There are (wonderfully talented) people here way more groomed on GTD and other organizational methods and such than I, but I think the challenge with what you are asking about is universal. I’m still working on yet another refresh of how I approach this.
Folks want to simplify (time and effort spent on) storage and access to data, while freeing our brains for easily and readily tracking anything, and all their relationships, and without stuff falling through the cracks and becoming lost or forgotten. (can be a tall order ;^)
That said… my conceptual take on this (fwiw) is:
I would suggest that there are essentially two ways we are handling data storage and (ability-to-)access management:
- hierarchical ordering: boxes with boxes of stuff, i.e., folder/sub-folder chains, and
- tagging: using various methods for flag/tag -ing things to empower access to things using filter-searches.
Like Louis_Kirsch is inferring …its all about best practices for leveraging good search-filter tools, assuming the data presentation method of the app also serves effectively.
Realistically there’s a bit more to it than that when you consider PIM (personal information management) framework includes things like calendar, mail and other communications channels [phone, texting, social channels, video conferencing and whatnot], notes and documenting methods [including prose, tabling, charting, etc], reminders, lists, research tools/moduses, and the like.
But getting back to terse-basics (to my mind anyway), each method (hierarchical + tagging) have their strengths and weaknesses. I think the bottomline is the better your search~filter engine tools are, and methods for leveraging that, then the less you have to rely on hierarchical methods (stuffing things into boxes within boxes promotes an unfortunate out-of-sight / out-of-mind phenomena, more work to keep organized and more-tendency-to-forget-and-lose-stuff situation). Tagging systems, then, should free us up to more free-form data storage and unhampered full-access, …which then should free us up to make boxes (folders) only when it “feels good” to do so.
With “listing/outlining”, there is a hierarchical framework for setting data relationships that is already like using folders/boxes. So if you realize certain data relationships already exist during outline/listing, then that’s a cue (to me anyway) for allieviating the need for yet-another-box/folder (or even a new separate list) …as long as my entries properly/effectively leverage tagging of course.
So while this doesn’t necessarily answer the tough question, I figured maybe it helps to talk about how I am approaching understanding what may lead to a best practice, and then thats hopefully good additional ammunition for the big gunners who may drop in and comment and critique ;^)