“In my view, shared by many blue-suiters, this marvelous airplane should still be operational but, alas, that was not to be. One of the most depressing moments in the history of the Skunk Works occurred on February 5, 1970, when we received a telegram from the Pentagon ordering us to destroy all the tooling for the Blackbird. All the molds, jigs, and forty thousand detail tools were cut up for scrap and sold off at seven cents a pound. Not only didn’t the government want to pay storage costs on the tooling, but it wanted to ensure that the Blackbird never would be built again. I thought at the time that this cost-cutting decision would be deeply regretted over the years by those responsible for the national security. That decision stopped production on the whole series of Mach 3 aircraft for the remainder of this century. It was just plain dumb.”

Ben R Rich and Leo Janos. “Skunk Works: A Personal Memoir of My Years at Lockheed”.

I felt something similar walking through Weird Stuff Warehouse and seeing old SGI workstations that I used to covet sitting on shelves for a pittance. Even if you bought them, you probably didn't have a 13w3 monitor or software to get them up. and now weird stuff is gone itself.

It's that "it's gone and can't come back" feeling.

When we look backwards in time, we see lots of tributaries that dried up for various reasons. That SGI workstation is still in your head, it is really up to us to copy and extend those concepts forwards in time.

That software is still sitting on a box of junk in someones basement somewhere, it just has to be found.

https://www.youtube.com/watch?v=NPvIs0I-gmg

First, the B-2 was the first large, complex product in history to be entirely developed digitally. Everything. was in CAD, and only the CAD geometry was authoritative. All drawings printed out were VERY clearly marked warning of this. IIRC, most of this was in Catia, so it should still be possible to read and use that data.

Second, if there's one thing Defense procurement Is good at, it's configuration management. This is doubly true of black programs like the B2, where not only do you want to make sure you never lose data or docs, but also that no bad guys ever have a way to get that data.

Either something went badly off the rails somewhere, or someone needs to rebuild something without the proper clearance. Otherwise, this makes no sense. BTW, the combo of the first complex CAD-only product multiplied by the hideous costs of any black program were largely responsible for the B-2's exorbitant costs.

Congress was stupid to cut B-2 production numbers, though, because the lion's share was fixed, sunk cost that remained the same regardless of whether one or a hundred airplanes were actually built.

(I think I've posted here before about how we spent many millions of dollars to develop a method to repair a major composite part in place. That part was pretty much what you started with when you put the airplane together, so even that was way cheaper than disassembling and reassembling the entire airplane!)

That was an innovative plane and the reality is that nobody is going to wake up and build it again once production stops. Once you break up a team, the team is gone — that’s why we always build submarines, even if unnecessary. You have to keep the people and plant working.

That tooling was going to be compromised and useful to the Soviets or some other adversary. There is a well established record of traitors selling plans or materials to the KGB/FSB/Chinese intelligence that have been caught. Also, from a PR perspective, a headline like “Air Force spends $50M/year to guard stealth bomber tooling untouched in 30 years” isn’t great either.

I mean if instead of discussing a topic freely on HN you were supposed to share your information using some I-1969 standard form with sections of one word answers, how many compelling and sensible posts would we usually see.

Many people are reasonable shrewd with their own money, and about stuff that directly impacts them. But people in charge of government organisations often have very different incentives and concerns.

(Other big organisations in the private sector can also go haywire. Especially when the threat from competition is low.)

Compare https://en.wikipedia.org/wiki/Public_choice

That being said, it destroyed a large part of the Canadian aeronautical industry since Avro ceased operations soon after and the well trained workforce was scattered to the States. There's a long history of Canadian governments doing shortsighted things like killing industries and then paying out of the nose to try to restart the industry every generation/ See shipbuilding in Canada and how Canada can't build ships for its own navy now, despite being one of the largest shipbuilders in the world in the early 20th century.

I agree though, the government has been short sighted and procurement of ships, planes has been a debacle for years. It seems that either way a lot of money is going to be burnt, either by propping up a non-competitive industry locally, or burnt buying something insanely expensive like F-35s from an ally.

I think this is why people like to refer to the Communinist Party of China (CPC [1]) as the CCP, because it evokes memories of the Soviet Union.

[1] cpc.people.com.cn

And let me guess, the 'pressure' comes from government decision makers that live in a perpetual revolving door with industry?

https://en.wikipedia.org/wiki/Revolving_door_(politics)

No, I think the pressure primarily comes from the DoD procurement process itself. First, you have to develop the requirements. The Air Force wants a plane, the Navy wants a plane, the Army wants a plane. If you say "yes" to one & "no" the others you get a political battle. A bright person says, "I know, we'll make a plane for everyone, and the volume purchase will make it cheaper!" "Great!" everyone says. They each bring in their laundry lists of what they want the plane to do... if you say "no", then they probably won't fund it, so you have this HUGE list of requirements. Now the plane (or whatever) has to do everything.

Then it goes out for bid. Only the lowest bid will win, so contractors will do their best to make it a low price for a massive number of requirements. They're under great pressure to underbid the real costs, and when you do new things it's usually more expensive than you imagine. Yet those costs will still look eye-wateringly big (because it's trying to do everything). Contractors will then say "don't require that 'bunch of useless technical data and those technical data rights... it'll be cheaper!". In addition, every time you learn something, you'll need to make a change, and that will result in a change request with a remarkably high price tag. That's partly because the contractors are motivated by profit, but to be fair, they were often pressured into unrealistically low starting costs by the process.

