thank you for collecting these sources together. It answers a question that has bothered me for 50 years. "Why did early americans use the Clovis spear head for 14,000 years? Why didn't they invent a better weapon?" Your materials provide the answer.

Why was the compound bow (the one with the pulleys) invented 200 years AFTER the last time a bow was used in warfare? Why didn’t someone invent it 20,000 years ago?

Perhaps I might suggest some further reading along these lines?-- starting with: "The TFX Affair" by Robert J. Art (the development of the F-111),"The C-5A Scandal" by Berkeley Rice, and "The Superwarriors" by James W. Canan (which describes the Pentagon mindset involved in much of this.) Read these three, and then we'll talk some more.....

Additionally, off the top of my head, two of the biggest reasons that are more or less background noise to both the doctrinal disputes among the service branches which affect this process and management problems per se are: (1) the tendency to "goldplate" each and every project to such an extent that the ideal becomes the enemy of the merely good
enough, and, (2) the fact that electronics advances far outpace design and production cycles such that expensive attempts to "shoehorn in" unplanned advances to the original design and/or production line are constantly being made--stretching out program length and raising costs.

The gold-plating is exacerbated as weapon development timeframes stretch out, and as weapon costs rise to the point that they must serve for longer because one can't afford faster recapitalization. Suddenly, you're trying to plan 50 years ahead because the F-35 will still be in service then. Moment of perspective here: 50 years takes you from the Wright Brothers to 5 years AFTER the first supersonic flight.

More gold plating equals more expense, and more risk of overruns because it's technically very hard to do right now. That lengthens the development cycle and blows budgets and costs, forcing program revisions down to fewer numbers bought. It also reduces availability rates because you're using more and more complex, cutting edge tech. Which makes the numbers crisis even worse. Smaller production runs raise the cost per item, AND lengthens the time older items must stay in service, to prevent sharp falls in available numbers. Those experiences lengthen time-in-service estimates for new items. Which brings us back to gold-plating.

Even as the rising costs and cutbacks in each new generation of whatever create shorter production runs, which makes it harder to keep design talent in key fields busy. But your industrial base will erode if they leave, and their accumulated lessons have to be expensively re-learned (or just go uncaught right into production).

And that's what we've had since the 1950s, in a tightening vicious circle.

So all these things do compound one another... but the Pentagon is indeed very American in its gadget orientation, and that does make a bad dynamic worse. Throw in a societal issue with any casualties, and that drives the need for performance again, and so up goes complexity, cost, et. al.

So, there's a lot going on here: a timing dynamic that creates problems (and worsens itself in a vicious cycle), societal pressures, a country's cultural mindset, certain fields where offense is improving rather faster than defense (big problem if you're the leader), the use of "political engineering" by politicians and industry to lock-in programs before costs are even understood, a political culture that seems immune to performance accountability (V-22 is the poster child - Pentagon tried to kill it several times for huge cost and poor performance, but Congress actively stopped them).

To which we add the same inherent problems with any public organization: ratchet effect, interest-group capture, public choice economics, and the rest.

The IEEE's prescription is a tall one, because it's unlikely that any President would make fixing the military procurement system their #1 or #2 Presidential priority... but that's pretty much what it would take to succeed, done by a popular President who remained reasonably popular.

With respect to item 2, think about whether you'd want to use a 10-year old computer for your work, and under what circumstances if ever. Now contemplate that most development cycles for weapons, which make heavy use of microchips, take longer than that.

No military procurement system I've seen has come up with a really good answer to this, and attempts to reconfigure the procurement process itself in order to address this have run into other serious problems (Army's Future Combat Systems, Coast Guard's Deepwater, British FRES armored vehicles come to mind).

At best, we have some promising examples - the Danish "Standard Flex Ship" and Absalon Class ships are both innovators here, without falling into the "frigate-sized coast guard gunboat" trap of the USA's Littoral Combat Ships... and "open architecture" electronics are becoming a military insistence these days. But no real solutions, yet.

