For over 30 years, I’ve been programming computers, all types of computers including no name personal computers, Apple II, Wintel PCs, midrange IBM, and UNIX systems. I’ve also done website work.
In my career, I’ve seen programs and systems fail. In fact, I was a developer on a project that, when put live, failed miserably. In a single year, the company lost millions of dollars.
I say this all to provide a foundation for the analysis below.
Anyone who says that will be completely fixed by November 30 is either outright lying or smoking crack!
The federal government spent 3+ years and $500 million dollars to produce the error-ridden result everyone saw on October 1. That wasn’t a couple of annoying bugs; that was a fundamental and wholesale failure of the code. Worse, it could indicate a design failure of the site, though without being part of the team, that’s impossible to know at this point.
Let’s assume that visible errors the entire country saw are corrected by November 30, that doesn’t mean the site will work. On the contrary, the odds are that a new wave of errors will be discovered in code that hasn’t had the opportunity to run because the prior errors prevented it. The sad truth is that the same people with the same coding philosophy created the entire code base. The code quality (or lack thereof) is pervasive. It’s conceivable that their error list will continue to grow as they fix more problems. It’s not uncommon to fix one problem and find two more.
Insurance companies said that the data they received from the site is being manually verified, oftentimes by calling the customer themselves to resolve data issues. With a small data flow, this is workable. Once more people can get through the site, the insurance companies will be overwhelmed, unless the data interface is corrected. This isn’t visible on screen but will prevent insurance from being bought.
So far, all the media attention is on the issue of code quality. Believe it or not, that problem is easier fix than if the problem is a design fault. In this case, they may have to trash much of the existing code and rewrite from line 1.
Either way, November 30, 2013 isn’t realistic. Maybe November 30, 2014.
Now, let’s look at this from a different perspective. They had over 3 years (1000+ days) to achieve disaster on the computer screen. Given this track record, is it likely (reasonable to expect) that 60 days is enough time for them to get it to work? I wouldn’t make that bet.
The government spend $500 million to get this result. That’s an incredible amount of money for what this site must do: have a shopping cart, create/maintain a user account, browse/search a variety of insurance plans based on criteria, and interface with various agencies and companies. This site isn’t breaking new technical ground, not by any means.
These things are quite doable and numerous private sector companies have done them and continue to do them very well, all for a lot less money. In fact, I’ve designed all of these functions and helped to implement them myself.
Plus, private sector companies do it securely, which seems to be a problem with the government site.
In the private sector, 3 years is enough. $50 million, not $500 million, would be more than enough. The private sector has every incentive to get a bang for every buck and to have a working site at the end. Failure is lost business, which is lost revenue. Lost revenue is lost market position, and in the extreme, bankruptcy.
For the government, none of the above concerns are in play. They can either tax more, borrow more, or simply print more money. No one’s job is on the line–they’re all protected by union rules.
I wouldn’t even venture a guess when the site will work from beginning to end. I highly doubt it’ll be November 30, 2013.
As an aside…
People shouldn’t wonder why I write the fiction that I write. Aside for national defense, I can’t think of anything the federal government does better than the states or the private sector. Every day the federal government intrudes more and more into ordinary citizens’ lives–it has involved itself in paying for healthcare now. (Face facts, payment dictates care–no payment equals no care. Low payments equals substandard care.) In the end, whether it’s 30 years or 300 years from now, there will be very little a person can do that the government doesn’t have a hand in. And if the government doesn’t do things well, how well can anything turn out?
This entry was posted in, software problems. Bookmark the permalink.

Leave a Reply