Pete Lacey’s Weblog

The problem is caused by the root culture of IT — project-driven funding models, a cobbler’s kids perspective on investing in infrastructure that helps IT (rather than a particular project), and a propensity to never decommission applications. IT systems have grown organically for the last 40 years. They’re a mess. It requires a fundamental change in the way IT operates as a service provider within the organization.

Now Anne’s my new hero (of course, we work together). This hits the nail on the head. Business is culpable. IT is culpable. I’m culpable. So are you. I would refine Anne’s point by saying that the ailments we’re discussing are not entirely cultural. I would add two other factors.

1. A sensible desire on the business’s part to maximize its investment. To not fix what isn’t truly broken. If a business in 1990 spent $2M and 2 years building a widget management system, and that system was perfectly designed and executed, then there’s an understandable reluctance to build it again. Similarly, from the IT perspective, if there’s an existing system that used to meet all of your needs but now only meets 75% of them, it seems sensible to extend the system, rather than rebuild it. It’s the same thinking that has me maintaining a house that’s over 100 years old, even though it leaks heat like a sieve and has antiquated wiring on the second floor. (You’re not allowed to beat up on that analogy.) This issue is deeply impacted by the incredible rate of change in technology, and all the things no one saw coming: PCs, universal networking, the Web, open source, etc.

2. People and groups of people are independent actors. They have their own biases, knowledge base, desires, needs, and motivations. Any time an overarching strategy tries to unify all the disparate players, it comes into friction with this, slowing it down and ultimately causing it to stop.

I suspect that there’s no way to make all these things go away, and if we want to drive better business through technology, our planning has to account for these three factors (Anne’s cultural issues plus my two). It’s likely that the cultural issues can actually be fixed. We can change funding models and processes. We can even effect a change in mindset. However, the other two seem to be intrinsic to the human condition, and I think our planning is simply going to have incorporate that. Thus, we will need to discover processes and technologies that allow systems to be built at minimal costs (time and dollars) and that can, in effect, be thrown away. And we need to allow IT and the business to act as independently as possible from some central governing authority. A very delicate balance in both cases.

From a purely technical POV (and recognizing my own biases), it seems to me that we can partially address the cheap-fast-and-gone issue by moderating complexity. This might entail things such as promoting dynamic languages; building smaller, minimally functional components (using your favorite technology, but erring towards the simplest); hiding the brittle things behind facades; making strategic bets on very few technologies or technology patterns, and so on. Regarding the people-are-people thing, I think this means that we cannot dictate many universal behaviors. We can only strongly encourage (preferably by example) that players do what they can to minimize friction between themselves and others (technology wise, that is.) via the use of standards and by the use of system designs that allow actors to evolve independently.

What that doesn’t address is how do we get IT to do what the business wants. Personally, I think IT does a pretty good job of that already. What seems to be the issue is that IT is building what business units want today. Both IT and the business units are not planning for tomorrow, nor are they are thinking about the rest of the business. And this is where competent CIOs and their minions, business analysts such as Steve [Jones] and Rob [Eamon], industry analysts such as Anne [Thomas Manes] and Nick [Gall], and consultants such as myself need to deliver on. That is, given the constraints above: project driven cultures within IT and without, rapid technological and business change in the face of sunk costs, and the fact that the enterprise (indeed the world) is an anarchic place, how do we get people to build systems that meet the needs of today and tomorrow? You can call this enterprise architecture if you want.