Sente and DevonThink Pro?

I’m using Sente for bibliography and have over 5500 pdfs in it. I also have DevonThink Pro and really like its indexing capability. Is anyone using these two applications together? Any advice?

I don’t use Sente, but if the files are stored in a directory somewhere, you could just have DEVONthink index that directory. This leaves Sente as managing the PDFs and you can have DT update its index via Synchronize when it changes. This is the simplest from the Sente side since it requires no changes in Sente. However, DT is just pointing to the PDFs.

Otherwise, you can select the directory and import them into DEVONthink. Now, DT is managing the PDFs and they are stored inside the DT database. Now that DT supports links into the database, you might be able to use that in Sente.

Probably the best option given the number of PDFs is the former.

Does DevonThink Pro change the files or the folder in any way that might mess up Sente’s activities?

DTP and Sente work very well together; I have been using them in tandem for nearly two years with no interoperability issues. My workflow is to use Sente’s search and import to default archive folder and have DTPO “index” that folder (which does mean that unless you have an applescript, you will need to “synchronize” from the DTP/O database). If you are moving between machines and your files are hosted locally rather than on a server, you will need to duplicate the archive on each machine or the indexed folders will have broken pathways. Hope that helps.

Yes, DT 2 changes the name of a file - even a synched file - when you change the Title of the DT document. Which of course is killing the link in Sente.

Text recognition doesn’t work with imported files. Well, you can convert a pdf into a searcheable pdf, but DT would create a new file. To avoid duplicates you would have to drag the new file into the respective Sente reference, delete the old Sente attachement, then resynch in DT. It’s a mess.

To your broader question how to harmonically use DT and Sente with your 5000 documents managed by Sente: Let Sente continue to manage these docs. You will presumably use scientific databases from within Sente, use its reference detection and pdf import feature.

If you rename a doc in Sente, the respective DT record will be deleted on the next synch and a new DT record for the renamed Sente attachment will be created. Not sure what happens to replicas in DT, but I assume that they are deleted as well and the new DT record will not be replicated to the folders the old document was replicated to. Any info you’ll add to the document in DT will be lost on synching when the file was renamed before.

You will lose all the meta-data of Sente’s reference records. If you have filed your Sente references into different collection or databases - I guess you have with 5000 attachments -, Devonthink’s synch feature doesn’t know about this meta-data, and you won’t either when you’ll later look into the synched group in DT. The attachements in DT won’t be grouped according to your collections in DT.

As your synched attachements either are in a single group in DT or are sub-grouped according to author’s names (depends on your Sente Link preferences), Devonthink’s crown jewel, AI, won’t shine, as your documents aren’t thematically grouped in DT. AI will be close to useless unless you invest a couple of days to group your files in Devonthink.

Theoretically, there is a way to properly group all your 5000 attachments according to your grouping into collection in Sente. Sente uses an sqlite database (.sente501 files) for data storage, the databases are non encrypted, not password protected (very nice indeed, thirdstreetsoftware). So, you could retrieve a list with all Sente references with attachment plus their respective collections. This could be done either with applescript or with the Automator. Not sure how to proceed then. The devonians might have an idea, here.

A decent roundtrip synch solution for DT and reference management software like Sente, Bookends, Endnote or Bibdesk would be very helpful for using DT in science or R&D. With Endnote and bibdesk being scriptable with Applescript, this might be doable. Sente and Bookends are a different story, though.

Apart from that, Sente and DT are both very nice.

Note that the AI features, See Also and See Selected text will work perfectly with content such as Indexed-captured material from Sente or any other such application database, as they are not dependent on the group organization of the material.

Re renaming Index-captured files in DEVONthink: that’s usually not a problem for material sitting in Finder folders, but would be a no-no if another application is managing the files by name or Path, in its own database. In such cases, resist the temptation.

Based on Andreas Schmidt’s comprehensive reply, it seems to me that it is too dangerous to attempt integration of DT2Pro with Sente at this time. (since I have 5500 pdfs with bibliographic metadata)
I will try a little pilot test using a small Sente test Library, and report back.

