It appears that the item counts for tag groups (those sporting the yellow icon) do not factor in any tagged folders. You can verify this as follows:
Create a simple group.
Add, say, 5 PDF documents to the group.
Tag the group itself (not each document individually) with a new tag name (e.g. “test”). This now adds up to 6 tagged items overall (the group itself plus the five PDFs it contains).
Note that the yellow “tag group” for “test” will indicate only 5 (assuming that counters are enabled in the DT preferences).
For comparison, create a smart group to display all items tagged as “test”. The counter for the smart group will indicate 6.
Not a catastrophic issue but a small and potentially confusing wrinkle. (I for a moment feared that my data is corrupted.)
I don’t think it’s a bug. IIRC, when tagging was introduced there were discussions back and forth about this question of including groups in the count badge for tags. When we “tag” a “normal” group we are making the child documents of the group inherit that tag – as long as the document remains a child of the tagged group. Because of inheritance, I think the logic was “when I look at a tag I want the badge to tell me how many documents are tagged”. Groups are not documents – they are containers. It’s a valid design choice.
BTW, with the same logic, the badge for “normal” groups also does not include groups. In the example shown in the image, each group contains one document.
I could see the possibility of a option to include groups in tag badge counts, but I don’t recall it being a frequent request. Personally, I wouldn’t see any information content in knowing how many child groups a parent has – either for normal or tag groups.
@korm, thanks for the good example. Fair enough. But in that case, why do SMART groups (with a tag-related search criterion) actually include groups in their badge?
Because Smart Groups are searches – finding things is a different data structure than listing things – and you can configure a Smart Group to exclude groups with the predicate “Kind is not Group”.
Well, this is becoming an ontological discussion, but I would argue that things are things and that, given a criterion for selecting a subset of things, the number of these things should not depend on whether one “lists” them or “finds” them. But I rest my case. Thank you for clarifying how badges in DT work.