I would go so far as to say all books should be read in an iterative manner, fiction or not. Good stories will reveal previously-missed tidbits again & again as you explore the deeper structure of the work.
Deep reading should be saved for books worth the effort.
There's far more written than you'll ever be able to read. The cost of discovery is a large impediment to that.
Skimming / scanning (contents, index, bibliography, intro/conclusion chapters) is part of the informational assessment. If the book holds up to that, keep digging into it. If not, set it down, and feel no guilt.
I'm reading a lot of material, though I'm not satisfied with my methods or progress. I've found Adler's book highly useful. The principle problem is that my interests have both breadth and depth sufficient that the volume of relevant material is enormous, and quality assessment is complicated. I'm doing the best I can.
I should clarify- I fully agree with your assessment. Sturgeon's law holds true with books as much as anything, and there is plenty that can be tossed aside after only a cursory glance.
My point was aimed along the lines that if a book is worth reading once, you will benefit from rereading it at a future date.
- First pass: Understand the structure, important words (from index), some interesting passages.
- Second pass: Read it all the way through, don't worry if you don't understand stuff
- Third pass: Look up stuff that you didn't understand from the previous pass and read it again in order to understand them.
- Four pass: Read it concurrently with other related books to try to truly understand the topic