My poppa worked for a minor defense contractor, a large part of his early engineering career spent on an Army project to come up with refinements to anti-tank shells.

His take on it was that there was an earlier age, where contract engineer (himself) and Army contract monitor had a professional, direct, and cordial working relationship of "let's solve these problems together", where they would discuss the technical issues during a shared rental car ride out to Picatinny, New Jersey to the Army base to run tests.

This has devolved into reams and reams of specifications and layers upon layers of oversight, where the contractors are spinning their wheels conforming to these requirements in place of doing much engineering.

And where has this contract conformance burden come from, but from the good intentions of the reformers, worried about contractors cheating the government. Each round of reform has added another layer of requirements and of oversight, and further bloated the cost.

This is like the financial crisis, where the current regime of oversight and regulation led to the crisis, so we are going to layer on yet more oversight and regulation, only this time we are going to get it right.

I don't know if there is any going back.

It's not just defense: the DoS is equally messed up, and I have good indications that Interior is even worse. One particularly fun one is satellite bandwidth used by contractors. They have to buy it largely on the spot market, which is the most expensive way to do so, because DoD and DoS refuse to commit to the multi-year buys, because their budgets can't anticipate that far out. So even when you know you are going to be in a place for three years, you can't buy a three-year contract up front at half the eventual total for buying a year at a time (or even less in many cases) three times.

You want to know what Eisenhower was actually warning about when he talked about the military-industrial complex? This is it. The Lefty fantasy that Eisenhower's warning was about more wars is amusing, but this is the real problem that Eisenhower foresaw and hoped to foreclose. Sadly, no one was listening.

One good thing that would have followed from McCain winning is that this kind of crap was a big issue for him, and knowing both the military and the Congress the way he does, he might have been able to make some improvements.

As someone who went from Silicon Valley to a major defense contractor, I can tell you that the problems in the defense industry are deep and systemic. While I see a lot of talent at the low end of the engineering spectrum, there is a nearly infinite reserve of talentless hacks in middle management who were promoted to positions for one of two reasons: 1) there was nowhere else to promote them based on their seniority 2) they were a diversity hire not qualified to hold the position in the first place. This dearth of good management leaves technical projects rudderless for large portions of their execution time and leads to substantial redesigns late in the game, when they are most expensive and difficult to execute. Furthermore, many of these hacks end up in positions of business development, where, with their complete lack of technical understanding, they promise their government customers features and performances that are not only unrealistic, but impossible.

Additionally, there is such an emphasis on conforming to regulations and following procedure that the room for innovation is extremely low. The “can-do” attitude of America has been replaced by union technicians who exit the building at 3:29 (who incidentally make more than many of the engineers) and woo be to the engineer who decided to finish soldering his test board together without proper union permission. Very early in my defense tenor, I distinctly remember staying late a few nights to complete a project that was holding-up another team of engineers. A college chided me with the words, “the company will never reward you for your hard work, so don’t”. Sad, but largely true.

As part of a research organization within the defense community, it becomes even more difficult. We literally have to have universities students run experiments for us because there are such elaborate safety procedures in place that it costs 100 of thousands of dollars to run basic experiments on rather safe materials. The safety protocol on dealing with a bottle of acetone is nearly adequate for dealing with plutonium, largely because the company is avoiding anything that could cost it a law-suit. Universities, along with China, don’t have to deal with this nonsense, which means that they can do research faster and better than we ever could.

At the end of the day, the defense contractors have optimized their system to avoid risk. Better to not innovate and not succeed than to hurt someone’s feelings (and get sued) or to allow any experimental risk that could also result in a lawsuit. In the defense contracting world, a company without the proper mix of ethnicities, genders, and disabled people is punished much harder than a company that fails to execute a multi-billion dollar contract. Is there really any doubt why the system no longer functions?

