If you’re one for the mixture of politics, policy and information technology, you must read [Re]Coding <America> by Jennifer Pahlka.
For IT professionals, every chapter resonates and has just enough technical tidbits to confirm Pahlka’s technical chops. The book details incident after incident in which institutional values overrule what has been best practice in commercial IT for at least 20 years. It’s a shame because, as Pahlka points out, the government was IT starting with Hollerith.
But what had me slapping my head in amazement was the fidelity the book has to my own personal experience in 2018 and 2019 working for the Commonwealth of Massachusetts. The project was to migrate its Unemployment Compensation system from an on-prem environment to AWS. This was Massachusetts’ first attempt at modernization — and it fell victim to all the pitfalls Pahlka describes in her description of the COVID-era backlog at California’s Employment Development Departmentm(EDD).
I saw first hand the resistance, the inability to understand that less is more and, what sticks most in my mind, the way management was cowed by the workforce. Management permitted behaviors, including harassment and humiliation, that would never be tolerated in the enterprise space I mostly consult for. For example, one enfant terrible was indulged because — get this — the Commonwealth’s extensive IP network was 100% manually assigned RFC 1918 IPv4 address spaces — no DNS. Management was scared stiff of losing him, it was a union position and he knew he had the state’s entire network in his head and no one else did.
I had taken the gig — at a discount — because I had a sense of mission along with a burning belief that migrating to the cloud would benefit UI claimants. But I was wrong. At the end of the project, the state had more of a mess on its hands than it started with. Mass’s UIOnline (from the same vendor as EDD’s system) fell victim to the same backlog and fraud issues that EDD’s system did in 2020 that Pahlka relates in gory detail.
I’d always wondered if there was a systemic reason for this experience. If there was something — something I couldn’t quite describe, something in the DNA of the way governments deliver IT projects — that led to what I consider a debacle but which everyone from the governor to the department secretary declared a success.
Now, in Pahlka’s book, I’ve found the answers. This page-turner finally expresses the totality of what I saw. From endless procurement to a lack of product management sprinkled with a great deal of ass-covering in between, [Re]Coding <America> details exactly what is wrong with IT in the government.
Where Pahlka and I disagree is that it’s fixable. I think that the victories she details are small compared to the lethargy of government. It’s not a David vs. Goliath story. It’s a speck of sand trying to stop a wave.
Datapoint: Pahlka’s beloved US Digital Service has suffered a political calamity and is now the US DOGE Service. The empire has struck back. Another: the book details the judicial bent of agency rulemaking. That’s not going to change. Yet another: the primacy of policy over delivery. As long as policy is the road to power, building good systems will never be elevated inside government.
Still, Pahlka’s optimism is at the very least a roadmap to what must change if we want our government to be more effective using technology.
Leave a Reply