Pentagon: Trillion-Dollar Jet on Brink of Budgetary Disaster | Hacker News Reader

Pentagon: Trillion-Dollar Jet on Brink of Budgetary Disaster

wired.com

48 pointsyahelc77 comments

We could have been landing on mars for what that cost to develop.

Meanwhile the enemy is low budget suicidal terrorists that this or the TSA can't stop.

I was watching Cosmos last night, and Sagan claimed that the (wealth-relative) costs of send a manned expedition to Mars is less than that of a Dutch expedition across the world in... the 1600s I think it was? And yet they accomplished it multiple times.

I need to research it better, but it's a compelling argument if it holds.

"At an estimated $1 trillion to develop, purchase and support through 2050"

People are going to spend more on video games:

http://venturebeat.com/2011/09/07/with-online-sales-growing-...

Look at the luxury goods market. It's huge:

http://en.wikipedia.org/wiki/Luxury_good

If you want to go to Mars, there's plenty of money in the US economy where GDP is about 15 trillion a year.

Crowdsource an X-Prize...

It's important to note that headline-grabbing trillion dollar figure is for a timespan that reaches 40 years from now and covers at least 2500 fighters. It's also worth noting that the JSF is funded by a consortium of nations, including United States, United Kingdom, Italy, The Netherlands, Canada, Turkey, Australia, Norway and Denmark .

it's also worth noting that 35 years of the Federal Aid Highway Act that build the US interstate system costed less that half a trillion in modern terms, created jobs, industries, and changed the landscape of the country.

it's also worth noting that 10 or so years of the Apollo program costed < $150b in modern terms, and did a lot of good for science, created jobs, and industries.

A few planes don't do much good. It would be cheaper to take that money and carpetbomb your enemy. ww2 costed ~4 trillion to the US in modern terms. A trillion dollar pricetag on any plane that doesn't travel through time is pretty ridiculous.

The only engineering Lockheed is doing is financial engineering.

Actually, the way to fight terrorism is to destroy states that condone it and/or look the other way, as well as any major terrorist infrastructure therein (e.g. schools, training camps, stockpiles).

And the right way to do that, is by painting a bright "thou shalt not cross" line in the sand, and then enforcing it (if necessary) with overwhelming airborne superiority.

As opposed to a massive ground invasion and occupation, which has clearly been shown to be self-sacrificial and counterproductive.

-----

To respond in one place to the most common responses I'm getting:

It's fairly obvious in day-to-day life that and military history that (1) you don't appease or tolerate bullies; (2) you don't try to "reason" with the enemy. What I'm saying is a straightforward application of this, and what you're all saying is a straightforward denial of it.

What?

The way to fight terrorism is to provide an environment where people not want to be terrorists. Terrorism requires no state sponsored infrastructure. Bombing places to hell doesn't stop the spread of hopelessness, of desperation, of taking comfort in extremism.

Note: Using the "anti-establishment guerrilla fighters" definition of terrorism, not "causing terror". That's a whole other bag of bagels.

There's no single answer.

In Palestine, most of the terrorists (but maybe not the leaders) are alturists. Israel has captured quite a few, and profiled them - http://utexas.academia.edu/AmiPedahzur/Papers/209890/Altruis.... The same may have been true of the IRA, and it may be true of the Tamil Tigers. People only blow themselves up when they believe they are fighting against evil oppressors. In these cases, you are right - you need to address the root cause.

In the case of 911, the terrorists themselves may have been altruistic, but they were given support by Afghanistan (and maybe a few Saudis, but we need their oil). Knocking out states which support terrorism (which the US can do quite easily) may be an effective way to stop state-sponsored terrorism.

In any case, it's partly the result of peer pressure. If everyone in a community sees terrorism as a good thing, a few people in that community will become terrorists. Threatening states forces them to fight this attitude. In other cases, there is no state to threaten.

The way to fight terrorism is to provide an environment where people not want to be terrorists.

