Why the Healthcare.gov fiasco SHOULD teach us to Open Source Government Application Development.

We Spent How Much?!

According to The Daily Beast the United States Government has spent $118 million to build Healthcare.gov and another $56 million in fixing it…and based on the fact that the site isn’t expected to be fully patched for some time yet I wouldn’t be surprised if the total cost in “fixing” exceeds that of building the system in the first place.

Image courtesy of OpenClipart.org and Iwan Gabovitch
Image courtesy of OpenClipart.org and Iwan Gabovitch

I’m not going to take a position on the Affordable Care Act (ACA) – I try to avoid speaking publicly on controversial issues…but I would like to suggest a lesson we can learn from the ACA that I don’t think will be (very) controversial across party lines – that the Government should utilize open source in the development of applications as a standard rule.

Now, I’m not particularly interested in arguing that every government project should be open source – I’ll be happy if 95-99% of them are. I understand that some people rightly or wrongly believe that using open source in sensitive areas could cause security risks. I’ll let Kevin Clough and perhaps Richard Stallman1Though Stallman would argue for free software rather than open source, but I leave that semantics argument, however important it may be, aside for the time being to focus on an area in which a relatively minor change in procedure (moving to open source development) could make a significant change in cost and efficiency. argue that point.2There are some excellent arguments on how and why open source technology can be more secure than closed source technology. Specifically, the additional security in closed source systems usually isn’t b/c the systems are actually more secure but a function of “security by obscurity” – in other words, security holes exist, no one knows about them (including those who wrote the software). But I digress… But for the vast majority of projects (Healthcare.gov for example) I can see no reason why the development should not be open source and believe there would be significant advantages to such a course of action.

Lets take a look at the specific ways in which open source development could have reduced or eliminated the issues involved in the Healthcare.gov launch:

Transparency

The government (not just one department, but its entirety – e.g. the white house and congress) and the public could much more readily have seen that issues were arising, deadlines were slipping, etc.  and made necessary adjustments.

It is a constant problem within organizations that individuals at higher levels make decisions without the proper knowledge base upon which to make such decisions. This can result in unrealistic timelines and even if the timelines are realistic, if unexpected issues arise and there is slippage, there is a temptation to “gloss over” the setbacks and “hope” that the timeline can still be met.

This oftentimes results in extreme pressure on those actually working on the application as they are pressured to produce more, quicker – which, especially in the case of programming – is unwise. The more you pressure programmers the more likely they are to make mistakes, to take shortcuts and the more hours you demand of them the less productive they will become and, again, the number of bugs will grow exponentially.

Bug Fixes

Open Source software is oftentimes very stable and secure because of the number of eyes looking over the code. Further, individuals who are amateurs can make small contributions that allow the programmers to development on system architecture and bigger issues instead of stomping out bugs and making aesthetic improvements.

It would make sense for the Government to take a similar approach to Microsoft, Google, and Yahoo! on this front – each offers cash rewards for the discovery of issues. This is a relatively inexpensive way to get folks to pour in their energies – and individuals receive (for them) a significant compensation (hundreds to thousands of dollars – depending on the issue discovered).

Load Testing

The failure to properly load test the Healthcare.gov site is shocking. An open source project still needs robust methods of load testing performed by the core team – but it also benefits from other individuals and organizations implementing the application and discovering bottlenecks.

An open source, distributed team, also could have easily simulated the significant load that the site experienced upon launch – exposing the load issues early enough for remediation.

Code Reuse

When a project is open source the code can be reused by others for all sorts of purposes. The code to this project would certainly have applications in other government projects as well as the private sector. Reuse of code can significantly streamline development timeframes and even if someone in an entirely uses a portion of code for an entirely different project in a different industry – they will oftentimes contribute their version of the function (with enhancements/bug fixes) back to the original project (resulting in better, more flexible, secure, and robust code).

Cost

I really am just spitballing here – but I have a hard time believing that the development of an open source system to perform the Healthcare.gov functions would have cost anywhere near the costs expended thus far upon this closed source system. I’d guess that $10 million could have completed the project in a more robust and timely manner via open source.

Lesson Learned?

Please, let us take a lesson from this fiasco. We want more affordable healthcare – we can start by not wasting millions developing an application as a closed system which lacks robustness and stability.

I know some areas of the Government are already working with open source (and that is great) – but this needs to be a greater emphasis. Perhaps (I don’t know) there should even be some legislation that makes the (required) standard for new applications be open source and any applications which are desired as closed source systems should require review by a panel to determine if there is actual, substantial reasons for developing in a closed source system.

[Apparently I’m not the only one to think OSS could have made a huge difference. See this article by Christina Farr over at VentureBeat. Not directly related, but still interesting is Dylan Tweney’s article “Healthcare.gov costs show that feds have literally no idea how to build a big web site” also on VentureBeat. Another article comes from NBC News staff writer Gil Aegerter and can be found here.]

