PDF reader as part of Zotero

JohnAndre

First, I am extremely new to Zotero (2 days). I decided to switch from Mendeley when I learned they were encrypting their database to make it harder to get my data out. So, if this has been discussed before, please forgive me. I searched but did not find anything.

My question is, has there been any thought of building a PDF editor into Zotero?

I know there is something to be said for focusing your core skills and depending on others to handle the extra stuff, but as someone who has used Mendeley for a few years and Zotero for a few days, I only find one great frustration when using Zotero: Annotating PDFs.

Mendeley is quite polished in this area. I mean, their UI is really easy to use. With Zotero, first I have to find a decent PDF reader that supports annotations. Fine, there is evince, okular, and others. First, I tried Okular but when I would save the annotated PDF, it would start on my desktop and I had to try to find the original file which seems to be in a folder which is calculated by a hashcode (or something similar). Then, when I do it with my next PDF, it starts in the the folder of the previous document and I must navigate to another folder. It's far from easy.

Evince is better at this, but there is still the issue that I need to save, overwrite, and hope I don't forget and do something which causes problems. So, I need to close and re-open to ensure that everything did work as planned. More time which feels wasted.

Short version: It would be fantastic if Zotero had its own built in PDF reader which could simply overwrite the PDF being annotated (or writen side-by-side for those who want to keep a 'clean' copy of the PDF). Also, it would allow some flexibility like making annotations searchable (which Mendeley could not do, though I wish it would have).

Is there any chance that this would/could happen? I don't have a lot of free time but I would contribute some of my free time to help build such a tool.

dstillman

This is a design decision, and it's not very likely to change.

First, I tried Okular but when I would save the annotated PDF, it would start on my desktop and I had to try to find the original file which seems to be in a folder which is calculated by a hashcode (or something similar).

If saving triggers some sort of filepicker dialog, that seems like a flaw in Okular — it should just save back to the original file by default (i.e., Save rather than Save As), which is how it works with, say, Preview on macOS. Double-click in Zotero to open the PDF in Preview, annotate, Cmd-S, Cmd-W.

dstillman

As for making annotations searchable, the third-party ZotFile plugin can extract your annotations to notes in Zotero, where they become searchable. We're planning to integrate similar functionality into Zotero in a future version.

It's also possible that reindexing a PDF in Zotero makes the annotations searchable, though I haven't tested that. It might be reasonable for Zotero to automatically reindex PDFs after they change so that an Everything search can match on annotations.

dstillman

Also, just to clarify, it's not like Mendeley built their own PDF reader — judging by their release notes, they just license a third-party framework. (For what it's worth, they also store annotations and highlights in the encrypted database and only let you export them one file at a time, though that's not inherent in a built-in reader.)

We sort of take the position that you use your OS of choice because you're happy with the tools available for something as general as annotating a PDF and it doesn't make sense for us to offer some third-party cross-platform option to try to compete with that. (We would also presumably want to use an open-source framework, and I don't think pdf.js is going to be comparable to macOS Preview anytime soon…)

JohnAndre

Thanks for the detailed answer.

About the Mendeley thing, my point was not to praise them for their hard work. I just think their end-user experience is quite good in that respect. Yes, their attempt at lock in is horrible (which is why I switched to Zotero).

I do think that having a more integrated experience would help bring over more users (and make their transition smoother). Of course, I get that you have considered that option and rejected it for now.

I will re-check Okular. I might be confusing the experience with qpdfview.

Anyway, keep up the good work. It is always inspiring to see others working so hard to help the world be a better place.

adamsmith

yeah, I think mainly what you're seeing is more about the sad state of PDF readers on linux than about the advantage of having a built-in one.

The freely available PDF readers on other OS, be it Mac Preview, Acrobat, pdfXchange, or FoxIt, offer by most accounts a better user interfaces and more flexible annotation options than Mendeley's built in one. I find annotation on Evince and Okular so bad that I've given up on annotating PDFs on linux.

fcheslack

Older versions of Okular by default saved annotations in a separate directory and required using a separate save as option to save to the PDF file. It seems that newer versions have changed this so that the normal save action will save any annotations that it can to the PDF file itself (and warn if there are modifications that can't be saved to the PDF).

emilianoeheyns

Foxit is available for Linux, although by the looks (and performance) of it they use something like Wine to do it. But it works.

emilianoeheyns

Also technically Acrobat for Linux is still available (ftp://ftp.adobe.com/pub/adobe/reader/unix/9.x/9.5.5/enu/) but it's not been updated in 5 years.

ZenonMarko

@dstillman "As for making annotations searchable, the third-party ZotFile plugin can extract your annotations to notes in Zotero, where they become searchable. We're planning to integrate similar functionality into Zotero in a future version."

True about ZotFile, and this would be a nice addition to Zotero natively. However, this still would presumably mean that one would have to reimport the annotations any time one edited or added annotations in the PDF. Perhaps some kind of auto-sync or auto-import? Not sure how technically feasible that would be.

rcastil

Visual cues.
I really like to Zotero, but I surely miss a built in PDF. What software designers miss, is not that the users are not happy with their external pdf reader, but people who rely a lot on visual memory, it helps a lot of time to identify the right paper just by seeing the thumbnail together with the reference, instead of having to open every single pdf.

dstillman

@rcastil: PDF thumbnails are a separate issue, better discussed in other threads (of which there are multiple).

flyredeagle

I think this thread mixes 3 things pdf viewing, reading annotations and editing them:

1 Portable pdf viewer:
I was quite impressed about pdf.js https://mozilla.github.io/pdf.js/ , should fit in the scope of electron here and can you agree that also license is ok ?

2 reading annotations:
They state that they want to read annotations they don't want to write them
https://github.com/mozilla/pdf.js/wiki/Frequently-Asked-Questions#faq-annotations
Is this good enough in this scope ?

3 writing annotations :
This may be a non standard thing, i.e. supporting all possible viewers and PDF formats etc., and may not be easy either, (in place, side by side, corruptions of original files, searching / exporting annotations). Maybe the solution for this can be a fork of the code of pdf.js or maybe is just a wish for now and left out.

I think with pdf.js you can make 80% of the people happy and then we think about the rest later .. also I suspect that the pdf.js can be implemented as a plugin, i.e. optional, correct ?

PS: all design decisions are born to be changed ;)

adamsmith

The main advantage of a built-in PDF editor would be the ability to natively annotate in Zotero. That's literally the first point raised by the OP and it's what invariably (and reasonably) comes up when people ask for PDF support. Without that, adding a PDF reader is going to add more frustration than it solves.