Are there any issues with (more or less) having all of my files in just a few folders (based on years) and then creating groups/ relationships between similar, ie by kind/type or subject/use case, using a smarter and less hierarchical approach.
Im thinking of weighing up the cons here in terms of speed/ stability more than anything else.
You say “folders” which implies you are indexing files. If so, then you are a little bit beholding to the performance of macOS.
I suspect you mean "groups (based on years) and the files are imported. I have minimum groups and little hierarchy with “smart groups” to do the organising for me. Works and I’m sure there are no downsides, really, as DEVONthink designed for that sort of thing.
It’s really down to you and how you want to view, interact-with, and perceive your data.
I tried having minimal groups and using tags and smart groups. I found I soon ended up with as many smart groups and tag hierarchies as I would have had using subgroups. I ended up with balance between the two. You can use the power of search in the minimal groups scenario, but you have to be strict on naming, tagging and metadata to avoid constantly trawling through long lists of hits - no quicker than navigating down a logical group hierarchy.
These things are very much individual. It depends on how different brains work, as rmschne points out
The groups you see in Devonthink are not OS folders, at least in the case of non-indexed locations.
For instance, if you have the option set to put all your annotation files in one group, they won’t all be in the same OS folder. Devonthink manages how many files go in any one OS folder without regard to how many documents are in a group.
Moving a file from one group to another generally doesn’t change its location in the OS. Even using the “move to trash” function doesn’t physically move the file, it just changes where the file maps to the groups you see.
Tags are really just groups with a little added logic.
Another option is to turn off the “exclude groups from tagging” feature. I don’t think that’s commonly used but it works fine. With the exclusion turned off, you can add a file to a group by adding the group name the file as if the group name were a tag.
It’s like Burger King except for all the calories. You can have it your way.
I’m not sure why this would affect speed/stability? DT indexes based on its own magic (=code) and I’m not aware that tags convey any special status over groups.
In any case, given that I seriously doubt there’s a preference from a software point of view (DT provides both options, after all, and most people use both in their databases), you should be considering this solely from a ”what works with my brain?” perspective.
There are many different options, and many people who swear by each option. E.g.
- Flat structure, rely solely on search for navigation
- Flat structure, rely heavily on tagging for navigation
- Flat structure, rely heavily on file naming conventions
- Hierarchal structure, still rely on search for navigation
- Hierarchal structure, navigate by eye
- Hierarchal structure, rigid file naming conventions, everything has a specific place to live
- Hierarchal structure AND tags, because life is short, work is complicated and sometimes we do want the best of both worlds!
And of course, I’m not even touching on the transient nature of smart groups, replicants, etc., which mean that DT can set up temporary hierarchal structures for you based on specific projects you’re working on, without affecting underlying file structure.
There’s no right answer here. For myself, I have a fairly rigid groups structure, up to 4 levels deep depending on the topic (I have a rule that it doesn’t go deeper than that). I only use tags to denote actions (high priority, write notes, etc.) and to denote file types, as that’s of value to me for my work (tags might be article, academic paper, report, note, etc.). But I prefer to navigate by eye a lot of the time and I like to limit searches to within a specific group. For others, they would hate this type of structure, and that’s fine!
Thanks for the guidance and reassurances.
As a result, Ive looked into how DevonThink stores files, as you say, it nests them in folders in its own way so the quantity of files in individual folders will have no impact on speed or stability as it is taken care of by the app itself.
I appreciate the comment regarding ending up with as many ‘smart groups’ than one would have if I were to create hierarchical groups myself.
The idea for me is that multiple files are being used as reference for multiple articles and reports and I’d like to create such groups for each article, pulling in resource and reference files accordingly each time.
I can then archive the smart groups when I’m down with each piece, without having to worry about where the original.
What a great tool this app really is!