[11/4: Good article from Matt Asay entitled, “Sorry, Open Source Isn’t the Panacea for Healthcare.gov” on ReadWrite.]

How Not to Scare Pro-Gun Supporters (Quite as Much)

John Morse served as President of the Colorado Senate until a recent recall vote ousted him from the position. What had he done to invoke the wrath of the people? Pushed for gun control restrictions in Colorado in the aftermath of the Aurora (Colorado) mass shooting.

I’m not going to take a position on gun control, but I do want to point out a major faux pas1Merriam-Webster defines as “an embarrassing social mistake.” I’m not sure if I’m utilizing this word in its correct literal sense, but I suppose you’ll get the idea. committed by Morse in an article he wrote for The Daily Beast (Rampage. Regret. Repeat.).

Man Shooting Gun, thanks to milan6 over at sxc.hu for making the image available royalty free.
Man Shooting Gun, thanks to milan6 over at sxc.hu for making the image available royalty free.

Morse’s article outlines five focuses of his endeavors in the Colorado Senate including (1) limiting new magazines to be limited to 15 rounds, (2) everyone must pass a background check before purchasing a gun, (3) the gun buyer has to pay for the background check, (4) training for a conceal-weapons permit must occur at least partially in person, and (5) proactive removal of guns from the possession of domestic violence perpetrators.

Whatever you may think of this legislation, Morse’s primary mistake is earlier in the article. At the beginning of his article he notes how as a paramedic he was called to a home where a husband used a shotgun to shoot his wife in the head (resulting in her death) and then turned the gun on himself…All the while, an infant child is in the room watching.

Why is this problematic? Because it reinforces pro-gun supporters fears. Let me explain. The great fear of pro-gun advocates is that we are on a “slippery slope” towards firearms restrictions which will become ever more stringent. The legislation Morse passed would do nothing to stop this sort of violence – and pro-gun supporters know this – and thus know if that if these sort of cases are the impetus for gun control, then it is only a matter of time till gun control advocates see that previous legislation didn’t eliminate these cases and they propose yet stringent measures.

“But wouldn’t taking away this man’s gun have prevented this violence?” Unlikely. Gun restrictions may help reduce the body count in mass shootings but is unlikely to eliminate or reduce (imho) crimes such as the one Morse outlines above. If the offender in this case didn’t have a shotgun, he could easily have used a baseball bat or knife. The casualties would have been just as severe – and the trauma inflicted on the child would be equal or greater to that the child experienced seeing his mother and father shot (a gunshot blast is fairly instantaneous, bludgeoning or stabbing can be a complicated process).

Nor, from the information Morse provides, would this individual have been prevented from purchasing a gun by the legislation Morse passed.

  1. Shotguns are generally low capacity – sometimes one or two shots, sometimes five.
  2. It is possible this individual had crimes in his background – but Morse doesn’t mention any.
  3. Paying for the background check is unlikely to dissuade anyone – I assume this would cost around $10-$15 for a background check. At the most $25.
  4. This individual was using a shotgun, not a concealed weapon.
  5. Again, it is possible that this could have helped, if the individual had a criminal background – but Morse does not provide this info. – so I assume he did not. Even if he did, proactively taking someone’s guns, even if legislated by law is a litigious action, and I highly doubt that the police force is going to be eager to undertake such proactive procurement except in the most certain of cases.

So, what am I saying? That gun control advocates shoot themselves in the foot when they use crimes which arise emotions in us but that won’t be reduced by the legislation they are proposing.

Pro-gun advocates might be able to find more common ground if they didn’t see red warning signs that this was only the “beginning of the slippery slope.”

Dave Enjoys Gets An Upgrade.

I sometimes go AWOL from posting on Dave Enjoys, so I’m sure no one has been wondering why I haven’t posted in the last two days or so…but there is a real reason why – upgrades.

I got really tired of Bluehost having downtime, running slow, and so on. I hope they get their problems straightened out – but DigitalOcean is offering sweet deals on virtual private servers (VPS) at prices lower than Bluehost offers shared hosting. So, I’m in the slow process of moving sites over. I already moved FreeWargamer and have just now finished moving Dave Enjoys.

What does this mean for you? Well, nothing much on the front-end, but hopefully:

  • The site will load a lot faster, making browsing the site a more enjoyable experience.
  • The site will be more stable – rather than the frequent downtime I’ve been experiencing with Bluehost.

In addition, WordPress released the latest and greatest Jetpack plugin which means you can now used your WordPress.com authentication with Dave Enjoys. If that doesn’t make sense to you – that is okay – it won’t affect your utilization of the site.

Finally, this does mean that there will probably be some hiccups along the way…so if you see any technical glitches with this or any of my other sites please let me know!