@ pvonk:
I would agree with you about combining the items tagged ‘Paris’ in your example organizational structure. In displaying ‘group tags’ DEVONthink separately lists each group named ‘Paris’.
But it’s not that simple. Many users have organizational structures in which there may be many groups with the same name, but containing quite different content. For example, a teacher might have a database with groups defined by topics, such as History, Geography, etc. — and within each topical group subgroups such as ‘Overview’, ‘Goals for the Semester’, ‘Course Outline’ and ‘References’. That organizational structure makes sense to the teacher, and it causes no problems for DEVONthink’s Classify feature in suggesting appropriate locations as new content is added.
But in the Tags view, each such group is repeated multiple times in the list. Yes, that’s confusing. But simply to ‘meld’ those subgroup tags by listing only one and combining all the contents of groups with the same names doesn’t appear very useful, especially for a group tag such as ‘References’, where the context of the enclosing group is important.
Perhaps another approach to simplifying the list of group tags might be to display only a single group tag for multiple instances, but to ‘split it up’ if that tag is clicked, in such a way that context could be retained.
Of course, there’s a second kind of tag — user-created and imported OpenMeta tags — where one should avoid using the same name multiple times. This is the kind of tag that I think you have in mind. In the final release of DEVONthink 2.0 one can choose, in Database Properties, to exclude ‘group tags’ from view for that database. Here, a tag such as ‘Paris’ can be applied to any item you choose, perhaps in conjunction with one or more other tags. If you are studying the sewers of major cities, for example, items about the famous sewers of Paris could have the tags ‘Paris’ and ‘sewer’ attached. This would be totally independent of the locations of items in the database’s organizational structure.