That's a strategy of altruism and failure.

The Arab Spring has probably done more to fight terrorism than any war.

Don't you think its a bit early for such a claim? Who knows where Libya and Egypt will be in 5-10 years from now?

To respond to your bullies bit:

Not "reasoning" with the enemy is absurd--that's how eventually you conclude a conflict. How many wars have ended with the complete annihilation of one side? How many of those victors are regarded with anything other than disgust or shame nowadays?

There is a very concrete point past which the cost of preventing an attack exceeds the damage inflicted by an attack. "Bullies" or "enemies" or whatever you want to call them can trivially exploit this for a civilization that cannot run numbers and face facts.

I believe that I showed in another thread that the TSA alone has wasted human-life-equivalent years equivalent to the 9/11 attacks.

On the schoolyard, in the pub, yes, you should stand up to bullies. But at a national level? As a civilization? You can't reason in the same way.

How many wars have ended with the complete annihilation of one side? How many of those victors are regarded with anything other than disgust or shame nowadays?

Nazi Germany? Japan? There was only one way to deal with them; we did it; the world greatly respected us for it; and it's a success story. (This is not to downplay the role other, allied countries also played in these conflicts.)

I believe that I showed in another thread that the TSA alone has wasted human-life-equivalent years equivalent to the 9/11 attacks.

Right - I completely agree - I'm advocating the alternative.

So you say -- "Peace for our time"?

(More modern: Please check the resource curse on wikipedia, or something. You can't reason with dictators which keep their population down so they can steal oil money.)

Edit: Hrm, Wikipedia says it is "for", not "in". Updated.

Wrong in two.

The way to fight terrorism is to act kindly and justly, and to promote the racial welfare to the extent that everyone is happy and well fed enough that they don't go starting trouble. Also, you get productive enough that the occasional bouts of crazies can be ignored in the long-run--a suicide bomber can only kill once.

Conversely, the way to fight a state such as you propose is to conduct the cheapest attacks possible while causing your enemy to expend lots of resources to fight an ill-defined threat, and to make public statements that are easy to flaunt and vilify.

EDIT: by racial welfare I mean human race, not Caucasian or whatever damn fool distinctions little minds want to draw.

> The way to fight terrorism is to act kindly and justly, and to promote the racial welfare to the extent that everyone is happy and well fed enough that they don't go starting trouble.

Do you have any evidence for this? The typical suicide bomber is middle class.

Well, we've hardly treated Latin America, most of the Middle East, or Iran fairly, now have we?

There will always be relatively isolated wackos who want to cause damage to you, for real or perceived injustice. As long as you are not provoking a movement against you, it would seem that those can be considered random events which don't need to shape policy.

No, the right way involves increasing economic co-dependence and providing better opportunities than drug smugglers and organized crime can provide.

And we undercut our ability to even attempt such things when we overthrow people's governments and occupy them for a decade, bring war-with-American-flags to their literal doorstep day in and day out.

How do you square that with OBL being heir to one of the richest families in Saudi Arabia?

You say that with the sort of glowing confidence that can only come from thinking in black and white.

That's because not tolerating and appeasing bullies, and not trying to "reason" with the enemy, are pretty obvious principles.

And when your principles, in practice, serve only to make more "bullies" and "enemies"?

If that happens, stand up to them too.

What else do you want to do? Lay down and die? Build more sewers and help them elect mullahs? Expand the TSA?

Anyway, you'd be surprised at the degree to which others will actually respect you for standing up to bullying, and winning; and the degree to which appeasement actually fuels more bullying (i.e., terrorism). In practice, what would happen is the opposite of what you think. (But if not, stand up to 'em.)

Your entire premise circles around this concept that some Evil Foreign Groups exist, always waiting for any hint of American weakness to pounce.

Have you ever considered that many of these groups develop as a response to our own neocolonial actions in their backyard?

