Should Obama Open Source To Fix The Fiasco?

Would open sourcing the site’s code help fix the mess? This man’s petition says “yes.”

Should Obama Open Source To Fix The Fiasco?
[Image: Flickr user Mararie]

Could the open-source approach fix Since its October 1st launch, the online health insurance marketplace has been plagued with technical glitches and error messages. A frustrated President Obama promises to have it up and running by the end of November, but at least one person thinks it could be fixed sooner––by opening its code up to the public and letting developers everywhere take a crack at it.


Matthew McCall is a health technologist who’s created a White House Petition to get the government to do just that. McCall works for the nonprofit OSEHRA, where he’s helping to build and grow a community around VistA, an open-source, government-created Electronic Health Record currently run by the Department of Veterans Affairs. He’s also a member of the Presidential Innovation Fellows, a program that brings private sector innovators into government to affect change. Here’s where he thinks went wrong and why open source can fix it.

Glitches aside, what do you think of the website?

Insofar as I’ve been able to get into the site, the design seems good. It’s pretty clean and simple. However, I’ve also only been able to get a couple of pages deep until I get an error message. I’m interested to see how well they have abstracted away the complexity once I can set up an account and apply, which I haven’t been able to do to date.


In your petition you say the problems with are “likely due to poor coding practices in components that are unavailable to the world’s development community to evaluate.” What does this mean?

I’ve read content from a few people who have done some speculative technical analysis, and while the system is clearly a complex set of interconnected solutions, it can be loosely divided into a front end and a back end. From what I can tell, the front end is fine; the back end is where the problems are. Obviously, something went terribly wrong and their solution isn’t scaling. I personally doubt it’s hardware issues, and assume it has coding issues, however all I can do is just speculate and assume the worst since I can’t see the code.

To me that is the real problem: The entire community is left to guess the issues and their resolution, so we wind up with Congressmen grilling executives over code, neither of which has probably ever written a line of it. By giving technologists a look, we can make informed opinions and help fix the real problems.

Were these problems avoidable? And if so, where did the administration go wrong?


There are always bugs on a product launch, so it would be naïve to say all of the issues could be avoided. Having said that, I think the “big bang” rollout was a bad approach, and a phased release could have helped find and fix issues earlier. It’s the same reason restaurants have soft openings.

The real problem, though, is a systemic one. Ninety-four percent of large federal IT projects have failed in the last decade, so I think all we could really do from the start was–channeling Obama–hope. This failure rate results in billions of lost dollars to the American people, and it’s in my opinion a direct result of a flawed procurement process. It’s very difficult for agile software companies to compete for contracts, let alone win them, so the usual suspects keep getting paid for not delivering.

I think if you asked any non-federal technology firm if they could build for a fraction of the hundreds of millions they spent on it, they wouldn’t just say “Hell yes,” they would deliver. If we can find ways to make it easier for them to get engaged, we can bring more innovation to government. This issue predates the current administration, but I think this failure is bringing the problem to light and they have an opportunity to act on it.

But the site is built, flaws and all. So how would open sourcing the code help resolve the problems?


To quote Linus Torvalds, the creator of Linux, “with enough eyes, all bugs are shallow.” Shared source code means anyone can pitch in, so you often wind up with higher quality software. Usually, it is done when people are solving a common need; say for a web server or database. I can’t imagine too many scenarios where private citizens would need to set up a federal healthcare exchange, but I know there are a lot of programmers out there who love our country and care about Obamacare, and might be willing to lend a hand with the issues. It’s an interesting opportunity right now for citizens to demonstrate patriotism through programming; I’d love to see what happens.

Why do you believe in open source so much? How is it more beneficial to its users than “private” code?

In the tech world, we are seeing an interesting shift toward embracing open source. Innovative companies are building open-source products, and offering a number of business models around them. It is paying off big for companies like Red Hat, Joyent, and MongoDB. They have essentially given their core software to the world because their business is not based on selling a product, but rather giving the product away and providing extensions and services with it.


The government doesn’t sell code, so what advantage could there possibly be to keeping it private? I actually believe it is a missed opportunity for the government, because they don’t really know what common need the private community could collaborate on with them for any number of projects.

It also has a massive benefit to contracting. There is a common situation in Washington called vendor lock. It often occurs when the government buys proprietary software and installs it, and from then on only one vendor is capable of providing support for it. They are then effectively able to set their own price point, regardless of the quality of service they provide. Software companies that work with the government covet it, since it is a source of stable income with large profit margins. Open-source code means that the software isn’t tied to one organization, so any number of companies can compete for its support, and competition is a good thing since it can bring down costs and improves service quality.

In your petition you say, “Code funded by taxpaying citizens should be made available to the general public as government funded development is generally public domain software.” I like the general idea of this statement, but isn’t there a cutoff point where this becomes unreasonable? I mean, our tax dollars have to help fund code for programs used at our intelligence agencies. Surely no one would expect the government to release that code, right?

You are right; there is a certain area of software in the government that shouldn’t be shared. Anything that has national security implications shouldn’t be shared for the safety of the people. Security keys and cryptographic algorithms shouldn’t be shared for network security reasons. In health care, anything that could compromise patient safety shouldn’t be shared. I personally don’t think the insurance exchange’s code falls into any of those buckets.


There is also the concern that sharing source code could let a malicious individual find a security exploit in the system. It is possible, but if you compare open- and closed-source projects, I find that typically open-source ones become more secure and reliable over time compared to closed-source ones. Security through obscurity is no security at all, and concealing source code constitutes that approach.

It is very important to also differentiate between open source and open data. I’m not asking for access to Social Security numbers, just the software that is capable of storing them.

Are there other areas in the government where they have open-sourced code before? How did that work out?

Yes, I’m very involved with VistA, the Electronic Health Record run by the Department of Veterans Affairs. The VA is arguably the largest single health care provider in the world; it has around 300,000 employees and hundreds of facilities across the country. Over half of the doctors in the U.S. have at one point trained in a VA facility.


Electronic health records are a big deal in the U.S. right now. The government is incentivizing their adoption through a program called “Meaningful Use,” but they are very expensive. To date we have spent, as a country, billions on their implementation.

The one VA runs is called VistA, is well liked by doctors, and is open source. Several companies have taken the code and built competitive products and services on it. They see the benefit of shared maintenance with VA of the core code, and are collaborating with the VA to do just that. Hospitals are running the system outside of the VA now across the world, and reaping the benefits of being able to customize their system as they see fit rather than being locked to any specific vendor. It’s a great example of how open source can not only help government, but also stimulate the creation of companies and economic growth.

Since the problems arose, has the Obama administration been making any steps in the right direction to fix this?

I understand that the current class of Presidential Innovation Fellows has been brought in to help; they are a very strong team and I trust that they will make a big impact. The other moves, such as bringing in Verizon, are interesting, but since I don’t understand the nature of the issues well, I can’t really determine if they are good moves or not.


I obviously think open sourcing the code would be a step in the right direction, particularly as they committed to do so previously in an article in The Atlantic. The code for the front end was previously available online, but taken down for some reason when the problems arose. In my opinion, that was a step in the wrong direction.

What is your prognosis for the long-term outlook of

It will eventually work. These are technical issues that can be worked and resolved, and I believe it will function as intended one day. Exactly when that will be though, I can’t say. Also, I feel as though people are linking technical issues to flaws in policy, which I just don’t see. Just because the website has issues, there is no reason to call Obamacare broken.


If you believe that open sourcing the code would be beneficial for Americans, be sure to sign the petition.