Many of the DoD's problems aren't from "evil government officials" or "evil contractors". Many really are trying to do good things. Sometimes the problem is that the DoD procurement process has some perverse incentives that sometimes results in very strange outcomes.

[0] https://en.wikipedia.org/wiki/Bradley_Fighting_Vehicle

[1] https://en.wikipedia.org/wiki/The_Pentagon_Wars

"The Best of the Journal of Irreproducible Results" (January 1, 1983) includes an article by William R. (Randy) Simpson about Murphy's Law, where he notes in a haha-only-serious way the many predictable problems that happen because of it. It's not online as far as I can tell, but the book can be purchased: https://www.amazon.com/Best-Journal-Irreproducible-Results/d...

Edit:

I'll answer my own rhetorical question: because budget forecast time periods are aligned with election cycles.

I don't think that policy (as written) would make sense for commercially-sold components (which were developed by the company without government funds and are being sold to the general public). Also, I think there are special cases where requiring the technical data isn't warranted. But I agree that policies could be greatly improved.

I think it would be appropriate to require technical data and technical data rights for all custom components by default (e.g., where the government funded part or a majority of its development). Waivers would be allowed, but only after being personally approved by a Secretary or Department CIO. That would provide some actual oversight; if program managers can approve making themselves look good, by lowering short-term costs (while increasing long-term costs after those PMs are gone), the rule will have no teeth. But government personnel want maximum flexibility, and while the current policies increase long-term costs, the policies also make it possible to get promoted when people are asked to do lots with relatively little money. So many government personnel are not very interested in that kind of policy.

A bigger problem is that contractors like it when their customer is stuck in an endless monopoly. Some contractors have been waging a long term effort to keep the gravy train going, and talk endlessly about the DoD "stealing our (contractor) data rights". The "stealing" they're talking about is allowing the DoD to receive the data that the DoD paid to have developed (!). Many contractors think they should have an exclusive right to the things they created using the government's money (and thus took little risk to create). For an example of this kind of absurd thinking, see: https://govcon.mofo.com/cybersecurity-and-data-privacy/break... Don't get me wrong, some contractors are actually quite reasonable, and the DoD cannot be successful without contractors. But contractors have a lot of incentive to gain exclusive rights over what they were paid to develop, and that's a huge incentives problem.

At least one past Secretary of Defense has tried to get more data rights through contracts, but it's very much an uphill battle.

There are some efforts to get more technical data rights in the future. I don't know what their status is, but my understanding is that they are minor nudges towards the government getting more rights to what the DoD pays for. These minor changes are considered by some contractors as the apocalypse, which should make it clear why big policy changes are hard. Here's some information on what was proposed, I don't know what has happened to it (I don't pay as much attention to this stuff as I used to): https://www.jdsupra.com/legalnews/dod-s-new-ip-rules-may-imp... I think these are at best small half-hearted measures to deal with a serious problem.

The whole point of hiring a contractor to sell a product is that the contractor is that it brings its own intellectual property to the table and “owns” the project.

Work performed as work for hire for the US government is public domain (subject to security restrictions). For some military products, the government “owns” the product. For example, artillery/naval gun tubes are manufactured by the Federally owned Watervliet Arsenal in New York, operated by a contractor.

The government owns other facilities relating to nuclear and chemical weapons that I’m aware of, and I’m not sure there is a material difference in competence related to the government “owning” all information.

If the data was classified (very likely) it costs a lot more to store it and audit it, and the pressure to "just burn it" is that much greater.

A few years ago I wanted to renovate an apartment. I dutifully applied for the plans from the local council. Only to be told they no longer existed.

I had this, "Well then, it was always like that" moment.

Didn't end up knocking out the wall, for other reasons, but if the plans no longer existed who was to argue that it wasn't always like that.

Nice tid bit for the B-52. These old birds, old in the sense of designed pre CAD, are quite hard to understand for modern manufacturing. Even manufacturers rely on retires sometimes to understand these old drawings for their own planes. Then again, the B-2 shouldn't fall in that category.

I suspect that people in the armed forces, at least some, understand maintenance but have no idea of manufacturing. A couple of years ago I had spontaneous "workshop" session where we discussed how we could leverage BoMs and data collected during production and assembly from the manufacturer for version and build tracking of military hardware. I used to work for that manufacturer in question, so I knew what data they collected during production. The only we would have had to figure out was which BoM to use. That was immediately shot down by middle management, because it would cost 10s of millions to get the rights for that data. Nevermind that it was there from day one, even in SAP so our client wouldn't even had issues with compatibility.

And none of the people I had these discussions with had any idea how BoMs and such are used. In these environments it is quite easy for crucial knowledge, especially of highly specialised parts and processes, to be lost.

As a CIM Manager for a major B-2 subcontractor, I can assure you they knew what they were doing. Yes, even 30-35 years ago. Yes, they knew as much or more then than we know today about BOMs, manufacturing, CAD, digital records and archiving, documentation and configuration management, and everything else. (BOMs weren't in SAP, but they were in mainframe DBs, ERP systems, and CAD systems, some of which are still sold and used today.) I have literally never worked on any other program in my entire career that was as well-run and orchestrated overall as the B-2 program. (We called it ADP-101 at the time, and for the first few years, we didn't even know what we were working on!)