I'm just giving the medicine to cure a particular disease. What caused the disease is irrelevant; all that is relevant is whether or not the medicine works.

That said, I very strongly oppose occupying foreign countries, the TSA, sacrificing our troops, providing foreign aid to nations that operate concentration camps (NoKo), etc. If I want to oppose these things effectively, I have to propose an alternative. I am.

By the way, I think the term "neocolonial" is stupid. I don't know precisely what you mean by it, but I'm assuming you mean roughly the preceeding paragraph, and I'm stating that my solution should not be considered a part of the imaginary "neocolonial" set of actions, but I don't know if you'll agree with that.

Many people in many countries view the United States as bullies. Should we never be appeased? Never be reasoned with? Never be tolerated?

That's the absolute opposite thing to do. We've been doing that for decades and all its done is bred hatred for us. Our interests are almost never pure.

The actual method is simple: You won't get stung if you don't poke the hornet nest.

The way they're beaten is through the people that harbor them, turn the citizens on the terrorists and they won't support them. Support of the population is one of the more important exploits that terrorists use.

The right way to fight terrorism is to educate the people in the countries they come from, by giving them schools and positive paths for their future.

How many poor teenagers do you think would join terrorists if they had better options and understood more about the world than the limited view they are purposely given?

Why do you think the first thing they do is destroy schools and ban women from attending them?

<sarcasm>Right on! Remember when we bombed Dublin and shut up those IRA dudes! Our unmanned drones totally delegitimized Sinn Fein, and the strategy of not reasoning with the Provisional IRA was such a great success! G.B. 4 Life!</sarcasm>

Terrific ideas. Can you cite any historical examples of when these strategies have succeeded? I'd hate to think you were just regurgitating neocon warmonger authoritarian dogma.

Hope it doesn't cut off the oxygen to the pilot when it detects a leak, or possibly at random times, like the F-22

http://abcnews.go.com/Blotter/22-raptors-suffer-apparent-oxy...

"Though they acknowledged the oxygen failure, Air Force investigators said the crash was Haney's fault for being too distracted by not being able to breathe and failing to either reduce altitude and take off his oxygen mask or to activate the emergency backup oxygen system."

Really? Who faults someone else for being distracted by an inability to breathe? I know that they're trained for high-stress situations but when it's the fault of the equipment in the first place, own up to it.

It's not a finding of legal fault. It's a statement about whether or not things need to change in the aircraft design. The oxygen system (presumably, obviously I didn't design it) isn't inherently redundant. If it fails (because Stuff Always Breaks) there are existing procedures the pilots are presumably trained on. The article lists two: an emergency backup (probably a second tank in the cockpit?) and the old standby of "dive to lower altitudes so you don't need oxygen, dummy". The finding was that these options would have been sufficient if they were used, so the design of the aircraft isn't deficient.

Obviously the software bug should still be fixed, and probably pilot training needs to be updated.

While the per-unit cost of an F-35 is quite high now (they've only built 60+) the comparative cost to the F-22 which was seen as expensive and too exotic will be lower in the long term.

While budget overruns are painful, the idea behind standardizing our fighter jets between the forces, thereby reducing costs is a wise move.

I always thought of the F-22 as a hand-built race car, while these F-35 are Porsche's that can be mass produced, are just as "fast" and will be more reliable due to the standardization.

Also deploying cheap drones are only possible when you dominate the skies with these expensive, complex, man-powered jets.

Also deploying cheap drones are only possible when you dominate the skies with these expensive, complex, man-powered jets.

Well, drones don't have to be cheap, and could hypothetically be just as capable as fighters in most respects (and much more capable in some respects).

The thing about drones is you can build them cheaply enough that it cost the enemy more to shoot them down than it costs to build more of them. Instead the military has decided to simply make fighter planes that don't need pilots but that's a separate issue.

Source? I'm not sure you could get a drone up and running for smaller than a multiple of the price of a surface-to-air missile.

