I am brand new to DT3 and am trying to figure out the best way to integrate it into my Zotero workflow.
Right now, I mainly download academic articles to Zotero using the Zotero Web Connector plugin for Safari. This both collects the citation in Zotero and the article PDF.
The major change I want to make is to be able to search through PDFs in DT3 (since searching within Zotero is pretty bad) and use the “See also & classify” feature to find similar articles in my database.
However, I’m not sure whether to index or import my Zotero PDFs to DT3. I’ve read the user handbook, and it sounds like there are two main reasons not to use indexing, namely the risk of accidentally deleting something and the risk of “deconsolidation”. On the other hand, if I decide to import, then I will need to move files downloaded via the Zotero Connector to DT3, either manually or via a Hazel rule.
However, reading through posts here (I’ll paste examples in a separate reply) there doesn’t seem to be a consensus about which practice is better for this particular use case. As a result, I’m feeling a bit of analysis paralysis and would love advice on which path to choose. Thanks!
The index-or-import debate has no final conclusion. DEVONtech’s articles and Jim’s postings here have gone into the guidance in depth, but ultimately you’ll need to decide your use case, if there are any risks, and how you’ll make sure to avoid inadvertently moving or deleting indexed documents. Lot of us have opinions; none of us have the answer you need.
Personally, I have indexed certain document collections for a very long time. (What they are is irrelevant.) I have never had a lose of data, or an inadvertent deletion. I have never had an issue requiring restoration from backup – though no matter what you do, make sure to backup backup backup.
Follow Jim’s detailed advice, and you’ll have the same good fortune. Also, there’s nothing stopping you from experimenting. Just don’t ask the community to decide for you, because your use case is just for you.
Thank you for the response, that’s very helpful to hear. While I realize that there is no single right answer for this sort of thing and that a community can’t decide what’s best for me, I just wanted to make sure there weren’t any considerations I was missing in terms of the pros and cons of the different approaches. But it sounds like as long as I back up my files frequently, neither approach is much better or worse than the other.
I have more recently come to terms with this analogy. Imagine your files are spaghetti. You have two apps to consume (work with) the spaghetti, DT and ZT. You have the following limitations:
ZT puts everything in its own bowls.
DT can either respect what ZT has for its bowls (indexing in DT), or DT can copy what ZT has in its bowls and store copies in its own bowls (importing in DT).
Indexing is a two-way street, and it is always on active mode. If ZT puts salt on the spaghetti, DT will taste the salt. If DT takes the meatballs off the spaghetti, ZT will not find the meatballs any more.
With indexing, DT will not mind have ZT put salt on the spaghetti. ZT may however throw a nasty fit if the meatballs go missing because it requires meatballs to even begin to consume spaghetti (and fish and steak and carrots and …).
Importing is a one-way, one-time thing. If ZT puts salt on its spaghetti, DT will not know. If DT takes the meatballs off its spaghetti, ZT will still have its own meatballs.
With importing, DT and ZT do not care what each other consumes or how each other arranges their spaghetti in their own bowls. When you want them to care (when you want their spaghetti to be exactly the same, let alone have some similarity in the arrangement of bowls), you have to (manually) have them re-connect (sync/import) with each other. This can meant that ZT becomes the master in defining what is on the spaghetti and how the bowls are arranged. Alternatively, you have to get rather good at setting up instructions (read AppleScript code) to enforce that DT takes things from ZT only in certain ways.
And by the way, the way that ZT uses to annotate is not the way that DT uses to annotate. So, you have to deal also with spaghetti with the sauce on the side (ZT) versus spaghetti with the sauce embedded (DT).
As others have said, you get to pick your own poison on this one.
You don’t mention what other tools you use in your work, and those might influence the choice you make. These days I make rather a lot of use of Obsidian and there is a plugin that allows you to turn highlights in Zotero pdfs into Markdown files in Obsidian. I index both my Obsidian notes and Zotero pdfs in DT. It gives me various different ways to get to the information. However, as others have intimated it really is a personal choice.
I do use Obsidian for notes (or am trying to get into the habit at least ), but I have never used Zotero’s built-in pdf viewer since I prefer Apple Preview (though perhaps not I’ll use the DT3 pdf viewer).