From Congressman who cut funding to "correct" programs one year and wonder why cost and schedule increase to government bureaucrats who demand every i is dotted with the correct ink and correct pressure to contractors who are optimistic because once they get the contract, they will more than likely get paid to make things work, the government (not just DoD) is screwed up. I can only come to one conclusion: The government acquisition system is designed to maximize inefficiency and cost.

I'm reminded of the way the F-86 Saber came into the AF inventory. As I remember, one day the President of North American showed up out of the blue at the office door of the Sec. of the Air Force (in the days when the Service secretaries were key players in the decision-making process) with a model of a plane he thought would be pretty neat for the Air Force to have. Solely based upon the reputation of N.A.'s success with the P-51,
the Sec. of the Air Force, after a brief conversation, decided to
authorize funding for test models--and the rest is history. If not for that we would not have had the F-86 in time for Korea and would have defense-less against the Mig-15.

Anybody want to bet what the odds of a similar thing happening today are?

I work for Lock-Mart. I'm sure we have our share of the talentless hack middle-management class, but it seems to me as if that's much less true now, in my division, than it was twenty years ago. Of course, that might be because I've sashayed into talentlesshackmiddlemanagement, myself.

Some good point, above. It's getting to be practically impossible to design for continuous upgradability, which is a necessity if you want to maximize product life. Say I have a targeting system, and that targeting system is evolving continuously to add more functions and capabilities. At various points in the life of the product, it's a given that you'll HAVE to upgrade the processor boards so that the constant algorithm upgrades are supported by more processing power.

To hit the point about shipbuilding, lightly, aren't we still in production on Arleigh-Burke class ships? Seems to me that we have one or two in the pipeline. And of course there's LCS-1, just finished a few days ago. Don't these qualify as viable shipbuilding programs? Or does "surface combatants" mean something different than I'm reading into it?

As far as "gold-plating" goes, my memory goes back to when the gold-plating issue was actually about gold-plating. I mean: horrors, gold actually has some applications where it's not easily replaceable, so you blow a few dollars, at most, in plating. In this as with practically everything else nowadays, the problem is not, in my experience, hardly ever building things better than they have to be. I mean, your competition is price-competing with you; unless there's collusion, you're all driven hard to cutting costs as much as possible. And then cutting them some more.

Also, there's the requirements-creep problem, but hardly anyone wants to talk about that. With regard to FCS, it's not creep so much as random walk. As far as I've been able to see, I mean.

I'm a systems engineer at a small aerospace company.

I look at this process, and I can only shake my head in dismay.

Everyone involved in development should read and understand Brooks' "Mythical Man-Month".

Research is an expensive activity. Per Brooks, converting something to a System typically triples the cost, making it also an expensive activity.

Systems Engineering is a minefield. Uncertainty can be a killer. It's essential to get the biggest uncertainties brought forward as far as possible.

I advocate separating the Research and Development activities. Each individually is expensive enough. Together they are even more expensive.

I like the A-10 a lot. It started with a role (close air support) and a well-suited technology (the Vulcan cannon). Systems Engineering work was limited to integrating the technology into its role. That worked very well.

The LCS went way over budget and timelines and the DDX has been shelved as it would cost a fortune to build as many components haven't even been designed or even tested. The OICW has had many remarkable failures, including an M-2 replacement... the Ma Deuce with modern upgrades and optics is still depending on barrels and frames made decades ago. Somehow the troops love a high rate of fire machine gun and like a separate grenade launcher and don't want to be swapping equipment in the field to turn one into the other, probably due to soldier lifespan anticipation while doing so.

I was at the end of a system inside DoD that was thought about in the '70s, designed in the early '80s, slowly produced in the mid-'80s and then delivered (only in pieces) in the early 1990's. It had a fatal Y2K bug. The computers on the street for normal use were better than what was delivered. That system had a price tag (on the white side) in the billions and had its own computer tracking system for changes, modifications and errors in the paperwork and design. The black budget cost was not to be revealed... no one ever talks about the black budget waste that goes on in DoD (and the rest of the security community, either).

