum, more than one of position in terms of Zachman-style layers. Architects mostly architect, and designers mostly design; yet the two roles will always meet somewhere within each person on that spectrum.
Next was a meetup with a director of the vendor of a mid-range enterprise-architecture toolset. I won’t say which vendor it was, for confidentiality reasons, but to me this was important: perhaps the first toolset-vendor to really ‘get’ the nature of whole-of-enterprise architecture, and the support that it needs from the the architecture-toolset. Like almost all of the vendors, they’ve come up from an IT-oriented base, and that’s still the core of their toolset; but they do understand about how all of that links upward into strategy and vision, and horizontally across the non-IT aspects of the enterprise – people and machines and non-IT assets and the like. Nothing else to report just yet, but definitely a Watch This Space, I think?
Leaving the Gartner conference-venue, a very brief meeting with two of the new generation of whole-enterprise architects, Gerold Kathan and Ondrej Galik. I had to run to catch a train at that point, so only enough time to talk whilst crossing Westminster Bridge, but good to know that the future of the field in Europe seems already to be in capable hands.
And likewise in Latin America. The last of this stream of meetings this week was with Roberto Severo, lead for the Brazil chapter of AOGEA (Association of Open Group Enterprise Architects). We met first at the Open Group conference-venue – I didn’t go to the conference itself, for reasons I’ve explained earlier. A long, rambling walk-and-talk through central London, covering a very wide rantge of enterprise-architecture topics – in particular, how to expand and embed whole-enterprise architecture ideas and techniques in the Latin market. One of the best ways to do this will be through a stronger emphasis on values, which aligns better with Latin culture than it does to, say, British or (especially) US business-culture, where – as I’ve discovered to my cost – it’s often very hard to get business-folks to understand any concept of value beyond money or the near-mythical ’shareholder-value’. There’s still a constant struggle to combat the baleful influence of IT-centrism in enterprise-architecture (and I’ll have to be blunt here and say that for the most part the Open Group and most of the big consultancies are really not helping us in this… ), but it’s probably somewhat easier to resolve this in Latin America than in mainstream ‘Western’ cultures. We’ll see: but it certainly looks like an interesting year ahead.
Interesting times, anyone?