I think you're greatly underestimating how difficult it is to preserve this sort of knowledge. A particle accelerator is not a case where you're going to build

I think you're greatly underestimating how difficult it is to preserve this sort of knowledge. A particle accelerator is not a case where you're going to build exactly the same thing again, like you might do with an F-22. Any new accelerator is going to inevitably try to do something new. Therefore, the most important kind of knowledge is the knowledge of the planning and development process, and of all the things that went wrong (or could have gone wrong, but didn't), and how the design was changed to accommodate those. I think this sort of knowledge is incredibly difficult to capture, in large part because people usually don't think about it explicitly; it's almost a reflex as much as anything else. There was a paper a while back about Google's software engineering practices, and one of the interesting points that I didn't see discussed much was that "most software at Google gets rewritten every few years" [1]. The paper includes some justifications, but I think one of the biggest reasons is that it's very hard to institutionalize the sort of knowledge that you get by building something yourself. For the same reasons, I think Lockheed's documentation effort might be sufficient to build an F-22 identical to what they do now, but wouldn't be sufficient to design a new plane inspired by the F-22; and similarly knowing exactly how to build an LHC wouldn't be sufficient to build the next big accelerator. [1]: https://arxiv.org/ftp/arxiv/papers/1702/1702.01715.pdf https://arxiv.org/ftp/arxiv/papers/1702/1702.01715.pdf (see section 2.11)