Obamacare site: It never should have been this bad

Commentary:

(MoneyWatch) When President Barack Obama today acknowledged that the Affordable Care Act website "is not working the way it should for everybody -- there's no sugar-coating it," he actually was dipping the technical disaster in powdered sugar. Officials have said that they don't know how many people had enrolled using the system, even as some state-run exchanges have been able to provide performance numbers, according to ProPublica.

What caused the website problems? From the looks of it, long-standing architectural and design mistakes, a lack of an adequate testing, a wildly inadequate rollout strategy, political infighting, and the need to create the systems for 36 states that could not or would not build their own.

When the massive problems appeared as the website opened, the administration blamed the results on higher-than-anticipated traffic, spinning that as a sign that the law was badly needed.

The traffic excuse is not entirely unreasonable. Because so many states had opted out of building their own exchanges, the federal government had to pick up the slack. Few major transaction websites like an Amazon.com or Walmart.com go from nothing to massive popularity overnight.

But the federal government's IT experts are hardly new to running popular websites with heavy traffic. They should know that you can't suddenly flip the switch on something large and complex and then expect it to work correctly. These are the conditions that create not glitches, as the administration wants to call them, but massive weaknesses and failures.

The potential of major problems was obvious to insiders in March 2013, according to a New York Times report. Major deadlines were missed, money was tied up by congressional Republicans, officials were changing features even at the end of September, and there weren't enough federal employees with expertise in large engineering projects assigned to the effort.

All of that should have raised an ocean of red flags, but the administration plowed on, possibly because it did not want to admit to problems on such a politically touchy subject.

In the private sphere, a large software project such as this one that had to support many millions of users would be rolled out in stages, testing the waters, finding problems, fixing issues, and learning how to ready the system for a larger number of users. A phased rollout on a state-by-state basis would have made much more sense. Get the system working in one region, then one state, then a few, and eventually implement it for all. Use testing techniques and systems that help developers know if there is an impending problem before going live.

To attempt to launch for everyone at the same time was sheer folly and unbridled hubris. What this project did was to turn the country into Microsoft. Eventually the kinks will be worked out, but it could take months and all the savvy people know never to use the 1.0 version of a system.