A Navy drone prototype, UCAS.

FALLS CHURCH: “I’m going to frame this discussion around the ‘three nots,'” assistant secretary of Defense Katrina McFarland said this morning. “Technological superiority is not assured, R&D is not a variable cost, and time is not recoverable.”

“Sequestration for us is horrendous,” she told TechAmerica’s annual conference here. “Funding for the accounts that exercise our design engineers have declined nearly 50 percent decline in the last five years. That’s not trivial. That’s… the foundation of innovation.”

McFarland and her boss, weapons-buyer-in-chief Frank Kendall, have their hands full. They must meet urgent pressure to cut costs and still protect research and development for the long term: When she says it shouldn’t be “variable,” she means R&D funding should stay robust even as the overall defense budget declines.

Katrina McFarland, Assistant Secretary of Defense for Acquisition

But what are McFarland, Kendall & co. doing to protect innovation in a tightening budget? A lot, it turns out, from software acquisition to requirements generation. (Indeed, Kendall himself is so busy he couldn’t make his planned appearance this morning and sent McFarland in his stead).

Kendall spent the weekend working on a revision of the Defense Department’s infamous Instruction 5000.02. An interim document has been out for some time, but the final version is coming out “soon.” McFarland was sparing with specifics, but after her public remarks she did tell reporters one important change will be to how the acquisition system manages information technology.

Historically, the department has tended to treat IT as just a part of a weapons program or some other big acquisition with a correspondingly long timeline. That’s not nimble enough for a world dominated not only by Moore’s Law, which doubles the computing power of hardware about every 18 months, but by software updates that may need to be pushed out in weeks, days, or — in the case of a critical cybersecurity vulnerability — even hours.  At its heart, “IT is software and we need to manage software a specific way,” McFarland said. “You’ll see it in the revision to 5000.”

The Pentagon’s new approach also seeks to bring the private sector — where most innovation comes from nowadays — into the innermost arcana of the armed forces: requirements. Almost everything the military buys must be justified by and judged against a specific requirement written by military professionals. Historically, it’s hard enough just getting different subcultures within DoD to work together, such as technologists, tacticians, and acquisition officials. Now Kendall and McFarland also plan to seek industry feedback on “early-on conceptual requirements,” she said, even on the draft Initial Capabilities Documents (ICDs) still in the throes of being thrashed out between the services and the Joint Staff.

Frank Kendall, Under Secretary of Defense for Acquisition, Logistics, & Technology (ATL)

McFarland also implied Kendall is seeking earlier input and tighter control over what the services do. That’s a decades-old tug-of-war between the services and Office of the Secretary of Defense (OSD).

“Mr. Kendall reviews RFPs [Requests For Proposal] before they go final on the street to… ensure that the acquisition strategy [set by OSD] is complied with,” McFarland said. (Witness the ongoing battle over the requirements for the Navy’s UCLASS drone). “That’s not something that they’ve done in the past, [when the service] came in, told you what they did, and you lived with it.”

OSD oversight of the services can’t be just a check-the-box formality, McFarland said. “You complied with the process?” she said. So what? “That’s not the intent here, you’re supposed to comply with the content.”

Ultimately, all these changes are supposed not only to cut costs but to maximize investment in high-priority areas. Repeated themes include cybersecurity, electronic warfare, autonomy (i.e. robotics), long-range weapons (especially hypersonics), and the military use of space for navigation, communication, and intelligence — either protecting our space assets or, if necessary, figuring out how to live without them.

So how much money is actually going into these priorities? “I can’t at this time say how much because we’re in the final throes of our budget deliberations,” McFarland told reporters. “That’s where we’re at right now — determining how much of those areas are going to be implemented into this next budget.”

Of course, no one knows what the 2016 budget will actually be, because there’s a $35 billion difference between the president’s request and the level mandated by the Budget Control Act caps. So far both parties have said they want a deal to avert sequestration but have come nowhere near reaching one.

Next week’s elections are not likely to break the logjam, the Pentagon comptroller said.

Pentagon comptroller Mike McCord

“I don’t think many people at the senior level expect we’re going to have instant enlightenment no matter what happens next week,” Michael McCord told the TechAmerica conference this afternoon. “We don’t really expect that there’s going to be some kind of a budget deal.”

“It doesn’t seem that likely for the rest of this year, for the lame duck session,” he said. “We’re looking more ahead to the debt ceiling expiration next spring as a more likely — nothing is certain — but a more likely time that we’re going get some more answers.”

Even if that materializes, the agony doesn’t end in fiscal ’16, or even in the current five-year defense plan. “The next couple of years are going to be tight for us,” McCord said. “[But] I think actually at this point we’re probably a little more concerned about the five years after the next five years being really harder for us as we face some hard bills to modernize the nuclear triad.”

McCord is talking about the replacement for the Ohio-class nuclear missile submarine and the Long Range Strike Bomber, which — along with new missiles — will strain the ability of the Pentagon to meet what are expected to be huge bills for the two programs. The Navy has already admitted it can’t afford the Ohio Replacement Program, formerly known as SSBN(X), and still build a blue water fleet that meets its requirements.