One of the few good systems is the Stryker, which is just modular enough to be upgradeable, and highly flexible as a platform for C4I on the move. A 'battlefield network' will not be top-down, but bottom-up for survival sake, Iraq has demonstrated that. Integrating across the entire battlesphere from individuals to ships and aircraft is a huge problem top-down, but becomes amenable to common systems interfaces from the bottom-up. Not easy,nor good, but easier than top-down.

Yet the top-down mentality is what needed shaking up, and when Rumsfeld started to get to the point of having to throw out multi-decade projects the inner ring started to fight back. Getting a FCS should be something that happens piecemeal and each part required to integrate with the rest: let the designers figure out how to do it and still deliver something good that can be easily sustained by the logistics pipeline.

That last has been one of the few brightspots in the military: logistics has improved hugely, to the point where supply problems are due to things not being made not lost in the supply chain. Of course we did have to end up purchasing ammo from our NATO Allies, but they weren't firing the stuff... batteries, dust filters, and a whole host of small items to keep all the equipment in the field working was and is another matter. All that fancy stuff needs a supply chain to work... and the troops, bless them, phoned home for the things that no one had supplied enough of due to Congressional mismanagement of the budgetary process during the 1990's. 'War reserve'? As if. That costs money to have, maintain and keep track of.

So now we have the strange specter of B-52's until 2050 while, at this exact same time, we are retiring a number of those aircraft. We just got rid of the F-14 and F-117, so I guess all our close allies already have better... on the flip side the M-1 has become something of a land-mobile starship, upgraded multiple times and still a battle-ready tank frame. The NLOS-C has taken all the bad parts of the Crusader and kicked them, while incorporating automation, automated defenses and a hybrid diesel system so that it can run on batteries at sub-40 mph and be very quiet, while being fuel efficient at higher speeds. Start putting the NLOS-C, upgraded M-1, Stryker and such together and we start to see how older concepts can be upgraded and integrated with new ones. Throw in the UAV and first gen UGV's and some more of that comes into focus. Then put in reliable systems like the B-52 and M-2, both having gone far beyond any expected lifecycle, and the objective becomes clearer: purpose built, but not integrated to one technology. The system must interoperate - leave that to the designers and stop trying to dictate design. Mandatory interfaces which can be upgraded are necessary, the actual, functional utility can 'float' outside of that. Doing that requires a change in outlook in procurement... one that has not been coming for decades and is still as far away as the early 1990's.

Low level, but you figure if they can't get that right.

I was a grunt in Carter/Reagan. The M-16, mostly fixed by then, although we would go off base to buy Teflon gun lube. We all wanted the Vietnam CAR. Almost twenty years later, the Army listened. We went and bought Aimpoint sights. Totally unauthorized. So we would blow away at night fire, shooting tracers( which would trace back to the shooter) over black iron sights at black silhouettes in blackness. Well, it was pretty. Another twenty years later the Army came through with the Aimpoint like sights. Ditto boots, breathable rain-gear, poly sweaters. Basically, the private sector was twenty years ahead of the killer elite.

This is great, Joe. Thanks very much for putting these links together.

I wanted to ask about something the first article touches on only lightly, and generally isn't touched on thereafter. How bad are our technical base problems? Is there a good public report or two on the demographics of the problem?

It's one thing to say that American manufacturing is in decline, that skilled technical workers in particular specialties haven't enough replacements in the pipeline (and are due to retire at record rates), that the number of engineers we train is flat (and increasingly sought after in non-defense fields), that the quality of our engineering curriculum is in decline (an anecdotal observation based on an examination of calculus exams past and present at a major American university), and that the number of American-educated PhD students in the hard sciences who can get a security clearance is headed for the toilet (I can recall several memorable graduate courses where I was the only US citizen present aside from the instructor). But hard numbers versus requirements and needs have been a little difficult to get ahold of. Has anybody done a comprehensive report on the scale of our problem?

I see three problems as well that are purely HR and would be easy to fix, if we had the backbone to get rid of them:

1. The defense bureaucracy generally: the amount of time my highly regarded software system engineer spends on BS is rediculous.

