What does DT examine when it looks for “See Also” recommendations? And for Classify?
I think that “See Also” just looks at the contents of each document (rich text document, plain text document, PDF, sheet, email message, etc.). “Classify” looks at the contents of individual documents, as well as the contents of ALL the documents in each group. And neither “See Also” nor “Classify” looks at the tags of a document, or any other metadata, such as dates, flags, unread vs. read, labels, etc.?
That’s correct. Matches are suggested by…
See Also = content
Classify = content and location
Since version 3 document properties and (custom) metadata are also used.
How do I not know this?
What non-custom metadata is used, things like Author ?
For both “see also” and “classify”? Does that include tags?
In both cases but tags don’t matter.
What document properties does “see also” and classify examine?
See e.g. Document Properties inspector.
Does this mean if I write tags into keywords (via exif tool) then they are used for classify?
Yes but only PDF/RTF support editing of keywords.