Why is it that everyone is so hard on RPGs for being similar, when there are a hundred identical FPSs out there... lets face it, FPSs haven't changed in any major way since before Doom
Of course it's an oversimplification -- there's always exceptions. But there's some truth in what he said.
For your first clause, because shooting things with a gun is inherently fun, and picking "attack" froma menu really isn't.
For your second clause.
Doom used what is considered a 2.5D engine, in the sense that maps were 2D in nature and the 3D aspect of them came from basically extruding them up. The shift to actual 3D was pretty big. So were additions like ducking and crouching because they made the whole process of killing things more varied and interactive.
Half Life and Quake 2 introduced the idea of linked levels that cohesively merge into each and allow for backtracking. Unreal introduced big outdoor levels with lots of organic touches and things like birds flying around. Also pretty big shifts. Level design matters a lot in... well, any game, really. The nature of level progression changed a lot, though, with Half Life and Q2, as you had OBJECTIVES. it wasn't just find key, open door, go to exit. End. And the terrain of the map became more important to how you played through it.
Also, the move away from hit-scan weapons to more physics oriented weapons also changed things pretty massively, since in Doom, point and shooting was pretty much all you had to do to shoot an enemy, whereas by the time you got to UT, you only had like three hit-scan weapons and everything else could be ricochetted off things or bounced around corners or took time to hit the target, which introduced both a level of strategy and a certain need to cope with, well, rockets being slow.
The funny thing with FF is that there's no real core underlying 'ruleset'.