But it’s a shame that we don’t have one truly integrated solution that accommodates all the requirements of a scholarly application, to include:

  • resource search and download
  • bibliographic data download and management
  • pdf and other content
  • full text indexing
  • tagging and categorization of records
  • easy note taking with good export

What have I forgotten?

Yes! Waiting for DEVONcite, front of the queue :smiley:

That’s correct, yes. However, you have to group the files from scratch in DT even though it’s been done in the reference manager before. And DT’s Autogroup/Group-Into feature won’t help you, at least in the beginning.

The implication of this is that you advice Devonthink users to not index pdfs downloaded with any reference management software on the market. Sente, Bookends, Bibdesk (not sure about Endnote) - they all rename files attached to references. This seriously blows DT’s usefulness in academics.

No, I gave no such advice. In fact, a number of users report successful sharing of information between a citation or other database and a DT Pro/Office database.

But I would think that the easiest approach to managing the Index-captured content of an external database would be to simply leave that DT database group in the state in which it was captured.

I would set that group as exempt from Classify. It’s not necessary to change the organization of that group. It’s not necessary to rename the items.

Remember that See Also, See Suggested Text, searches and smart groups are fully functional regardless of the organizational structure of the database, or the Names of the documents – the text content of the documents holds their most important information. By incorporating the Index-captured files from the external citation database into a DT Pro/Office database, their information has been melded into the other information resources in the DT database. Nothing has been lost. Much has been gained.

I know you didn’t and I would have ignored such an advice, anyway. :slight_smile:
It is possible to “successfully” share information, but it’s not possible to do it elegantly, seamlessly, effortlessly. You lose information, have to avoid caveats, go great lengths, you have to do things twice. This integration is just not as supreme as DT is in general.

By the way: Apple has nicely integrated Pages with Endnote…


Right. But sometimes you would like to rename the DT document (“no author, no year, wrong title from google scholar or so.pdf”). DT 1.x didn’t rename the file after renaming the document, DT 2 does. DT 1.x was acting better in this particular use case.

I guess, the academic workflow starts with searching literature. You would be foolish not to try to do that with your reference management software, nowadays. Why? I think I’ve written earlier in this forum about the difficulties of getting a pdf or webarchive into a new reference in Sente or Bookends.
I would assume that everybody using a ref management software categorises the references in that software. This information is, you’re right, not lost in your reference management software after importing it into DT. However, it’s just not imported into DT. So, when working in DT, you gain DT’s fabulous abilites, but you lose all that profane referential metadata of a pdf, which could be so valuable if it would make it into DT.


I just posted this on the Sente forum on the same subject, so I thought I could re-use a few electrons :wink:

I’m using Sente and DT Pro together. I’m actually trying to figure out the best way to do it, and I still find a few things that need to be improved.

For some time I kept my pdfs in DT and linked them to references in Sente. You can drag and drop a file in DT on a reference in Sente and it gets linked. I’ve changed the way I use them by letting Sente file the attachments, especially because it can change the names automatically (you can do it in DT, but you’ve got to do it by hand). I have DT index (not import) the Sente attachments folder, although you have to tell DT to synchronize it every time you add a file. I think there’s no way to automate syncing.

Once you have the attachments indexed you’ll have them in DT, so you can replicate them to a specific group (when I start working on something I usually replicate the docs I want to have at hand). And of course, DT will search and find in their content.

A few things that might help to work better:

  1. I like the note-taking system in Sente, which would be even better if it would mark in the pdf the fragments that have been extracted to notes. But it would be great to be able to import/index those notes into DT easily. If DT would be able to manage sheets so they could be seen as forms (it’s not possible right now, but could come back in the future according to the DT people) it would be great to be able to import the whole group of notes for a reference into DT that way.

  2. Metadata. If sente would be able to tag its references and attached docs with keywords readable by DT (maybe through OpenMeta) it would ease greatly the workflow (attachments tagged with a specific keyword would automatically appear in a Smart Group in DT once indexed).

