Zotero is to Bookends, as Evernote is for DEvonthink

[Dellu]

This is just to note that people should not consider Zotero and Bookends as competitors. They work together best. That is how I use them.

The relationship between Zotero and Bookends is just like the relationship between Evernote and Devonthink.

Zotero and Evernote are specialists on the WEB. They are best for collecting data. BE cannot compete with Zotero's translators. They are so many: it is unthinkable for a private developer to cover them all. The translators are developed by a large community of developers. The same is true with Evernote. The web clipping tool in Evernote is the best. A huge resource has been invested on it. Devonthink team admitted that they cannot make the clipper as good as the Evernote because they don't have the resources that the Evernote team has.


But, we know that BE and DT are the real beasts to manipulate, process, store, retrieve your data (references). They are the ultimate powerhouses of the power user to control your material. But, getting data from the web is not their best strength. That is why we need to supplement them with the Web specialists such as Zotero and Evernote.

I have tried hard to completely eliminate Evernote from my life forever, for a number of reasons. But, I keep on falling back to it, because the web clipper in DT is no where close to it.

So, what I do right now use Evernote as a temporary inbox to clip web pages precisely the way they appear on the web. As I get time to process these clipes, I move them to DT. I am using the same strategy with Zotero. Many of the web pages covered under the Zotero's translators are not covered by BE. I use Zotero as a temporary inbox, to scrap references from the web. I finally move these references to Be for permanent storage and manipulation, if I need to do so.

This is my strategy. How about you guys? Do you completely rely on a single tool, or combine the powers of multiple tools?

[iandol]

I suppose it depends on the field of study. For my field (Neuroscience), almost every relevant web page will have a corresponding DOI, PMID or preprint address (and PDFs are how we read the actual papers). Zotero provides no added value here I think, Bookends quick add (automated by applescript) does everything I need. But I assume for more heterogeneous sources, Zotero, which started out as a web plugin, comes into its own...

One other really important use for Zotero however is online reference sharing. For collaborations Bookends server is not as flexible as Zotero's online interface. I find manually syncing from Bookends to Zotero really clunky, and would really welcome a sync service between Bookends and Zotero. This may be coming for iOS Bookends (viewtopic.php?f=7&t=5487), and if we have iCloud to sync macOS and iOS and iOS to sync to Zotero this would be wonderful. I think having a sync would improve this concept of Zotero and Bookends working together, each providing their advantages to a unified workflow!

[DrJJWMac]

I specialize in the physical sciences (physical chemistry) and engineering (chemical engineering). I find Zotero currently to be only of nominal added value to improve how I process references for my research. I do not clip Web pages for the literature review portions of my research. I assemble and read PDF journal articles. Most recently, I read them on my iPad. In this workflow, Bookends outshines the competitors. As for working on the desktop (macOS), I have more friction working with Zotero for PDFs than any other of the numerous citation management apps (Papers or Mendeley are the other two that I use). Perhaps this is because Zotero "hides" the PDF that I want to read as content inside a top level container-link. I have to click into a link to open it to find the PDF that I then have to click on to open to read ... too many steps. Having garnered documents across a few decades and a few dozen different fields in my career to date, I appreciate the ability of Bookends macOS to allow me to split libraries into smaller chunks. I know that Zotero allows a user to split off a portion of a large library to share to the iPad, but the process in Zotero seemed so much more difficult to followed compared to what I do using Bookends macOS<->iPadOS syncing. Finally, I generate documents directly in LaTeX. I appreciate the ability to build a smaller library in Bookends that is specific to the document that I am creating, and then export that library as a BibTeX file to use directly in the compilation of my document. I finalize the integrity of the \cite{} keys nearly as the last step in the compile process to create the final document.

I also know that Zotero allows you to create smaller libraries in shared folders. This option entices me the most about Zotero versus Bookends, primarily because I collaborate with folks who use Windows.

As a note referencing the title analogy, I've not opened Evernote in what now seems like forever (although I have a lot of stuff there from 10-15 years ago). I also have only started appreciating DevonThink primarily as a tool to help me categorize through an extensive set of content from courses taught over a few decades with the hope to compile a textbook.