In terms of workflow, I was thinking something like this:
Web article Zotero connector Zotero (to download the article PDF and capture citation info in Zotero)
Zotero reference info Better BibTeX .bib file (to make citekeys available in LaTeX)
Downloaded PDF from Zotero Zowie/Hazel? index in DevonThink
Read and highlight in DevonThink PDF reader (or maybe just stick with Preview)
Take notes in Obsidian
Index Obsidian notes in DevonThink
If there is a way to export annotations from DevonThink Obsidian, that would be amazing. I know that the “Summarize Annotations” tool exists in DT3, but it seems like would take a lot of processing to get those into obsidian since each individual highlight gets a DT3 url like “x-devonthink-item://8582F393-CFAB-4752-984B-3BD255323D63?page=0&annotation=Highlight&x=318&y=482”. Imo this makes the annotation summary a little difficult to parse.
As an alternate to @korm, I’d say that both approaches are equally tedious. But then, if given the need, I would prefer working with my spaghetti using a spoon and fork (Bookends and DT) rather than a knife and fork (ZT and DT).
I am working with the author of this Zotero translator - he has done lots of work with Zotero in the past.
My goal is an easy way to export individual Zotero citations or entire Zotero collections or library in a nested folder/file structure. Such a structure would easily be added to Devonthink.
Thus this would be a way to export not just individual PDFs but the entire Zotero record for each item - notes, annotations, PDFs, metadata, etc.
Once this is done I don’t see any benefit to indexing. And I think it would be superior to what can be done now with Bookends export.
The app is a work in progress but I anticipate it will be completed within the next few weeks at the latest.
A second reason for me to choose Zotero over Bookends is the well-regarded Notero integration between Zotero and Notion:
This lets you sync your Zotero collections to Notero, where you can create a custom database including AI queries to summarize or search or query or rebut Zotero citations. It’s an extremely useful tool.
Wow, this looks like exactly what I’m looking for—I can’t wait to see the finished project! Would this also automatically copy Zotero links to the imported copies in DT? I know Zowie already does this but it seems like a useful feature to me.
Thanks for the time you’ve put into developing what looks like an amazing tool
I won’t stomp on this thread that has significant questions about the pros/cons of indexing versus importing regardless of which app is on the other end to pair with DT to dwell on my reasons for BE versus Zotero. Spoon/Fork versus Knife/Fork. Plastic versus balanced stainless steel cutlery. Whatever helps you enjoy your spaghetti meal the most.
Just to throw some extra dust in the air, there is an Obsidian plugin called DEVONlink that allows some integration between the two programs:
I tend to think that when one settles on a workflow or a program, one is basically choosing what kind of irritation one is prepared to put up with. None of them is ever perfect, and one can usually arrive at the desired final outcome (book, article, etc.) whichever method is used. Sometimes it even comes down to quirks of the user interface, I believe. I know there are some very good programs out there which I don’t much like using because of the interface.
Never worked with Zotero, just with Bookends. So, I can say what I love BE for:
Deep integration with Mellel (powerful word processor for manuscripts, with live bibliography)
Wide AppleScript support
Very flexible formats manager
I import all articles in DT and connect them with corresponding refs in BE. Script does it all in one click. Another smart rule is run on a timely basis to sync them if there are any changes from both sides.
The purpose of importing over indexing is not just search and stability - it is organizing power of DT: groups, tags, reps, annotations and etc. Many of it is impossible or may brake if you rely on indexing.
Interesting - I’ve looked at Mellel but never quite saw its advantages over Word except for languages which are written right to left. Is that a main reason you use it or are there other major benefits over Word not related to language choice?
live outline (reordering the outline, reorders the text), allows additional metadata for outline topics (notes, cross-ref targets)
powerful search and replace with a very nice integration of RegEx
live bibliography (which I told in the post)
different approach to styles, which make it possible more automation with Streams (auto-titles, footnotes and endnotes streams etc)
One of the most important things apart from Bookends integration was a pure XML realization. Before I turned from Word to Mellel, it was a mess to make a nice document in Word more than 100 pages - it was heavy, slow and with some polterGates unexpected things, that could drive anyone crazy, esp. when you make some more or less good formatting, citing, cross-ref and other word processing features. I don’t know if it is already changed with the Word, but since that times I use Word for short-writing and compatibility. Long academic writings goes to Mellel.
As for me, it is a more niche, but more professional word processing tool.
), but I have never used Zotero’s built-in pdf viewer since I prefer Apple Preview (though perhaps not I’ll use the DT3 pdf viewer).
Zotero connector