2. Incompetency in middle management - many of these guys are ex-military hires brought in to politic the customer.

3. "Diversity" - I have a friend who complained to me about two black women who were hired on his job, because the army contact (a black woman) was complaining about how it was all white me.

Building cutting edge systems is hard enough (hell most software projects are small versions of what the DoD is experiencing.) Beyond that getting some true market based competition is critical and won't happen under the next president. It's too bad, it would be a way to cut defense spending without endangering our troops.

Clarification: #3 -- the hires were also highly unqualified, which is what happens when being qualified for the position is not the number one priority.

The gold plating is not just in the final product. I worked on an aircraft subsystem for a large defense contractor. The program was already in trouble due to time and cost constraints. The usual test set cost $1 mn to make and took a year and a quarter to build, program, and test. We were behind the eightball because we had 4 months. No one had any idea. For one thing the internal hardware design dept was booked.

I said lets use COTS for the control electronics and some very simple custom hardware for the interfaces. I did the board designs at home. Got them made. And the techs at the contractor built them. We got the job done in 3 months with a two person team (hardware and software) at a cost for the first at $.5 mn and replicas at $100K.

One of the things that helped was that it was not a deliverable. No gvmnt paperwork.

And why couldn't the company figure it out (it wasn't hard - I was a contractor)?

Mind set.

I had done a lot of commercial work in my career. So I wasn't encumbered with the usual defense/aerospace mind set. And what is that mind set? Every detail needs to be tested. However, that is not what they needed. All they needed the test set for at that stage of production was minimal hardware verification and total software verification. They really didn't need to know the contact resistance of the relays and switches - just open or closed. They didn't need to verify signal voltages - jut there or not there.

And don't even get me started on software. "C" is a disaster because it is not good for quick designs. I had a 4 person team on a radio redesign at another contractor. We could turn out a complete redesign (the DoD guys were trying ideas) in a month. Hardware and software. I used FORTH. A contractor we were competing with was using "C" with a 30 man team. It took them 6 months to do a cycle. They hated us.

So is FORTH allowed in defense projects? Are you kidding me? Why? Not enough built in protections for the mediocre programmers and of course when you have 30 man teams you are going to have a lot of average programmers. So design teams are bloated and slow.

The answer? I haven't got a clue. The software mandates were above my pay grade.
As long as DoD mandates a belt and suspenders approach at every level stuff is going to cost a fortune.

Everyone involved in development should read and understand Brooks' "Mythical Man-Month".

Word.

Belt, suspenders, and tell me at every point of the belt/suspenders design and selection cycle exactly why you're doing what you do, and what value the belt and suspenders add, and what your basis was for those figures.

Oh, and get me a materials safety datasheet on that belt, ok? I've got a laser eye exam scheduled.

Every government contracting job I have worked on failed or succeeded on one thing: were there good, stable requirements at project launch?

If not, we had budget and schedule problems.

Some of the best projects the DoD ever ran had one page of requirements, like SR-71 and U-2. Disasters like the A-12 probably had a couple of hundred pages. That changed every month.

Requirements creep: Brilliant Pebbles. Comanche. A-12 as mentioned earlier. FCS. Probably there are dozens of others. I recall a program that was supposed to be a replacement for Tomahawk, but the requirements never gelled.

Which is not to say that a program can't evolve. The program I work currently has added a ton of additional capabilities beyond the initial requirements, but those were sold as p-cubed-i. We didn't have all the improvements preplanned, precisely, but we knew what direction we wanted things to migrate in, and there's still plenty of room for further improvement. Most of it is software. I realize this doesn't apply to all programs, but it applies to ours rather well.

I think we need to take into consideration that many of these weapons systems never got fielded for the primary use envisioned- defeating the Soviet Union in a conventional WW3.

What seems like gold plating now might look significantly different in a shooting war with an enemy that could field a legitimate AF and navy.