  3. Inserting references. Some kind of easy way of switching back and forth between DT and Sente (the way Cmd-Y works between Sente and Mellel) would be great. I don’t think being able to scan and replace citation markers is essential, since I usually export the rtf docs and work with them in a word processor, so the final compilation of the bibliography can be done outside DT.

I think that making Sente (or Bookends) work nicely with DT is the way to develop an academic working environment, rather than hoping for somebody to build the perfect app.



Hi, Hendrix:

Thanks for your post! I have a question about your process. You wrote:

But what happens if when Sente renames the attachment based on your system (e.g., Author, Date)? Doesn’t that break DT’s index?

Nothing changes in DT until you synch the folder with the changed file. DT will then kick out the old documents (all replicas), and add the renamed file to you DT database.
I guess, Sente doesn’t change file names automatically (any longer?), only if you click “Refile attachements” in the action menu in the bottom link pane. So you would want to avoid that command and, above all, the “Apply to Existing Linked Files…” command in the Link section of Sente’s Preferences. And never rename the title of a synch’ed document in DT that is controlled by Sente.
And cross your fingers that neither devontech nor thirdstreetsoftware come up with new features that will screw up this somewhat fragile, imperfect coexistence. (Like DT2 did with its forced renaming of files after changing a document’s title.)

I always have Sente rename the attachments (usually pdfs) the very moment I file them. Once its done there’s no need to change their names again. So they’ll keep their names when I work with them in DT (where I don’t change their names either).


Unless someone is able to design the perfect all-in-one app for this kind of work (which would probably look a lot like Microsoft’s bloated products, anyway :slight_smile: ) the way to go is collaboration between apps that do best what they’re supposed to do. For this to be possible the very basic request is “don’t step on the toes of the others”, don’t introduce features that break the connections.

Of course one can dream of developers looking at the others and opening up connections (via AppleScript, Automator flows, plugin architectures… what else). I’m sure they’ll come true. 8)

I have to agree with Hendrix’ comments: I’ve been using Sente and DPTO2 together for a while and have few complaints. I’m very productive with the combination.

If I had any requests, it would be to Sente to open up a little more of their application with Applescript.

Best, Charles

Yikes! Why does it do that? That kills the workflow I was aiming for :frowning:

Use the synchronize script. Attach it to your indexed attachments folder, and whenever you open it, DT will resync automatically. This script used to be on the site, but I couldn’t find it. I’ve attached it.

I remember bringing up this whole conversation a year or two ago (see here) Since then my thoughts have changed somewhat— for 1 do like Sente more than Bookends now, and DT2 has brought Mellel collaboration a lot closer. For what it is worth, here is what I think would be cool as far as DT and Sente integration:

—Have Sente treat local links just like an actual attachment. This may mean Sente has a dedicated field for local links or just using the URL field. Either way, the local link would be opened and viewable right in the Sente window just like a Sente pdf attachment. This is something ThirdStreet developers would have to do, and would be useful beyond DT.
—Sente would have to also have the ability to launch the local URL if the user wanted to view it in the “native” application (in this case DT)
— The above would also require that DT add something to the edit menu, “copy link as plain text” (I imagine this is a rather easy thing to include)
—Once DT2 is complete it will have the Preview annotation tools. I’d like to see Sente with these tools as well. Annotations would be seen in either Sente or DT on a PDF
— A new DT script which would take available data of a PDF (presumably its filename) and find the resulting citation in Sente, copy its local link from Sente, return to DT and paste the link in the PDF’s URL field in DT. This would allow us to go swiftly back and forth between a PDF in DT and its citation in Sente. The script could also, perhaps, be tailored to bring over other info. For example, I’d love the abstract of the article to also be pasted in the notes section of the PDF in DT.
— to facilitate better cross-linking between items in DT and Sente, I’d like to see Sente improve its linking abilities, specifically I’d like the Sente URL to be customizable.

That would take of a lot of it, though 2 of them are pretty big items. This would bring some great interoperability between DT and Sente.