The way I see it, there are three types of paradigms in this topic.
Awaclus has a qualitative paradigm, where he says that any deck/strategy can be categorized in the basic types: BM, Engine, Rush, Slog and Combo/Other, and that there is a theoretical sixth deck type involving coin tokens that never works.
Chris and I have a more quantitative paradigm, where we suspect that there are underlying variables that distinguish these deck types from each other, opening the doors for hybrid decks to exist because they might fall somewhere on the borderline regarding one or more of these variables.
Finally, Traces Around has a nihilistic paradigm when it comes to deck types; they say that these graphs are useless in practice and we should only use terms like engine and BM to give strategy advise.
The nihilistic paradigm can be understood from a pragmatic point of view; if it doesn't actually help you win games, why bother discussing it? My answer to that would be: because I find it interesting. Another reason: by investigating this, we might get new insights into Dominion strategy that we otherwise wouldn't have gotten.
When choosing between a qualitative and a quantitative paradigm, I feel that the qualitative one has a big flaw in that it groups wildly different decks together under 'Combo'/'Other', as well as ignoring any decks that might fall on the borderline between two deck types. I'm not saying my personal quantitative idea is necessarily better: maybe the variables I distinguished don't say enough about the strategy of the decks themselves, which is why I encourage people to come up with alternatives. So far, only Chris has done that, which is nice, but I don't think we've hit the holy grail just yet.