The author is a CPython core developer so I expect he has some idea what he is talking about. Nothing guarantees that CPython will stick to semantic versioning. There is precedent for that in the way the Linux kernel is doing version numbers now.
elegant?? It's 2014 and Guido is talking about event loops.
beautiful??? Atrocious more like.
Python is terribly put together (I'd say designed but it's clear no such process has taken place). It's a language that,
like PHP and to some extent Perl before it, rewards idiotic behavior, makes it extremely easy to end up with horrible
unmaintainable code and requires strict discipline and a lot of experience in order for it to "work" in anything that's not exploratory/throw-away in nature.
Of course the (mostly?) clueless masses love it, because they feel empowered. But really ask yourself, what has Python brought to the table in terms of paradigm shifts? It's no Lisp, Haskell or Ocaml.
Hell, even Java looks to be better, especially for beginner/mediocre programmers.
Good decisions are really irrelevant -- being a core developer (and, more importantly, Director of the PSF) means that Coghlan has considerable insight onto what kind of changes are likely to to be accepted in Python versions with particular version numbers, independent of whether those decisions are good or not.
