Tuesday, October 22, 2013

Software: Making It Work

First, there is a law of Nature that is immutable: Government never makes any software work the first time, or the second, or the third. The Corollary is that any Government software projects take five time longer and costs ten times as much as any commercial project.

Thus the obvious collapse of the ACA sites is hardly unexpected. Add to that a Canadian company with I gather a less than exemplary record and what would one expect. Now I started programming in 1962, a while ago, and worked on a few projects over the years. They all had their problems. But they all went through independent validation and verification, IVV. That way we knew if they dd what we expected them to before they were released. Also they were incremental.

Thus when I see people who apparently have had no experience opining on the topic I just am not surprised that they do so, after all they apparently have been doing it all their lives. Now an example is in the Times today:

What went wrong? My diagnosis is that there were three big mistakes.First, the ... administration acted too slowly. It waited too long to release specific regulations and guidance on how the exchange would work. It also waited too long to begin building the physical Web site. These delays were largely because the administration wanted to avoid election-year controversy. 

Yes, but.  Was there an architecture, architect. Did they get the best people or was it purely a political mess. I suspect the latter. So what have we changed to eliminate this factor? Nothing. Just a new political player.

Second, the ostensible quarterback of the federal health care exchanges, with responsibility for integrating all the various components, is the Centers for Medicare and Medicaid Services. While the agency has expertise in issuing reimbursement rules and overseeing large-scale claims-processing operations, it has little expertise in creating a complex e-commerce Web site. 

First this is not a Web Site. This is a transaction processing system interconnected to other existing systems with a front end web page which is tertiary at best. Also good web design means simple not one filled with useless pictures. Look at Google. Simple. Form follow function and having smiling people does not reflect function, especially if it does not work.

Finally, this was not the first health insurance exchange ever created. Massachusetts has had years of experience with its exchange, and there are private exchanges, like eHealth, where individuals can shop for insurance. In addition, many states, like California, Connecticut and Kentucky, had already spent around two years building their exchanges, gaining experience and proving it was possible to create a good customer shopping experience. It does not appear that the Centers for Medicare and Medicaid Services or its contractors spent much time reviewing these models and adopting best practices.  

And by the way CMS will be the same set of humanitarians controlling people's lives. Yes there are many functioning sites, even the Government has its own for employee registration every year. Why not just expand that? Because they wanted total control to reflect their own agenda.

So what would one do in the commercial world? Fire the management, get people who can do the job, and reward success and punish failure. And get the Government out of the way. But alas it will go in the other direction.

So despite what the author of the above suggests it will not happen.