Advice on using Devonthink to Organize Article and Book. PDFs

How to people use Devonthink use to organize PDFs? I’m trying to see how I can make use of Devonthink to help with the storage of PDFs. What features would you recommend I look into? Are there any advantages to using this with software like Bookends vs alone (I currently use Papers)?


I use Papers extensively as a citation manager. Devonthink will only compete against a citation manager to the extent that both store PDF files and both can allow you to annotate. The ability to use EZProxy to pull a PDF directly and the ability to see comparable references directly as via Papers is missing in Devonthink, as are any capabilities to build BibTeX exports (i.e. for LaTeX compiling of references to a journal article or proposal).

I use Devonthink to index a hard drive of files from over a decade as I seek reference files to prepare a textbook. This is where Papers will fail by comparison.



I think the advantages of a reference manager over Devonthink are huge. You get data structures that are pre-organised properly for the data you need to store, integration with search by DOI etc., integration with your writing tools, ability to search e.g. by author without relying on OCR, autofilling of linked references, export templates, keyword coding and search etc. etc. I use BibDesk but all this is pretty generic.

I’ve tried dumping all my article PDFs (copies) into Devonthink on several separate occasions over the last 10 years, and it never produces anything of value. I think if you’re filing receipts and utility bills DT’s “smart” features are probably helpful, but for semantically rich and complex content it seems pretty useless. To give an example, it decided that all my documents downloaded from JSTOR were closely related, presumably on the basis of the JSTOR cover page which has absolutely nothing to do with the content of the document.

1 Like

I started using DTPO before a reference manager (Bookends in my case), and was therefore already deep into DTPO, which saw me not move over to BE for the actual ‘management’ of my PDFs.

Had things been the other way around, I might have never ventured into DTPO.
That said, the obvious benefit of DTPO, is that it manages virtually everything on my Mac, so its my default ‘start location’ for everything work, and research, related.

But, in only speaking to my research workflow, as a result of this, my approach is rather superfluous, in that I essentially have 2 copies of all my research PDFs.

Since I am a heavy annotator of my PDFs, I was met early on with the ‘problem’ of wanting to share articles with colleagues, but not wanting to send through my heavily annotated versions.
Again - my approach is no doubt an outlier - but I drop the original ‘clean’ PDF into the BE library, where I either fill in the necessary, if BE cannot autocomplete it.
DTPO is also watching that folder, so as soon as the PDF drops in, DTPO’s window pops up, asking me where I want to place the PDF.

So I have one primary Folder with all +/- 3500 PDFs (in their original state) - which is also synced to BEoT on my iPad, and I have my DTPO DBs, where the PDFs get stored and annotated.
I do zero annotation, keyword, tagging, noting inside BE – that all happens in DTPO or DTTG2.

That said, DTPO allows me to splice and dice those PDFs in ways that – for me – are not simply possible inside BE, primarily through the use of Replicants, Smart Folders, tags and the various scripting options that are available (i.e. RTF content files that link back to a multitude of ‘grouped’ PDFs).

Oh, and because everything inside the Pro Office version is/can be OCRed on the spot, I find the ability to be able to search the content of everything to be particularly valuable. It’s something I guess we all take for granted, since it is so ubiquitous now, but it is a point worth mentioning.

**A final point that I don’t recall seeing spoken of here, is a feature inside DTPO called “Workspaces”.
Not sure how many Reference Managers have this, but it is something I use all the time.
Essentially, if I am focusing on a particular aspect for my writing/research, and I have opened 10/20/30 files/PDFs, that I am working through – and have set up my various views/search criteria, you can ‘add’ that configuration as a ‘Workspace’, which can be individually named.
It then appears as an ‘item’ that can be selected under the ‘Go’ menu, with a shortcut key allocated.
At any time in the future, you can select that same Workspace, and DTPO returns to that state, with all those PDFs and views in the same condition as they were when you created the Workspace.
And it can be updated, to allow you to ‘add/remove’ open files etc.
The usage possibilities of the above, are endless - and apart from research-based Workspaces (for ‘active’ topics I am busy with), I also have them set-up for specific work related tasks, so that all is ready for me, at the click of a button, to ‘enter data’ for often-repeated tasks (like interviews/meetings/budget review/performance review etc.)

So my paradigm sees DTPO as the place where I work in my research, whereas BE is the place I work with my PDFs, if that makes sense.

When writing over in Scrivener etc., I then use BE extensively to insert the temp citations etc., and over in Word, I use it to produce the final product, and if required, the bibliography.

Obviously, any number of Reference Managers can do all of the above, without there being a need for DTPO. But in my case, I quite prefer the notion of having two separate repositories for everything, and it works well (for me).

@dfay’s comment on his not finding the AI inside DTPO to be particularly useful, is completely valid. I didn’t have much use for it initially, but as I started splitting up my DBs, into specific, refined stores (even if temporarily for the purposes of specific projects), I found it started working quite well.
It’s certainly not something that I would cite as a primary reason or motivation for using DTPO as a research tool, but it has yielded interesting results, drawing me into a related article that it suggested, that I didn’t initially consider, on quite a few occasions.

That said, it certainly does come into its own when I am managing all the other areas of my data/information, with the Auto-classify features having saved me several hours (I am sure) by large ‘automating’ the placement of frequently stored files etc.

Bottom line, if you are happy with what a Reference Manager offers in terms of ‘access’ to your research information, then there is surely no reason in using DTPO. It is, after all, largely just another information manager.