Trial by Website

Here’s a shock to anyone who’s ever done a buggy technology rollout.  Waiting a couple weeks for things to get better only makes the problem worse.  Ed Morrissey at Hot Air has a post this morning noting that even the New York Times can’t cheerlead this thing to success.  I noted this paragraph in the Times:

In interviews, experts said the technological problems of the site went far beyond the roadblocks to creating accounts that continue to prevent legions of users from even registering. Indeed, several said, the login problems, though vexing to consumers, may be the easiest to solve. One specialist said that as many as five million lines of software code may need to be rewritten before the Web site runs properly.

In a well-written software deployment, you can generally ID the number of bugs in relation to the number of lines of code.  It’s different for any company based on policies, but let’s be clear: If you’re re-writing five million lines, you have much, much more than that.  And re-touching that many lines pretty much means you’re on your way to introducing new bugs.

I get that the code has to be this complex.  However, it complexity can be done in pieces.  A modern approach to secured web services enables you to build code in big chunks that can be tested against the interfaces of the other areas.  As long as your changes in one place keep the tests green elsewhere, then you’ve contained the issue to one area.  But again, that’s modern code, and it doesn’t appear to be architected that way.  Let’s not get started on the security checking that has or has not gone on in this deployment, we’ll just wait for that fun to come out in a month or so.

By the way, this whole hindsight bit about how the shutdown distracted us from the roll-out… I don’t agree.  In fact, the administration appeared to not do anything about the problem for the three weeks, which is causing even greater problems now.

Oh, and I did like the comments early on by the administration about how we should cut them slack like we do Apple.  Let’s look at what happens when there are significant problems in marketplace software.

Microsoft has pulled temporarily its recently released Windows RT 8.1 update from the Windows Store due to a vaguely explained “situation.”

Microsoft pulled the RT 8.1 update some time in the morning of October 19, as reported by WinBeta.org, two days after it first made it available for download by existing Windows RT users.

Microsoft also posted a recovery image to fix the damages that the first release broke.  I’m not suggesting that the government pull the entire deployment.  At this point, live testing might be the only way they find how the issues are manifesting.  Oh, right, we’re dealing with real people who are really at risk.  This might not be so bad if it weren’t for the fact that many of those people lost the insurance they already had, and might not have insurance in time to replace it.

Government healthcare is likely here to stay.  The last chance to significantly effect change was probably the 2012 elections, and that didn’t happen.  This has moved from a pure political problem to a lesson in technology that will be studied for years, because anyone else doing what happened here would be fired and likely never hired again.