I remember once reading some campaign reports from people going back and playing old-old-school 1st edition modules. Village of Hommlet, Against the Giants, that sort of thing.
One of the things the DM mentioned was that the more experienced players jumped right into using strategy and tactics to overcome stacked odds. I don't know for sure, but I'm willing to bet that there wasn't quite as much room for maneuvering around and optimizing characters, the further back into editions you go.
The idea of a Optimization/Strategy rift amuses me, if indeed one has steadily become more popular over time as the game has evolved.
More accurately, this is probably a false dichotomy--4e seems to stress inter-party cooperation very strongly and has a strong focus on the battlegrid and tactical movement. In that respect, it could be seen as a move back to DnD's wargaming roots. But I think it's undeniable that the companies that published DnD have noticed that, 'players like more options'--that is, more material to optimize. Since designing characters has become another part of the fun of the game for my people, books are published in keeping with that. Perhaps the ease with which one can design a Wizard, CoDzilla, or some other game-breaking build to be effective in any situation lends to less need for such cooperation? I don't know enough about the history of minmaxing in 2nd edition, and my knowledge of likewise 1st edition optimization is limited to a bizarre piece of fiction I read once.
DMs should probably tailor the levels of strategy and optimization used by opponents to those used by players. It need not be exact--enemies fighting smart, like the classic Viet Kongbolds example, might be a frustratingly interesting encounter for a bunch of cocksure minmaxers--make them use those smarts and powerful characters up against a challenge more complicated than numbers, eh?
On the other hand, it's somewhere between rude and sadistic to spring that sort of encounter against rawer players.