There's a bit of subtext in the title that may be worth drawing attention to. (I dunno, maybe it's all obvious to everyone interested enough in this stuff to be

There's a bit of subtext in the title that may be worth drawing attention to. (I dunno, maybe it's all obvious to everyone interested enough in this stuff to be reading the dissertation. But perhaps not, so here goes.) In the 1970s, after their paper introducing Scheme, Sussman and Steele published a series of papers showing how to implement a bunch of things "on top of" a functional-programming substrate, many of them having titles or subtitles of the form "Lambda: the ultimate X". (The first one was "Lambda: the ultimate imperative", the second "Lambda: the ultimate declarative", others "Lambda: the ultimate GOTO" and "Lambda: the ultimate opcode".) The idea of these papers wasn't merely "look, you can implement X in Lisp", but something more like "You might be inclined to think of X as simple and primitive and cheap, and Lisp as sophisticated and complicated and expensive -- but that's mostly prejudice, an accident of the particular hardware we happen to have built first, and using something like Lisp as your lower level and implementing these things on top is better than doing it the other way around". So writing something -- especially something about a Lisp implementation -- with a subtitle of the form "X: the ultimate Y" where X is not lambda -- is pointedly one-upping that. Especially as the thing that $vau is allegedly providing the new best way to do is ... function abstraction, which is exactly what lambda is for.