How is knowing where the published standards diverge from de facto standards precious and important work? You say that's where the value is, but the subset of people and organizations who would pay for that (if it's valuable, people will pay, right?) has to be pretty small.

>(if it's valuable, people will pay, right?)

No? There are tons of valuable contributions, pure and applied, that "people" (markets) do not pay for at all, or pay pittance relative to their usefulness.

It makes it much easier to build new engines in the future. Even if only a few people are interested in this knowledge, they can make a big impact with the software they write.

I don't think important work necessarily means that people will pay for it. The team thinks it is important work therefor it is.