Getting started - grouping by content or type

I’m setting up my first DTP database with the goal of combining and organizing documents collected over time from several computers. The documents have various topics, file types, uses and reasons for being kept. Most of it is notes and practical knowledge in several subject areas, but there are also many smaller categories of info. Not having my head completely around devonthink quite yet, I was first operating under the assumption that documents with related content should be carefully grouped together, combining dissimilar file types in subfolders as they relate to a common subtopic. But in looking through the tutorial database it seems that organization is generalized by subject or usage at the upper levels, then subdivided by file type at lower levels. And in the paperless office sample db the upper level seems to be by source with little consideration given to subject. Having looked through the sample databases it seems that there is organization by intended use, content, origin, doc type, and output with methods often being mixed within databases. Does grouping affect the way information will be searched, or does indexing erase all boundaries and negate the need to base structure on content? Is it more important to just build an intuitive structure that makes it easy to classify new material without having to analyze content relationships each time? I’d love to hear discussion from more experienced users on the strategy of building the most effective folder structure from the start.

I’m new to DTP too, but I have indexed my Documents folder into DTP, and after trying my default folder structure (folders for different subjects) I went back and took almost everything and put it at the ~/Document level - I only have a few sub folders left. Instead of creating “hard” folders I now have several smart groups in DTP to get at what I need. HTH.