[Philologist]

I use Zotero as a temporary inbox, to scrap references from the web. I finally move these references to Be for permanent storage and manipulation, if I need to do so.

Zotero for Mac is actually restricted to MS Word, LibreOffice and GoogleDocs, so I guess you must be using MS Word.

The problem is, if you save in Rich Text format and have an author whose name contains an umlaut or a diacritic, Zotero will fail to execute the RTF Scan command from the Tools menu, thus making Zotero impossible to use with other programs, such as Nisus Writer Pro, Mellel, Scrivener and TextEdit.

The problem is well known among Zotero developers, but they are not interested in fixing it or are unable to do it. The answer I got from them in 2020 was:
"Yeah, it looks like this was just never implemented — I think you're the first to report this in the 10+ years that RTF Scan has existed. (RTF Scan doesn't get a lot of use.) I've created an issue to address this, though I don't know when we'll get to it."

Next year, i.e. 2021, I spent a week or more with the developers trying hard to find a way to scan RTF documents with diacritics in the above mentioned applications, but to no avail. In the end they were suggesting I should write my paper in Nisus Writer Pro, then use LibreOffice to scan it, which I find is absurd.

According to my information, no Nisus user is using Zotero. A question in the Nisus forum, viewed 6730 times, received 0 response.
https://nisus.com/forum/viewtopic.php?f ... 37&p=37671

Beside these restrictions, Zotero also requires that you install Java Development Kit (JDK) for the ODF Scan.

What export format do you use when you export references from Zotero? Just now I tried Endnote XML, RIS and CSV, but when I tried to import into Bookends nothing was imported.

[Philologist]

So, what I do right now [is to] use Evernote as a temporary inbox to clip web pages precisely the way they appear on the web.

What browser are you using?

If you clip/save sites as Web Archive with Safari they should appear in DEVONthink exactly as they appear on the web.

Evernote requires you to be permanently connected online. Have a power failure, …and suddenly you can't access your notes anymore. I find this a big drawback. :–(

[Philologist]

I assemble and read PDF journal articles. Most recently, I read them on my iPad.

I don't have an iPad, so what's the advantage of reading them on an iPad rather on a Mac with PDF Expert, PDFpenPro or Skim?
What apps do you use to take notes, highlight text, etc. when you read a PDF on an iPad?

Papers or Mendeley are the other two that I use

Papers has the same RTF-scan limitation for me as Zotero, and Mendeley is also a reference manager for only MS Word, OpenOffice and LibreOffice. If you work in the humanities and *writing* good text is an important issue (versus just presenting hard data in science) you will inevitably want to use other word processors, such as Nisus Writer Pro, Mellel or Scrivener, and then you're completely out of luck with Mendeley: there seems to be no easy way to enter citations and automatically create bibliographies in anything else than in Microsoft Word and LibreOffice. You can do it manually, by copy and paste, but who would want to do that?

[Philologist]

Zotero [is] best for collecting data.

It seems you use Zotero for scouring the Web for references. If I don't completely misunderstand, this can only mean that you are looking for brand new publications not older than, let's say, one week. Such publications are not (yet) to be found in any library. And this also means that your academic field is not in the humanities—in the strict sense of that word—because in the humanities the date of publication is usually not the most important criterion for whether a book/article is good or in any way relevant for your own publication.

My fields of learning are literary scholarship, history and philology (not formal linguistics, unless topics are treated diachronically; my soulmates are Jacob and Wilhelm Grimm)

Most of my references are *books* (not journal articles, because I write for teachers and students, not for expert researchers in the field.) Therefore my hunting ground are books (mainly) and these books can all be accessed in a good online university library.

I posted this about my favorite online source in 2017:
"For the humanities I have found Germany - GBV (GVK).fltr to be the best online source. It's a Library Network of the German States Bremen, Hamburg, Mecklenburg-Vorpommern, Niedersachsen, Sachsen-Anhalt, Schleswig-Holstein, Thüringen and the Foundation of Prussian Cultural Heritage. GBV stands for "Gemeinsamer Bibliotheksverbund Verbundzentrale. Here I've found many English books that I couldn't even find in the Library of Congress. […] The import filter Germany - GBV (GVK).fltr *combines* university AND public libraries."
viewtopic.php?f=2&t=4300&p=19444

You can find more about this library here:
https://en.wikipedia.org/wiki/Berlin_State_Library

[Dellu]

Yes, you guys have a point.

