"They just can't wrap their head around the fact that someone might build something because they like building things."
Even if that weren't true, I find value in documenting the old file formats and having readable implementations of the games that used them. From the point of view of the developer, it can be incredibly satisfying to see the functionality come together, and to get a picture of how things were done 25 years ago.
I'm obviously biased in favor of efforts like this. Part of the reason is that I'm going through another DOS game with a disassembler and trying to understand, document, and reimplement the game engine (it's going glacially; I'm no reverse-engineering expert, and I don't have much experience in assembly). To me, it's a fun subject for a hobby. And I don't like that a game that would run on a 25MHz 386 natively won't run at full speed in Dosbox in a computer built 10 years after the game's release (or on my various ARM machines, a modded game console, etc). Something feels "wrong" about that. I have the same feeling about the game being unmodifiable. I've seen how much came out from the open-sourcing of various id engines, and it would be great to see modifications, enhancements, and remixes of the other classics, too.