Comment by japanuspus | Hacker News Reader

Comment by japanuspus | Hacker News Reader

Parent Full thread

If the town is "online reference managers," you are probably right, but I would argue that reference managers are one of those areas where you really want something offline:

- Offline ensures that you do not suffer an externally caused downtime just before a deadline

- Offline ensures that you have a path for keeping your database throughout your research career, and to do system updates when _you_ want to.

- Offline ensures that if you leave academia, you will always have access to local copies of the academic papers you have referenced.

My favorite offline/local reference manager is `jabRef` [0] which stores all metadata directly in a bibtex-file. The GUI has an excellent pdf-integration, and everything is local and super fast.

Case in point: after a decade in industry, I am looking to get back into my academic fief. All the papers I ever read are in my Dropbox, and all I had to do to pick up where I left was download a current version jabRef and point it to my database which it read without any issues.

[0]: https://www.jabref.org/

Just to be clear, Zotero is primarily a local program that saves all data and files to your computer. You don't need an account to use it. There's optional syncing and a web library so you can use it on multiple devices and collaborate with others, and there's a browser extension (which JabRef uses as the basis for their extension) that integrates it deeply with the browser, but in no way is it an "online reference manager". We strongly agree with all the reasons you give, which is exactly why we designed Zotero the way we did.

(Disclosure: Zotero dev)

I use Zotero and sync the attachments to my Nextcloud instance via WebDAV. The database is synced to the Zotero servers, because that's the only option. With a bit of care on initial setup for new machines, it works wonderfully. The real value are the PDFs, some of which are of course really hard to get by. And those are "local" in the sense of "local server", yet available to all clients.

There was a forum entry that self-hosting the database is a topic being worked on. That's my number one gripe with Zotero, it just seems weird: if using multiple clients, users have to sync to the Zotero servers, yet the storage space for attachments there is minimal (which is okay; Zotero pricing looks reasonable but of course, own storage via e.g. Nextclodu is much, much cheaper). This basically forces a hybrid solution, with bring-your-own-storage. If all of Zotero's data could be self-hosted, now that would be a streamlined, great experience.

Then again, I am anal about my entries' metadata and often correct faulty data found online. Syncing that to Zotero servers is kind of like giving back? I always thought that data might help improve Zotero's own auto-fill functionalities, which I'd like helping with.

See my comments on [1] for why we don't support self-hosting. Short version: it would pretty much just be an entirely different project.

> Syncing that to Zotero servers is kind of like giving back? I always thought that data might help improve Zotero's own auto-fill functionalities, which I'd like helping with.

No, we don't expose your data to anyone else. (And tools that try to do that usually just end up filling people's libraries with junk data.)

You should sync Zotero if you want to 1) access your library on other devices, 2) collaborate with others, and/or 3) have a real-time backup of your Zotero data.

[1] https://github.com/zotero/dataserver/issues/105

I've been syncing my full-texts via syncthing for years and years now and I've never had to think about it after initial setup.

Zotero is great! :)

Thank you for clearing this up - and apologies for not researching!

Thanks for posting this - JabRef looks perfect for helping me organise and track papers to read, I had no idea it existed!

Reply on news.ycombinator.com