I use Devonthink as the entry point in my scientific scholarship, reading and annotating, and then using it to find relevant citations when I’m working on a new project.
I update my Notebook by importing my new PDFs every few months, then assigning them to one of 8 categories, each with 3-10 subcategories. The effort to classify (even using the auto-classify feature, which helps a bit…) takes hours.
Question: Does this help the AI? IOW, when you index, does this information get attached to the item?
Years ago on this forum, the answer seemed to be “no”. If that is so, is their any utility to creating these folders?
Welcome @MikeKaspari
I update my Notebook
Do you mean database ?
then assigning them to one of 8 categories, each with 3-10 subcategories.
Do you mean groups and subgroups or are you referring to tags?
The effort to classify (even using the auto-classify feature, which helps a bit…) takes hours.
Why are you waiting months to put data into DEVONthink? Smaller incremental additions/modifications are more efficient.
Does this help the AI?
See the Help > Tutorials > Understanding the AI.
IOW, when you index, does this information get attached to the item?
What information are you referring to?
-Database (called “notebook”)
-Groups and subgroups (i.e., folders)
-Because I don’t start new projects every week. And there is an approach avoidance conflict given how much work it goes to assign 80% of the PDFs by hand, even using the classify function.
-If I type “Understanding the AI” into Help, I get directed to DEVONthink Simplified. There is no subheading or direct discussion of the AI here. Could you send me a link?
-Does the name of the group/subgroup inform the AI of an PDF’s content and relevance to any search I run? For example, if I type in the search bar
“Insect declines are associated with higher temperatures” does DT automatically target any PDFs in the Abundance:Insect Declines group/sugroup as relevant, then refine and rank output using DT’s word concordance?
-OR is this the equivalent of adding one byte of metadata to an item that has 1000 bytes already? If so, then all the effort to forming and maintaining subgroups is probably not worthwhile, and I just stick with AI’s word concordance feature.
-In the same vein, in the discussion of groups, tags, and group tags, it hints that assigning tags does go into an item’s meta-data (which presumably is used by the AI?), and that you can also assign tags to groups. My current setting in Database properties has checked “Exclude Groups from Tagging”.
-But then it says that “Groups tags are applied by DEVONthink automatically.” So am I correct to infer that if I added the tag “Abundance” to the Group “Abundance” that would be redundant information?
Thanks for the help with this. I’ve been using DT for at least 15 years but must admit this question has bedeviled me from the git go.
Thank you!
Mike