Edit: I don't work with such things, but posting the first version of this felt unwise so I cut a few things.

Stinger's cost ~38k and are not guaranteed to work so multiply that by the probability of failure.

You can buy drone control system can cost below 1,500$ ex: http://www.micropilot.com/products-mp1028g.htm, granted there are good reasons to buy / build something far more expensive but we are looking at low cost baseline. After that it's a question of range and payload for 10,000$ you can get an RC jet that can do 200 mph and carry a camcorder for surveillance. Call it 15k for a fairly respectable surveillance aircraft. As to being able to fight back just add a spike.

PS: Now figure out what you can do with say 10 billion$ aka 600,000 of these things. Worse yet that's just an opening, they are cheap enough that you have plenty of budget for other things.

I'm really looking forward to the "Mythical Man-Month"-type book that comes out of this disaster. But since it's the government and likely "Top Secret", we'll learn absolutely nothing from this. $1T and no "lessons learned". Our (USA) tax dollars at work, folks.

"Top Secret" projects do share information with other "Top Secret" projects and over time many things do get declassified. IMO, the real problem is drones are simply more useful and a lot less expensive but chances are they can take a fair amount of R&D related to this project, and apply it to drones.

Also, total defense spending has crossed the 1 trillion / year number a while ago so cost wise this is not that unreasonable over the next 40 years. (Including things like nuclear weapons research at the DoE, GAO oversite, pensions, survivor benefits, and healthcare etc.)

The US Air Force design program is laughably bad - it would be more so if it wasn't our tax dollars being pissed away.

The same thing happened in the past to the B-1 and B-2 bomber projects, which cost $200m and $700m per unit respectively. Massive scope creep leads to huge unit cost inflation, followed by a reduction in number of units ordered to keep total cost below the Congressional mandate for the project. The low number of total units ordered leads to an increase in maintenance cost per unit, which eventually leads to shortening the plane's lifespan (too expensive to maintain, each airframe receives double the hours it was intended to endure).

Compare the F22 (at 150m) to the F16 (at 10m). Aside from the stealth capability, this jet is not 10x better than the F16.

While I agree with your overall point, the cost of a new F-16 these days is more like 40-60 million USD, not 10 million.

The cheapest 'good' 4th-generation fighter is the Saab Gripen, which was designed in a single country to intentionally-limited specs, but even those cost 30 million USD a pop for the standard C & D models and even more for the NG units under development.

To some extent modern fighter aircraft are inherently expensive as a result of all the technology that has to be crammed inside them. They're multi-role networked flying systems these days.

In essence that's part of what has derailed the F-35. Technological advances have allowed for the extensive use of software to create capabilities that just could not have existed before now, such as the ability to see 'through' the plane, an extremely high level of sensor fusion and integrated systems, etc. But massive software projects are inherently risky and can fail massively if not managed very carefully. They've been doing things with software in the F-35 that have never been done before in this context, which only increased the risk.

Also do not underestimate the costs of redesigns following the theft of F-35 design information by China from both Lockheed-Martin and Pentagon computers. I doubt that's the sole cause of the cost overruns, but it definitely hasn't helped.

Obviously lines of code is a poor metric, but wikipedia (http://en.wikipedia.org/wiki/Source_lines_of_code) would suggest it is not as complicated as:

Linux kernel 2.6.35 13.5 million lines of code

Windows Server 2003 50 MLOC

Mac OS X 10.4 86 MLOC

Debian 5.0 324 MLOC

<EDIT>

And alternatively:

"if you bought a premium-class automobile recently, it probably contains close to 100 million lines of software code":

http://spectrum.ieee.org/green-tech/advanced-cars/this-car-r...

Yes, but this code is low-level, life-threatening on crash, hard real-time, etc.

Not really. Those counts include totally unrelated components (like drivers) which you can optionally turn off, and just forget about.

