The slow death of government IT

I read the wired article on the lateral move of Todd Park from US Government CTO to technology advisor to the whitehouse. It's a fascinating deep dive into the mind set different between modern technology companies and our political administrations. What's interesting is that there has been some leadership down south that has yet to materialize here. And the question is why?

The article, and my experience, cite a mentality rooted in caution, security over experimentation and adherence to bureaucratic procedure as the major causes of the current failures. Fundamentally, thought, for me the issue is rooted in the government's definition of fairness. Fairness, you might ask, how could that be a bad thing? Well, much like I tell my children every. single. day., fair is getting what you need, not making sure everything is equal.

In a government procurement process, in Alberta at least, the conversation goes something like this:

Bureaucrat: We really need to solve problem X

Procurement officer: Ok, please define the criteria of a solution that will solve that problem.

Bureaucrat: But how can I define discreet criteria that will evaluate if we've got the right solution when I don't know what the right solution is?

PO: That's not my problem, we have to be able to evaluate if the proposed solutions, in their entirety will solve your problem. It needs to be fair.

In theory, there is a good reason behind this process: to ensure that we only aware contracts based on the merits of the proposed solution, and not to our best friends. Appropriate use of government funds and all that. But what this does, in practice, is has the people who have no idea how to solve a problem determine criteria for determining of a solution is the right one. When a solution for a problem arrises that doesn't fit the mould of what the bureaucrat envisioned might be the problem, it cannot be considered because it doesn't meet the criteria.

To make the best use of governement money, we would want developers (or whomever) to be nimble, quick, risk-taking with relentless testing to find and mitigate errors quickly. I've always said that the list of requirements changes as soon as any part of the solution is put in front of the client. When you've RFP'd a huge solution, the winner is bound by contracts to deliver that solution even if you change your mind half way through. Sure there are change requests and so forth, but fundamentally the governement processes expect that nothing of the success criteria will change until you're done when in reality they change constantly and immediately.

I wish Mr.Park and all those working in the US administration the best of luck. I hope they are able to take the disaster of healthcare.gov and transform the mindset of governement bureaucracies. More importantly I hope that that trend moves its way up here. For me, since I tend to work in that environement all too often. And for the people of Alberta. The current system is wasting time and money and liberally burning people out. It is time for a change, and maybe this time that change will come from an unexpected source, the US federal governement.