It might actually be easier to just implement regular tectonics, reverse engineering them is only really interesting for the world editor
Looking at the blog Ninjabread linked the actual code itself is not super complicated, the problem is rather the data storage, which is a common problem for DF already. Like, DF has thinks like drainage, vegetation and elevation, now image there's several extra of those maps(blog mentions crust density/age/thickness, but I also imagine direction and speed are necessary). From the fact DF is already able to name islands and continents, I think there's already code that could potentially recognise and name plates. Because everything would be in a grid, it might not even be hard to reverse engineer fault lines, hell, if it's gonna work like the blog post, then tectonic plates could work in more mystical worlds too. (Imagine a world with a massive 'black hole' in the middle and the tectonic plates and ocean water just kinda drifting off into it, taking whole islands or even continents, and everyone dreading the day their home drifts off into it.)
However, the post does make clear that tectonics move in speeds of several million years. Thoth is describing a world with super fast tectonics, to make the latter possible you'd have to ensure it is still running post-worldgen.
I don't think it'll be super complex, but it might in fact be so uncomplex that it won't even show up during the first myth release, but instead in a dedicated 'earthquakes and volcanos bonanza' release. The most complex thing that is necessary is that the world itself can be modified post worldgen, which seems something that the myth map rewrite would take into account. (The actual hard map rewrite involves basically everything with coordinates, it is waaaay bigger than anything discussed so far.)
I'm sorry if this is rambly, but I am sorta thinking aloud