26 January 2018

THE NEXT NEW MILITARY SPECIALTY SHOULD BE SOFTWARE DEVELOPERS

JIM PERKINS

The words “soldier” and “airman” do not immediately evoke the image of workers in grease-stained coveralls turning wrenches on tanks, personnel carriers, trucks, and aircraft. Most people predictably imagine an infantryman or pilot, even though the U.S. military’s “tooth-to-tail” ratio — the number of direct combat forces compared to support personnel — is around 1:5. Despite this common misconception, it’s easy to understand why vehicle and aircraft mechanics are critical members of the military team. These support roles and countless others enable our military to be constantly ready to “fight tonight.”

Militaries have always relied on technical experts to maintain their technological dominance. Blacksmiths supported sword-wielding warriors and industrial mechanics have been critical assets since self-propelled weapons and vehicles entered use in warfare. With each technological improvement, modern militaries have become more reliant on the support personnel who ensure that these advantages are ready for war.

For every innovation, militaries have required trained experts to maintain and repair it. The Navy runs its own nuclear power school to train sailors who will maintain the reactors on ships and submarines. All four services have cyber forces to protect (or attack) military networks and public infrastructure. Yet the military’s increasing reliance on digital technology over the last 40 years has not resulted in a corresponding growth in software experts.

To be clear, cyber warriors are not software developers. A common misperception is that cyber refers to anything involving computers. Cyber teams design weapons, not user interfaces and reports. They write code, but not in MySQL, Python, or Ruby. They are not creating new tools and systems to improve efficiency within their services. The services have created no occupational specialties for software coders or even skill identifiers for qualified personnel. To remain a dominant force in the Information Age, the U.S. military — all four services — must create a corps of software developers in uniform.

The Source of the Problem

The Defense Department’s failure to keep pace with the world around it is startling. Each military service is the size of a Fortune 100 company, but collectively they have zero software developers — an impossibility in the private sector. How could this happen? The problem stems from software procurement. Procurement uses a one-size-fits-all approach, even though hardware and software are fundamentally different. Unlike hardware, software’s intellectual property is all in the code. After an initial contract-based project was completed, software contractors may have recognized the long-term benefit of being retained to maintain and fix systems. However, the contracting officers may have viewed the work as finished and been less keen to keep the contractors on. Even more likely is that contracting officers may not have recognized the importance of the source code and may never have asked the contractors to hand it over – meaning access to the code ended after the project was completed.

As annual funding has ebbed and flowed, long-term maintenance contracts have been cancelled or left to expire. If the service wanted a new suite of features, a new contract may have been awarded to a lower bidder than the original developers. In this way, the services have both avoided training their own developers and also created countless layers of applications that do not talk to each other — patches on top of patches.

Instead of single sign-on portals – a hallmark of integrated systems – all of the services are well-known for having IT systems that require users to repeatedly log in as they move between them. In an attempt to remedy a particularly absurd instance of this problem, the Army is currently developing personnel management software known as the Integrated Pay and Personnel System — Army, which is promised to replace the 50 disparate personnel systems in use today.

Readiness Failure

With hurdles between the military and its code, the military’s ability to fix and maintain its own systems will be consistently inadequate. Just like a vehicle made from spare parts, a patchwork of IT systems will be fragile and require constant attention or risk being frequently broken. This a true readiness issue in the digital era.

Why is software readiness any different than equipment or personnel readiness? Tactical-level maintenance on vehicles and other equipment is a critical component of modern warfighting, and it is tracked at every level. Military mechanics are experts who often know the design and limitations of their systems better than the manufacturers and take unrivaled pride in keeping their equipment to the highest standard. Yet when software breaks, the services are unable to even diagnose, let alone fix it. System downtime is a bureaucratic nightmare, not something that can even be addressed at the user level.

The time for kicking the can down the road is over. Although the Defense Department has launched an initiative to promote open-source software, that is not enough. The costs of maintaining the status quo are higher than investing in the people who are the backbone of the military. The military depends on software just as much as it does on its equipment and the dearth of developers is a readiness failure.

No Tactical Innovations for Software

Rapid adaptation of hardware for new environments has yielded countless tactical innovations. In Iraq, tactical-level mechanics and welders enabled units to invent protective cupolas, turret armor, and even simple devices created to defeat IEDs using passive-infrared sensors. As a former combat engineer who cleared roadside bombs, I can personally attest to the value that our “grease monkeys” provided.

But when soldiers, sailors, airmen, and marines encounter an information problem, their only option is to try and solve it with repetitive, manual work. Even in the cases where the military does have access to the source code, there are no coders to create new reports or new tools, or to automate processes. As if to deliberately remove the option of purchasing other software, software procurement is so heavily regulated that it’s nearly impossible for tactical units to do, so they are often stuck with the basic Microsoft Office suite installed on their computer.

Eric Schmidt and other members of the Defense Innovation Board have repeatedly remarked about users (especially tactical-level leaders) spending tens of thousands of hours each year manually moving data from one system to another or creating duplicative tools — one to meet the Defense Department or service requirement and one for day-to-day use. Uniformed software developers will allow units to generate better, more accurate reports and allow leaders to focus energy and attention on the human issues instead of spending hours each week updating reports by hand.

The Force of the Future

In stark contrast to the military’s lack of digital experts is the large and professional corps of militarylinguists. Like they do for mechanics, the services recruit, assess, and train cultural and linguistic experts to be fluent in a myriad of languages including Chinese, Indo, Pashto, Russian, and Urdu. Many of these linguists spend years studying at the Defense Language Institute in Monterey, California. Additionally, any service member with a demonstrated talent for a language can take an exam and often get a monthly pay bonus.

Training coders could be less expensive than training linguists. Linguists attend immersive training for 18–24 months, whereas software developers could complete the equivalent of 3–6 month coding programs that are running in every major city in the country. Moreover, the value of gaining such skills could enable the military to offer equal or longer enlistment contracts and generate even more value for the force. One doesn’t need to look further than the Israeli Defense Forces for an example of how those units could function.

One year ago, the Defense Innovation Board recommended to then-Secretary Ash Carter that the department make computer science a core competency. How can the Defense Department be perceived as taking innovation seriously if this board of experts is not being supported on something so obvious? Since then, what progress has been made?

The Defense Digital Services and Army Digital Services, often cited as progress toward this goal, cannot be counted as wins. They both existed before the board’s recommendation and are both intended to circumvent the current system. Rather than build a core competency, DDS and ADS are described as SWAT teams or the Rebel Alliance with authority to by-pass regulations in order to fix major problems. For this same reason, more recent efforts towards open-source code are also insufficient. The services must invest and truly build this capability into the organization.

It’s 2018 and the services seem to care less about maintaining their digital systems than their vehicles. Perhaps this is because the military’s own lack of technical knowledge makes this problem too abstract for leaders to get passionate about. When the national security community thinks about computers today, the assumed context is cyber war, but the lack of nuanced language conveys a deep misunderstanding of the role of cyber, electronic warfare, and software.

New technology and weaponry is soon worthless without the requisite maintenance and repair. The services have been procuring software for tactical-level use for over 30 years without a single mechanic to look after it or to improve it for the new challenges. The services must immediately begin identifying, training, and employing software developers across the force. The U.S. military needs people who can fix its digital personnel carriers and turn wrenches on its information systems

No comments: