Quite new to DEVONthink, like it a lot.
The Data > Group feature particulary interesting. It would be even more useful if there was some way to determine why things were grouped together. I’m not seeking proprietary algorithms. I would like some better indication of the meaning of the group other than the name of the first file.
The proprietary algorithms analyze the contextual relationships of each of the documents that have been selected, and if similarities are found among the terms used and the patterns of word usage, documents will be grouped together. The name of one of the contained documents will be applied to the new group as a possible aid to the human, as it will likely provide a basis for expectation of what will be found in that new group.
No human intervention is involved. Formation of groups is pretty conservative, so that most new groups created will contain reasonably obvious relationships of the contained documents, and a sizable fraction of the selected items may be left ungrouped.
It’s as simple (or complex) as that.