Looking at a real world scene directly, with our own eyes, is very different from looking at a scene that is pre-filtered for us by photography.

Our eyes don't function like two traditional camera lenses.

Our visual cortex, and various levels of our brains, is what provides focus, really...

...and that focus is generally highly dynamic... and non-linear.

http://www.cambridgeincolour.com/tutorials/cameras-vs-human-eye.htm

So I'm with AnAardvark in preferring to avoid too much forced focus, if any, in an image.

Generally, if I try to implement any kind of depth of field effect, it is in post, using blur filters... and it is non-linear... by which I mean, it is not objectively photographic, or physically accurate (relative to how a photographic lens would focus)... but rather, more subjectively applied... and usually with multiple focal points, compiled / contradictory depths of field... etc...

...because the effect I want is something more like you would get from being in the scene directly, in person...

...if that makes sense.

Generally I like to leave plenty space for the viewer to look around and focus, themselves, on each and every element I've taken the time to cram in to the little bubble of made up reality I've portrayed.

For me, that's a more "realistic" end result than if I tried to simply make a facsimile of a photograph.