I have a copy of the Adobe PDF Specification document (as a PDF) in my DT database. I have noticed that the AI associates every PDF I import to the reference document, regardless of content. The PDF reference document is usually one of the first 3 matches in See Also and the folder the reference document is in is usually the first suggested folder in the Move To window.
This applies to vastly different (by content) PDF files and seems to be making the association only because the new document uses PDF codes which are documented in the PDF SPEC document. For example, the PDF manual for the BMC Remedy Help Desk package isn’t at all related to the PDF specification, but DT’s AI thinks they are.
For a time I had a PDF of the owner’s manual for an automobile in a database. It was hundreds of pages in length.
It was a favorite of See Also, as it covers almost any topic one can think of. Air quality, chemical hazards, climate, energy efficiency, music, GPS, maps, metallurgy, regulations, electronics, fuels, Internet, wireless communications, voice recognition – and many more. Although it was in its own group, it also attracted Classify suggestions for a wide variety of documents about a great many topics.
In DEVONthink 2, one can exclude a document like this from Classify and See Also, or even from Search. That’s what I did. Whenever I wished to look at that owner’s manual, I knew where it was. Later, I moved it to a separate database that contains several other owner’s manuals, from my coffee maker to my TV, etc. Problem solved. I don’t need Classify or See Also in that database, but it’s a convenient way to keep a collection of such items along with related material such as warranties, purchase invoices, maintenance records, etc.