If the reference has been published, and has a DOI, BE is sufficient

But, in my field (Linguistics), it has been a movement in the last couple of years to either publish in open journals: or share your pre-publication material in a shared repository. These repositories are now the primary means of communication for new ideas, pre-published, or even published articles, dissertations, and books.

Look at Lingbuzz: (https://lingbuzz.net) for example.
There are also many open journals that don't assign DOI to the articles.

When these kinds of sources appear, what I do is simply request a Translator. It doesn't take more than a couple of days to get a Translator. I am then good to go.

[Dellu]

What export format do you use when you export references from Zotero? Just now I tried Endnote XML, RIS and CSV, but when I tried to import into Bookends nothing was imported.

I havn't tried the scanning part of Zotero. As I have noted, I often export my references to BE. But, the BetterBibtex is also great add-on for bibtext users.

But, for the export, the Endnote XML does a great job for me. BE is able to import the complete reference as well as the attachments. You probably need to try the latest standalone version. You need to consolidate the attachments after you imported the references. They will remain in Zotero's database otherwise.

[iandol]

Can you explain what a translator is; is it a custom HTML scraper to get information from a page?

For very new articles sometimes the DOI does not resolve in Bookends, this does drive me crazy as I tend to jump on new papers, then can't quickly add them and they may get lost unless I keep them as a tab in my browser or add them to a note for later import. As Bookends is scriptable, it wouldn't be too difficult to build a custom importer, using Python or Ruby or something to scrape & clean up a web page and extract the data to add to a new reference, but it sounds like this is more streamlined via Zotero...

[Dellu]

Yes, translators are metadata extracting scripts. I think they are written in Javascript.
I believe they are two types:
The first type uses the metadata already embedded in the webpage (if there is any). That is the easy one. But, the most important ones are the second type which extract references and their associated attachments from the html pattern.

You can read the documentation here: https://www.zotero.org/support/translators

You can also search the Translator repository in Github (https://github.com/zotero/translators) if the webpage you are getting your references from is covered

This is incredibly efficient system. For unpublished articles uploaded to Lingbuzz, I used to download the pdf, then autocomplete in BE. The result was hit and miss as GoogleScholar is an unreliable source. The Translator gets all the metadata. It is so magical.

[Philologist]

Look at Lingbuzz: (‪https://lingbuzz.net) for example. […] When these kinds of sources appear, what I do is simply request a Translator. It doesn't take more than a couple of days to get a Translator. I am then good to go.

I can understand that Lingbuzz is a valuable site for you as a linguist.
Here is an article from this site, chosen randomly:
https://ling.auf.net/lingbuzz/000927

How would you directly import it into Zotero? Do you need to ask the Zotero developers to provide you with a translator? And if so, would that translator be valid for all articles from Lingbuzz? Or only for the above mentioned article? I ask because I'm not familiar with 'translators'.

[Philologist]

for the export, the Endnote XML does a great job for me

I changed the import settings in Zotero and now it works. :–)

[iandol]

Yes, translators are metadata extracting scripts. I think they are written in Javascript.

Ok thanks. The developer info is here: https://www.zotero.org/support/dev/translators — as this all is open-source, it would be technically possible to run these javascript translators using a third-party tool triggered from Bookends and import the data. That requires someone with minimal javascript skills and motivation enough to build it. The Zotero project has made some major open-source contributions to bibliography tools, including CSL (used by iOS bookends, Pandoc and many other projects) and this translator framework, so for that even if those of us who don't use Zotero much should still be grateful for its existence!

[Dellu]

How would you directly import it into Zotero? Do you need to ask the Zotero developers to provide you with a translator? And if yes, would that translator be valid for all articles from Lingbuzz? Or only for the above mentioned article? I ask because I'm not familiar with 'translators'.

You need to install the Zotero add-on to your browser. Lingbuzz is already supported. The add-on uses the translator to detect the articles in Lingbuzz webpage. You need to open the individual articles for the addon to detect the references data. Batch download is available in many webpages, but not lingbuzz.