It seems mainly to be the difference between people like me who want to learn "bottom-up" and have an understanding all the way down to the metal, vs folks who prefer to learn "top-down" and want to focus on OO theory first then work down to its applications in Perl (e.g. via Moose), without ever wanting to go much lower than that.
Of course, both world views are reasonable and useful and maybe they just reflect the differences between us old-timers with an engineering background vs those folks with a more mathematical approach.
I can see your point about Conway's book from the "theory" point of view. I just disagree with your assessment of the irrelevance of the "esoterica" in the last third of the book. But it's easy to forget that not everyone sees (or assesses) the world the same as I do and it's good to be reminded of that. So thanks.