Why isn’t all government software open source?
The federal government is the single largest purchaser of code in the world. So why is this code — taxpayer-funded and integral to the day-to-day working of our democracy — so often hidden from public view? There are two sides to answering that question: Why does the government so often build on closed platforms, and once built, why isn’t the code released to the public?
Using open source
It’s a lot easier to contribute to open source when you’re building on an open platform. While it’s possible to open source a VBA script, you’d likely have more momentum and receive a warmer reception from a platform with a more vibrant online community like Ruby or Python. Yet more often than not, the default in government is to look to “enterprise-grade”, proprietary platforms from the onset, which send the government on a closed-source trajectory.
The demand for “enterprise” solutions
There’s a good deal of FUD — fear, uncertainty, and doubt — in government CIO shops that open source is less secure, buggy, or more costly, and that you’ll be in for a lifetime of pain if you don’t invest in a real “enterprise solution”. For one, if an agency writes a check to a software vendor, they know what they’re getting. The contract spells out features, upgrade schedules, and allocates liability in the event that something goes wrong. More importantly, the vendor provides a phone number that the agency can call if they need help. “Post in the support forums and someone will reply” can be a scary proposition for a CIO.
There are fewer suits behind open source
Before that transaction even occurs, the closed-source platform likely has a flashy marketing page and a cadre of federal salespeople calling up the agency and tabling at conferences, things open source platforms traditionally don’t do, save Red Hat and a few others. And when the CIO’s office asks for “enterprise” features like audit trails or meeting certain compliance requirements, you can bet that the closed-source solution will make sure those features make it into the next release cycle.
Closed source contractors
Last, these closed source platforms are what government contractors know, because it’s what’s taught in computer science programs, and what they’ve always been asked to supply. If an experienced development firm has a bunch of ColdFusion developers, when they bid on a contract, they’re going to recommend that the thing be built-in ColdFusion. Not to mention, the government’s RFP may be scoped to favor the legacy technology they already know. All this means that before the first line of code is ever written, the odds are stacked against the project from ever seeing the outside of the agency’s firewall.
But even if the agency’s using a closed-source platform, there’s no reason their custom code can’t still be open sourced.
Contributing to open source
With the exception of 18F, CFPB, and a few others, government doesn’t actually write code. In fact, it rarely has the human know how to do so if it wanted. Instead, the agency traditionally plays the role of a non-technical program manager, providing specs for the functional requirements, and selecting a contractor to deliver the end functionality. The points of contact at the agency overseeing the contract are rarely engaged with the open source community, let alone passionate about open source. As a result, open source traditionally isn’t even part of the conversation. Why would it be?
Closed source workflows
In terms of the actual software development mechanics, the contract likely calls for a throw-it-over-the-fence workflow, where the agency doesn’t even see the code until it’s already in production, if at all — a long way from open source. Even if asked, the contractors may not have experience with more modern, open source workflows, or with participating in the open source community, creating a bad experience for all involved and further chilling open source efforts across government.
Reinventing the wheel as a business model
I also suspect that federal contractors have a disincentive to open source their work, considering that technical requirements likely don’t vary much from agency to agency. A FOIA request is a FOIA request and a press release is a press release, regardless of whether it’s on FAA or FDA letterhead. Open sourcing these solutions the first time around could potentially decrease the demand for that same code being written a second time at the taxpayer’s expense.
A culture of “no”
Once built, it takes humans to bring that code to the escape velocity necessary to overcome the agency’s guarded inertia. Security is going to likely say no. Legal is going to likely say no. You’ll have to get the code hosting platform approved. You’ll have to procure an ongoing maintenance contract to review pull requests. You’ll have to create a developer engagement policy for how you accept them. In a world of competing priorities, government employees will likely choose to move on to the next citizen-facing project, rather than spend potentially months combating the bureaucracy’s immune system to change.
Clash of cultures
Even if the agency manages to open source the code, the open source community follows a set of norms vastly different than the rigid hierarchy of government. Government agencies don’t always know how best to engage the open source community, or how to integrate an open-source workflow within their own command-and-control culture. Supporting an open source community takes time, something government employees are traditionally short on. And when the agency doesn’t follow the open source community’s unspoken norms, the naysayer’s worst fears become a self-fulfilling prophecy.
Transparency as a liability
Open sourcing code exposes the agency to the potential for criticism from millions of highly-technical, critical eyes, with little perceived upside from the agency’s perspective. The non-technical agency team may not have the ability to evaluate the craftsmanship of the code internally, and it’s often preferable to sweep things under the rug, rather than potentially air their dirty laundry to some of the internet’s most skilled trolls. Not to mention, the benefits of open source that advocates espouse are often not realized, and thus cannot serve as a counterbalance, if the code is so purpose built so as to render it unusable outside of government, thus attracting no outside contributors, or if the project is poorly managed, so as to scare those contributors away.
Meanwhile, unreleased source code poses absolutely zero liability in today’s political climate. Which one would you choose?
What needs to change
To flip the default, three things needs to change:
First, government employees need to be empowered with a better understanding of and appreciation for the virtues of open source. Those agencies who have open sourced code do so because of individual cheerleaders spearheading the charge. Successful projects are scoped from day one with the intention of being open source and serve to reshape market demand. Initiatives like 18F and the PIF program can seek to inspire and educate the next generation of open source advocates within government.
Second, even when the agency doesn’t explicitly call for it, as subject matter experts, government contractors have a duty to explore open source alternatives and to educate the market as to modern industry standard development practices. Any casual observer can see the direction the market’s heading, and the smart firms have an opportunity to get in front of it. Create internal competencies around today’s hottest technologies and grow government demand. Make it more practical for government to do the right thing, rather than the thing it’s always done.
Last, the open source community needs to step up its game, both in terms of what it offers — moving past the perception that open source is written by hobbyists — and the reception government code receives. On the supply side, there’s a tremendous business model in being the suits behind any one of the internet’s most popular open source projects — Automattic, GitHub, and Red Hat being a few examples — combating the FUD and providing “enterprise” support. On the demand side, the community needs to make it a liability not to release code (“what are you hiding?”), and make the agency’s return on investment clear, by breaking down the “us vs. them” mentality, and supporting, not criticizing, government efforts to learn open source.
Why isn’t all government software open source? Because you’ve got an entire value chain designed to produce closed source software, a system at equilibrium, with few incentives to rethink itself. Technology has made collaborating in the open easier in recent years, and as a result, the open source ecosystem has exploded. Yet, like all technology, government is still a few years behind mainstream adopters.
Hopefully, with your help, that can change.
If you enjoyed this post, you might also enjoy:
- Towards a More Agile Government
- Securing the Status Quo
- Why open source
- 19 reasons why technologists don't want to work at your government agency
- Five best practices in open source: external engagement
- Twelve tips for growing communities around your open source project
- The difference between 18F and USDS
- The three biggest challenges in government IT
- Eight reasons why government contractors should embrace open source software
- Ten ways to make a product great
- 15 rules for communicating at GitHub
Ben Balter is a Staff Technical Program Manager at GitHub, the world’s largest software development network. Previously, as the Senior Product Manager overseeing the platform’s Trust and Safety efforts, Ben shipped more than 500 features in support of community management, privacy, compliance, content moderation, product security, platform health, and open source workflows to ensure the GitHub community and platform remained safe, secure, and welcoming for all software developers. Before joining GitHub’s Product team, Ben served as GitHub’s Government Evangelist, leading the efforts to encourage more than 2,000 government organizations across 75 countries to adopt open source philosophies for code, data, and policy development. More about the author →