Should I move to DevonThink?

I have been following several discussions, and I get the impression DevonThink might be something I could use. I downloaded it to do some testing, but software like this requires to spend quite a while with it to understand if it fits your workflows in the long term.
I have some questions, and you might point me in the right direction:

  • should I use multiple databases and separate stuff (work, personal,…) or is it robust enough to handle everything in a single database?
  • How does is handle books, both ebooks and scanned (OCR) ones? I have a vast collection and never found a dependable way to manage them. I use them for research, and Spotlight isn’t doing a great job.
  • Is the mobile version suited to annotate PDFs? What would be the best way to use it to go through large PDFs?
  • I use Papers 3 for citations. I press CTRL twice and can insert a reference into a document. DevonThink seems to be better to search through papers and the feature that suggests similar ones (the “inverse top hat”) appears to be very powerful. So, how could a new workflow look? I don’t want to add papers to both of them, because in the long term this would lead to documents stored in just one of them. You know, we forget.
  1. Personally I would split work and home, so you can filter out one or the other.
  2. I use Calibre to organise my books - though these are not ones I need to search the contents of to be fair.
  3. Yes, it can. But it’s a document provider too so you could use many apps to edit the PDF.
  4. I use Zotero, where I keep the reference, and then the file goes into DevonThink. If you can store a link to the file without messing up your referencing than that would be one way, I just copy the bibliography entry into the notes field.
1 Like

I’d say no. I gave it a try again (maybe the 4th time) last night and remain unimpressed, especially for work with PDFs.

I was using it with a folder of about 800 PDFs I had accumulated with no DOIs, i.e. that would require manual intervention. After about 1.5 hours with DT, about 200 of them are now in named folders. Because of the See Also… feature I probably filed about 30% more files than I would have otherwise. But

  1. They’re only filed in one folder, i.e. they only have one keyword, and every one will need multiple keywords to be classified properly
  2. They’re still going to need manual intervention to get the metadata into my reference manager (BibDesk)

So in the end DT sped up my triage a little bit. But the first issue above is still a deal-breaker for actual use - PDF rendering is the worst of any PDF viewer on my Mac (Preview, Skim, Highlights, Adobe Reader) and unusable in some cases.

Also it doesn’t display epubs at all, just displays an icon, though it does seem to index them.

And when you triage, and use right-click > Move To to put a document in an existing folder, you just get a long list that you have to scroll through, no autocompletion or suggestions.

Echoing Rose (whose workflow is very similar – also a zotero user, also do not store PDFs in Zotero, rather I keep them in DEVONthink):

Aboslutely separate work and personal in different databases. This will help the “AI” work better and if you are using mobile, you may want different syncing settings for work-related databases and personal databases.

I am surprised that you find PDF rendering so poor. The DT app on the Mac uses Apple’s PDF rendering engine. If Preview loads it quick then DT should load it quickly. That said, because they share a rendering engine, they also share problems. This has been discussed on the DT Forums:
And a workaround in the Hidden Prefernces section of DEVONthink:

DEVONthink has dealt with a lot of PDF-related issues for years. I am sympathetic to this struggle. In theory they could replace Apple’s PDF renderer with their own, but it is not easy to build a PDF rendering engine from scratch. They could, in theory, license an engine from another company (like PSPDF and the PSPDFKit), but there may be other reasons they don’t do that. But it is for this reason that I rarely view PDFs within DEVONthink. I just cmd-shift-o to open it in the default application (in my case PDF Expert) and only view a PDF in DT when I want to copy a link to a specific page in a PDF Document (You do know about the amazing “Copy page link” right!?).

1 Like

That’s one of the huge time-servers with Papers 3. If I have a paper, I can pull the data from publication databases, saving me a huge amount of time. So, while searching for similar papers in DevonThink (“Zettelkasten”) is easier, I still will need Papers 3.


And, of course, DEVONthink doesn’t insert citations or generate bibliographic references!

An information manager and a reference manager, as I see it are two different tools that can (and should) be used together. A reference manager will never replace an information manager like DEVONthink, nor will an information manager be able to replace a reference manager (although in theory it actually probably could if someone were to build that feature, though personally I wouldn’t want it, I switched reference managers about 6 or 7 times in the last 6 years and I like having it not be tied to my information manager, which I’ve switched once in a decade).

1 Like

They had particular issues with one of the macOS versions, was it Mavericks(?) - but that has been mostly sorted now. I hardly ever have rendering issues now.

@dfay - what OS are you on?

I need to think through this…

  • I scan a paper from a journal
  • I then have a PDF with a random filename
  • I pull it into Papers, use the match funtion and get all the fields (author, name, year, journal,…) filled
  • Now I can cite it
  • Papers 3 search features suck…so I’d also like to have it in DevonThink…
  • And annotation on the iPad is done where?

Even if there’s no DOI in the PDF?

Sierra. But the same PDFs render fine in other apps.

Just copy/paste the title into the search field and Papers gets everything else (Author, Source, DOI) from the databases (PubMed,…).

ok that makes more sense…same degree of manual intervention as with my existing Alfred/BibDesk workflow

Is this a function of the quality of the OCR? If so, every tool is going to have the same issues I’d expect.

Nothing to do with the quality of OCR. DevonThink has more extensive search features than Spotlight or Papers.

If getting started with DT. Pro Office, Pro or Personal?

@Yaakov Here is a useful comparison page.
Note the tabs at the top, for all the feature sets. If memory serves, they have a 150 hour trial as well.

1 Like

Honest question, but you seriously scan journal articles in 2018? Isn’t that immensely labour-intensive compared to just pulling the article off of the journal site? (Unless, of course, the article predates that and hasn’t yet been scanned, but nowadays that seems like it’d be fairly rare, so much of the back-catalogue of most journals has been scanned!)

That’s what undergrad student employees are for.:stuck_out_tongue_winking_eye:

I had the same Q actually. It’s pretty rare that I find a journal article that’s not available electronically, though there are definitely some where my library has a long paper run & thus has decided not to subscribe to the electric archive.

1 Like

Would be Pro, I guess.

Not a student and we have several subscriptions to paper editions at my company. But also the ones I get online (conference proceedings) need to be properly entered in a “citeable” manner.

Ah, I had assumed faculty or staff actually, but I see, private sector is a bit of a different ballgame so that makes sense!