Hmm. This discussion is NOT unique. I recall the same discussion in a Thief thread in another forum. Essentially it was about whether something was lost when all items on the screen are glowing? Some posters commented in the original Thief items didn't glow and important items could be buried in a desk. Some said they liked it that way, but it seems most agreed it just led to try-every-single-container-posible and were thankful it was abandoned.
That's the problem with bad and thoughtless design. Rummaging through a hundred of identical desks/crates/barrels is neither challenging, interesting, nor fun.
The problem isn't whether this hundred of containers is searchable or not, though. The problem is that there is no external information to be obtained regarding their contents. Player should be safe to assume that a typical barrel contains only typical barrel contents and not, for example, a +5 sword, unless there exists some external clue that such a sword has been hidden in some particular barrel. When this "contract" holds, player will be able to treat barrels as generic sources of typical barrel contents rather than potential sources of easy to miss phat lewt. When such contract fails you're basically patching one failed design with another.
Well hidden items are not those hidden in unreasonable places. They are those hidden in perfectly reasonable spots player won't think of.
You know I think this boils down to minimaps, radars, automarkers on maps, glowing paths, linear maps and other such things. I may get flamed for saying it, but I think the crucial element in exploration is NOT being shown where things are and furthermore not being excessively nursed by the game designers with the intention you follow the correct path and hence don't get lost.
Basically this.
Pretty much any mechanics is made or broken by how involved it is (or how involved the bigger whole it's part of is, but then we can't really say that the game X has good Y, if Y's only role is being a component of some Z, possibly as insignificant as an attention sink minigame - can you say that Wizardry 8 has good lockpicking system or System Shock 2 hacking even though they do their job in those games?).
When it comes to exploration the involvement generally comes from interpreting information.
So compasses, automaps, highlights, and so on have ability to destroy exploration that directly correlates with their usefulness for it, as they are all gameplay aids (not be confused with AIDS, although, come to think of it...) translating complex and often ambiguous information you get from looking at environment, into basic stimuli such as simple, easily recognizable tokens.
When you're told to find something incriminating NPC X, you have no precise idea what it might be, so you basically have to take all your environment in and constantly make and modify hypotheses regarding what might it be and where it might be hidden - is it a document? in desk? *behind* desk? hidden in shoe? it might even be scratch mark in some particular spot.
Enter quest compass and you no longer have to take anything in, just note relative position of pizza slice, approach and click on whatever is pointed by it.
Similarly, when exploring in raw 3D or pseudo 3D you don't really have any idea regarding what "interesting" means. It might be anything - some particular detail about environment, seemingly unreachable scenery, cryptic textual clue, obscuring clutter, or some peculiarity in large scale layout of the area. You have no shortcuts and no luxury of filtering information out as irrelevant.
OTOH, if you take something like BG, you have very clear idea of what constitutes "interesting" - blackened areas indicating unexplored terrain and anything that flashes teal or changes your cursor on mouseover. You can parse everything else out, especially given that not only are aforementioned stimuli reliable and unambiguous, but anything else can be just not helpful - an area of interest may be an obvious container like barrel or chest, but it may just as well be several pixels in the middle of copypasted rock, while many obvious containers may fail to be hotspots.
tl;dr
IE style makes for good (or any) exploration if and only if QTE makes for good (or any) combat system.
In both cases relevant information is reduced to simple, unambiguous stimuli and relevant gameplay to conditioned reflexive response to them.
If you adore one when condemning another, you're a hypocritical moron, if you adore both - just the latter.