So how much is a silver bullet worth? It seems a total extravagence when it never leaves the chamber, but when you need it can you put a pricetag on it?

And specific threats may receed but others appear. If we go to war with China in the next decade or two, our view of the F-22 may change quite rapidly.

Slarti,

DDG-1000 is dead, for all intents and purposes. They're just arguing over the 3rd ship/corpse - and my bet is that it will become DDG-51s.

Which would restore viable surface combatant programs to 1, by ordering more. DDG-51 is currently in close-out mode, and lead items had to be ordered to keep the option of more open. An active array SPY radar would be a big step forward if they reopen production, because the current ships can offer targeting level tracking of just 3 targets at any given instant - and fast switching is problematic against the new breed of maneuvering, supersonic anti-ship missiles. Bad news: it puts more weight higher in the ship, creating balance issues. Hope they figure it out for DDG-51 Flight III.

LCS? It's almost a great ship, and Lockheed's LCS-I for Israel is far more viable than the US version. A ship that is supposed to be the US Navy's low-end, but can't do anti-ship combat or fleet air defense, is a failure. A ship that can't be upgraded to do those things (for instance, via 16 Mk41 vertical launch cells), is simply insane. It's a program that was supposed to field 9 prototypes, and will field 2. The Navy just declined to use their FY 2008 funds to buy another, as planned. Its acquisition program is in shambles, and doesn't meet my criteria for "viable." My prediction is that it will end up canceled within the first half of the incoming administration's first term, or sent completely back to the drawing board.

So, let's see. At the high end,
CG (X), which was supposed to be part of the DDG-1000 family and intended to replace the CG-47 Ticonderoga Class built under Reagan, will not be underway in any seriousness until 2015.

DDG-51 in close-out mode. DDG-1000, which was the tonnage of the Graf Spee but still called a stealth destroyer, canceled (and at $3-5 billion per ship, not affordable anyway). No replacement for the current fleet of AEGIS destroyers, which are the backbone of the US Navy.

At the low end, LCS procurement program has been totally redone once, and even that new plan from last year is coming unglued. It cannot perform basic fleet functions of a low end ship even if it survives, and there is no complementary ship planned to do anti-air or anti-ship combat at less than $1 billion per (DDG-51s). Other navies can field low-end ships with the ability to fill these roles while performing other tasks, for $400-600 million.

That wraps up surface combatants, and not a viable program in sight.

CVN-21/ Gerald R Ford Class sper-carrier and LHA-6 America Class escort carriers are naval aviation, not surface combatants per naval terminology... good news is, they seem more or less on track. Though CVN-21 is very dependent on a new electromagnetic aircraft launch technology (EMALS) that works fine on some roller coasters but has yet to be scaled up for F-35s. If EMALS works, big savings in mechanical complexity and bucks. If it doesn't, serious trouble because a refit for all that steam piping would sink the project.

I don't have a quarrel with most of your comment, Joe, although it's very, very hard to say where DDG-51 is from online sources. fas.org, which once lay fallow for a couple of years, got updated once but the DDG-51 section is...well, politely, behind reality. They've only got names up to DDG-93. Globalsecurity.org says there are orders for three more beyond the Dunham, which just finished, and there are orders for nearly a dozen more planned over the next decade. You might know more than they do, though.

Oddly, I got hired back into LockMart to work DDX, which became DD-1000. Guess we didn't win that one. The first time I hired in here, we lost a missile defense program I was hired to work the day I reported. Guess I shouldn't hire in here anymore.

No replacement for the current fleet of AEGIS destroyers, which are the backbone of the US Navy

Earliest decommission date is for Arleigh Burke in 2026. I think we have time.

I was present for the commissioning of the Mason, by the way. I have a bud who was the CSO; he's since moved on to bigger things. Impressive pieces of equipment.

[Xpresssions, Ms Know, etc.: You're banned for nick changes (sockpuppetry), cant, and lack of substance. Please have a nice day, somewhere else. --NM]