I suppose there's a lot to be said for having a concrete idea of presumed context. I've often found this to be an irritant in fiction: the author provides just enough detail for me to get a good mental image of the characters and their context, but just when it's relevant, new details emerge. This isn't along the lines of "she saw the reflection of her blue eyes in the pooling rainwater" when all along I had thought the character had brown eyes. It's more like having the rainwater appear in the first place when I had understood the weather to be clear. Is this the fault of my own assumptions, or can I reasonably expect authors to provide me with enough detail to minimize assumptions?
The same issue is true with writing scientific manuscripts. That's what I'm doing now. It's a collaborative job, so at any point a fellow author may add or remove details and data. The collaboration is helpful when it works toward the same goal but that goal often remains blurry, a granite crag looming in the distant hills.