In OD&D, only a third (a roll of 1 or 2 on a d6) of the rooms contain monsters. Of the monsterless rooms, 1 in 6 of those contain treasure, which should be hidden or otherwise hard to get at (in a trapped chest, for example).
So what is the purpose of all these rooms that contain nothing? Are the PCs expected to search them all, in hopes they may contain hidden treasure? Do they look interesting with furnishings and statuary and so forth or are they featureless?
First, I think you're making a mistake in assuming that the only interesting things in a dungeon are monsters or treasure.
Second, it's a matter of pacing. If every single door you open has something exciting about it, then there's no uncertainty. It's as if every single door on
Let's Make a Deal had a new car behind it. The "boring" rooms provide context for the interesting rooms. (And don't really chew up that much time.)
Third, even empty room can be strategically valuable to the PCs in their explorations. Does it provide a short-cut allowing them to bypass that tribe of ettins? Can it be secured to give them a place to rest and recuperate? Et cetera.
Finally, resolving empty rooms doesn't really take that much time. Even if the PCs decide to search it high-and-low, if the resolution of an empty room is taking more than 5 minutes then something is wrong.
And even if we (a) conservatively estimate that your typical combat takes 6 times longer to resolve than an empty room; and (b) absolutely nothing interesting happens outside of combat -- then you would still need 54 empty rooms for every 1 room containing monsters or treasure to give you "20 minutes of fun in 4 hours". In other words, non-empty rooms would need to make up only 1.85% of the dungeon. Which is considerably lower than the 44% suggested by the OD&D guidelines.
I agree that this is the fastest way to get the players to what the DM determines as important. In other words it would be more like how RE Howard, Tolkien or Clark Ashton Smith would describe a ruined city -- from an authorial point of view.
And the problem with the approach is that it
is authorial: The DM is taking control of the PCs away from the players.
Now, a friend of mine is experimenting with designing the ruins of an abandoned city using the structure of a hexcrawl. He's also talked about using a similar technique for large underground complexes with lots of empty rooms and very few points of interest.
IOW, if you've actually got a complex where only 1% of the rooms are interesting, then the format of the traditional dungeon crawl may not be the best way to handle it.
But I'd still want to adopt a format where the PCs are in control of their expedition, not the DM.