Whereas the total count for the jet code is (I presume) of highly relevant code all around. It's not like it has 100 video drivers code included, and support for 2000 network cards, or several filesystems.

Could the F35 choice of C++ be the problem or heavy use of OOP?

It appears they have struggling with the complexity of their software implementation. It might just be their requirements but I wonder if it is similar to Ericsson's massive failure with their AXE-N project. No one knows exactly why that software project became too complex but it has been speculated that:

"One particular problem area identified is the system's software , especially for using extreme attention to object orientation (which at the time was the golden hammer ) in both computer programs as databases , [8] which can be difficult to combine with parallel programming, and that they chose to develop and make use of largely new, completely untested methods."

Ericsson attempted to heavily use OOP for the AXE-N, might that be the cause of the failure? The failure spawned the Erlang (functional language, with strict evaluation, single assignment) project which went on to be a massive success in their next product.

http://translate.google.com/translate?sl=sv&tl=en&js...

The reference to one trillion dollars in the headline is accurate, but lacks context that makes it a bit sensationalist. The article elaborates the cost as "an estimated $1 trillion to develop, purchase and support through 2050". Or about $25 billion a year. Whether that's pricey or a bargain depends on how many aircraft are purchased and supported; TFA states that the "Air Force, Navy and Marine Corps are counting on buying as many as 2,500 F-35s", which would be $10 million per craft per year for the next four decades.

This is not to say the thing isn't an overpriced, failing project. If the GAO's analysis and predictions cited in the article are correct, it sounds pretty bad. In particular, it's hard to imagine that four decades from now, the US armed forces would want to still be using early 21st-century aircraft.

You are omitting the time value of money which can make a huge difference in such a large time frame. http://en.wikipedia.org/wiki/Time_value_of_money

What realistic scanario has the following characteristic:

We are fighting against an enemy for which fourth generation fight planes are inadequate and said enemy does not have nuclear armed ICBMs?

Meanwhile the Air Force is bent on taking out of service the one type of aircraft that is most useful to actual, likely war-fighting - the CAS planes. The AC-130 was originally designed in the 60s and troops still cheer when it arrives over the battlefield.

The boys that run the fighter jet mafia consider such "bus driving" planes beneath them, so we instead of upgrading the A, C and perhaps B lines we have to spend hundreds of billions of dollars buying the kind of F line toys they prefer to play with.

I was following the (at the time) YF-22 and the proposed JSF when I was a kid. I'm now 28. These things take so long to build and there are so many hands in the pot that they can't help, but be ridiculously complex.

Talk about changing requirements during development. Imagine working on a software project that took 20 years to develop.

Add that to the fact that any large military project has to be spread out over as many congressional districts to keep as many congressman happy as possible.

what is the primary language for fighter jet software?

https://en.wikipedia.org/wiki/Lockheed_Martin_F-35_Lightning...

quote:

Unlike previous aircraft, such as the F-22, all software for the F-35 is written in C++ for faster code development. The Integrity DO-178B real-time operating system (RTOS) from Green Hills Software runs on COTS Freescale PowerPC processors.[211] The final Block 3 software for the F-35 is planned to have 8.6 million lines of software code.[212] The scale of the program has led to a software crisis as officials continue to discover that additional software needs to be written.[213] General Norton Schwartz has said that the software is the biggest factor that might delay the USAF's initial operational capability which is now scheduled for April 2016.[214] Michael Gilmore, Director of Operational Test & Evaluation, has written that, "the F-35 mission systems software development and test is tending towards familiar historical patterns of extended development, discovery in flight test, and deferrals to later increments."[215]

I think they are doing the JSF in C++. Here are the coding standards. http://www2.research.att.com/~bs/JSF-AV-rules.pdf

Those standards are a pretty good read for anyone wanting to write good code in that language.

Or indeed, anyone who wants to write good code any language. I don't work with C++ but it's a clear and well structured guide to OO development in general.

