TThe Post Office Horizon scandal has long been a frustrating one to follow as a technology reporter. Because even though it stems from the failure to deploy a large-scale government IT project, it’s not about technology at all.

In such stories there is a desire to uncover the specific fault lines that caused the disaster to occur. Taking Grenfell Tower as an example, the entire system was flawed and the investigation into the fire revealed gory details, but it is also clear that the fatal error was in covering the building with combustible panels. Identifying that fulcrum leads both ways to further questions (how were the panels deemed safe, and was the building able to be safely evacuated despite their flaws?), but the catastrophic It is clear where it is.

I feel like there should be comparable focus points in the Horizon system. “What happened at Horizon that led to so many false accounts?” is a question I’ve asked many times over the decade since I first learned of the scandal. Thanks to Computer Weekly for the coverage. I searched for systems in the hopes of finding some important crux, a terrible decision around which all subsequent problems swirl, that could be sensibly explained to provide a technical foundation for a very human story of malice and greed. I’ve been looking into architecture.

Still, the conclusion I’m forced to draw is that Horizon was really, really broken. From toe to toe, the system was terrible. Each postmaster had fundamentally different flaws, so a plethora of technical errors, worst practice decisions, and lazy cutbacks were probably part of the reason the Postal Service continued to fight for so long. Masu.

One system continued to accept input even when the screen froze, writing transactions to the database invisibly, while other systems simply had edge-case bugs in the underlying system that caused transactions to change. It just couldn’t lock when it shouldn’t have. There was also a problem with the network with the central database, causing transactions to be dropped without warning whenever there was a problem with the data connection.

Still, if you want to trace the point in time when bad IT became a crisis, you need to look completely into the technology past. The Post Office declared Horizon to be functional as legal tender. Everything that happened after that was a logical conclusion. If Horizon works, the cause of the error should be in the subpostmaster operation. If they say they haven’t made a mistake, they must have committed fraud. If they committed fraud, a conviction is morally right.

But Horizon didn’t work.

Today’s big technology companies aren’t so cocky as to claim that their software is perfect. In fact, the opposite is accepted as reality. The phrase “all software has bugs” is repeated too often and casually, implying that users are demanding too much of the technology they rely and work reliably on.

But they often still act as if they believe the opposite. My inbox is constantly filled with unmanageable people who have been falsely flagged as spammers, scammers, or robots by Facebook, Google, Amazon, and Apple’s automated systems. These people have lost years of shopping, lost access to friends and family, and lost the pages and profiles on which they built their careers. I can’t help them all and still do my day job, but strangely enough, the cases I decide I can contact a large company for are almost always easily resolved. It turns out.

No one would argue that even the worst software Google has put out is as broken as Horizon. (The Post Office says the current version of the software, created in 2017, has been found to be “robust compared to comparable systems.”) But the real culprit is broken software with flaws. If you’re acting like something isn’t supposed to be there, that’s serious. The tech industry may have more lessons to learn from this scandal than it’s willing to admit.