My Journey with Groups and Tags - To be continued

In this post I want to describe my little journey through the lands of grouping and tagging. - The old question remains: Should I file my records into groups or should I tag the data?

I started out with a mixed approach: Some high-level groups (like “Home”, “Travel” or “Formal”) and everything else I categorised through tags (like “to-do”, “notes”, “work”, “to-read”, etc…, ). This worked well, because the few groups give a good overview over the data, while the tags do all the detailed categorisation.

I had told myself to file each item in only one group, which was strategically satisfying, but got me in the real world into trouble: Some items I had filed in one group, but later I assumed they were in another group. (A travel insurance PDF I had filed under “Formal”, but later looked for it under “Travel”.)

My second approach was using only one group and organise everything with tags. I used one master group and not the top level in order to have the option to “collapse” my items away. All high-level groups from before ("Home, “Travel”, etc) became now tags. I structured the tags hierarchically, which was a great tool to auto-tag an item with multiple tags: E.g. I made the tag “Canberra” a child tag of “Australia” which gave me the benefit that everything tagged “Canberra” automatically is tagged “Australia” too!

Playing around with this scheme I realised I could now use groups for something else which they are very good for: Grouping similar items together. So for example I created (inside the one master group) a group for an event where I had lots of photos of plus a story document in Markdown. Or I have the webarchive of a special turntable and a review of this turntable in a group with the name of the turntable.

This is my current state, but I feel it isn’t just right either: When looking for items in DEVONthink it appears to be quite tedious to find tags in the long, long tag list given taht you have many tags. I haven’t come up which anything better yet - so the journey continues.

I use a similar combination of tags and folders. It doesn’t feel polished, mostly because - like you say - it is rather cumbersome to retrieve tagged items in DEVONthink. Yes, we can make saved searches (called Smart Groups), but still I’d like to see more features to make use of tags. Things like suggesting tags, for example, or better use of tags in ‘See Also & Classify’.

What helps me, however, is:

  • Put tags in tag groups, like you suggest, but then exclude those tag groups from tagging - so a child tag doesn’t automatically imply a parent tag. This helps to limit the number of tags on files, while still allowing me to organize tags.
  • If you want to have an item in multiple locations, use duplicants.
  • Use short-cuts to view an item (and its duplicants) in its different locations (including revealing it in tags), with ‘opt-cmd-left/right’.

Currently, I often link files (mostly markdown notes) together by manually including links (i.e., devonthink-item://). This also works to link to groups, smart groups, and tags.

As long as you understand, you are no longer Tagging with this excluded group.

Yes, and I too do this with my tags. As example, I have a database that has a heavy date context. I want to tag years, but I don’t want to see many years of tags at the root level of the Tags group. I put them in a group named /Date and then exclude /Date from tagging. As a FYI, I use the ‘/’ prefix to name groups in the Tag group that have been excluded from tagging so I can tell by the name that the group is only for organizing tags. It’s helpful to me.