Used to be Ada, which for all its flaws was a great DSL for it (type-safe, bounds-checked in the type system, etc). Wasn't popular tho' so now people use C and formally verify it in OCaml.

Woah. Is there any available material on how they're doing that/how it's working out?

It seems to be Java: “Software providing essential JSF capability has grown in size and complexity, and is taking longer to complete than expected,”

Just kidding. Don't stone me.

It's all relative. As hackers, you can't even imagine what its like to code for the govt (I used to work for navair). Even the most corporate java programmer would find the govt program ugly and overly-complex.

Oh, I've worked for the military sector, with those huge NATO specification books. Also with JSF, the Java thing that happens to share TLA, but in another job. Honestly, I'm not sure which one was worse.

Thank you for that.

Ada is the standard language for fighter jets.

Old and Arcane it is, but its the most functional...

Usually ADA

Here is Canada our government is insisting they are going to buy some F-35 jets.

People have complained other than its complexity the single engine is bad if you are over the Arctic and the engine fails, most modern Canadian jets have dual engines.

People have complained other than its complexity the single engine is bad if you are over the Arctic and the engine fails, most modern Canadian jets have dual engines.

I wonder if the F-35 has a particularly reliable engine.

I'm not sure but if it's a vertical take-off and land (VTOL) it must be pretty durable.

Found this gem on Wikipedia under Lockheed Martin F-35 Lightning II

Canada had previous experience with a high-accident rate with the single-engine Lockheed CF-104 Starfighter with many accidents related to engine failures. Defence Minister Peter MacKay, when asked what would happen if the F-35’s single engine fails in the Far North, stated "It won’t"

Looks like the F-35 is the SAP of fighter jets

FYI: The eurofighter has about 1.2 million LoC (but they use other languages too) and is written in the same programming language (Ada). They have about 500 coders working on it. Impressive..

I hate when companies under bid projects and then the public wonders why its so much over budget.

Government contractors do this all the time. They underbid, and then 1 year later, ask the government for more cash because they are soo far in and know the government will do it.

Government contract work is sooo lucrative too. Most of that initial investment into Lockhead probably just went to pay salaries of the higher ups and did no actual work.

Sorry for Ranting.

I have never quite understood why so many powers (the United States, Canada, Japan, etc) have lined up to try and buy F-35s. Every plausible enemy our countries could square off against have extremely old technology (or no technology to speak of). In fact, I would be prepared to argue that slightly updated F-15s are superior to anything that any plausible enemy could attack us with.

Am I wrong? If so, help me understand why!!

I concur. IMO all the US needs to do is upgrade/update its F-15, F-16 and F/A-18 fleet to stay ahead of potential enemies for several years.

While certainly not cheap, the B2 was brought in under time and under budget. I also think we got our monies worth out of it in the several wars it has been used in. Source; the project manager. Disclaimer: he's my brother---we are both a little biased :)

Kill it. We don't have the money. All those alternatives that other countries are looking at are alternatives for us (US) as well.

some basic math...1 trillion over 40 years is about 25 billion a year....US budget in 2012 was 3.796 trillion FOR THE YEAR....so that is 0.6% of the budget. I think we can all agree that it is certainly militarily advantageous to have air superiority keeping in mind that it is mainly for the wars we dont fight and will never fight rather than our current conflicts.

Also these articles fail to mention the planes that will be retired over the next 40 years that the JSF will replace. Lots of costs savings that arent included.

Military superiority is supported by practical every political base including libertarians. Large, complex aircraft cost a ton to produce. Is there waste? Sure but some of the waste is due to our own democratic political system. Easy to cast all blame on greedy govt contractors.

We could replace those planes with cheaper alternatives.

The total program cost may be a small part of the budget, but our economy would do better if we spent that money on things we can use, not things that sit in a hangar and suck up money.

Brink? How is this not a budgetary disaster already?

What 'technology' costed a trillion dollars to develop? Certainly not going to the moon.

Reply on news.ycombinator.com