Hao's Zotero Workflow best practice

My obsolete tutorial: [Tutorial] Zotero for PhD or general researchers

Initial processing:

1. Add new papers to Zotero and tag _st/New_Unread (immediately);
2. Skim all added papers:
   1. Write a summary note in PDF
   2. Extract annotations to Zotero using Zotfile (the above summary will be included)
   3. Add tag _st/Skimmed
   4. Add tag :todo:further_study if worth further reading the details.
3. Tag :todo:skim_first to prioritize remaining papers if cannot finish skimming.

Later when have time:

1. Skim :todo:skim_first
   1. Add summary and tags as above.
2. Read :todo:further_study in detail and write a detailed summary note;
   1. Tag _st/** accordingly. (see my_tag_system)
   2. Tag _st/Printed if printed
      1. Replicate all the annotations on paper to the original PDF files.
3. Skim _st/New_Unread
   1. Process as for initial processing

Use Okular for PDF reading and annotating, because:

• cross-platform; use standrad PDF annotations (saved to file instead of externally); customize shortcuts; support opening multiple files and silent loading external changes (be careful).
• Known issues: only under Linux no Chinese input in some cases, a fcitx and qt issue. Workaround: use English, write Chinese notes somewhere else and copy-paste.

Set up shortcuts as

• j k h l, default for down, up, left, right. (no need to set manually)
• d u, page down, page up, set page overlapping to 40% for half page scrolling.
• G, G Shift + G, go to beginning, go to end. (same to Vimium)
• v, text selection, same to Adobe
• b, browser, only in this mode can I click links and view notes.
• 1 2 3 green, blue, red, with opacities all 30%. (need to set annoation first)
• 4, pop-up notes.

Keep simple and only use these four types of annotations:

Tags are portable, but collections are not.

Therefore, tags are less likely to lose or crupt.

For example, if you use collections to classify items to different topics, you may need to move items here and there often. It is very possible you could accidentally mis-drag multiple items into somewhere you didn't even notice. Then it's a disaster to recovery everything. This happened to me many times before and I didn't realize it was a defect of my organizing system. Differently, the tags go anywhere and you won't need to drag them,

In other words, “collection system” is more suitable to be a more static hierarchical structure, avoiding frequent changes of the categories.

Upgraded/renamed tags to be compatiable to Obsidian tags, using only letters, nubmers, _, and - (/ for nested tags as well). So that I can easily sync the tags in Zotero and Obsidian.

Don't set color to these tags. If I want to see good references in the current collection, I could simply select the corresponding tags. I don't need to let these tags visually pop up when I'm scanning all the items.

Usually, I just need to use 2, 3, 4 to rate a paper. 0 and 1 means that paper does not worth more effort. 5 indicates a classical paper that usually is a fundamental or groundbreaking work.

Updates 202109221513: Rating tags were changed from :rt:?/5 to _rate_? to be compatible with Obsidian's tag system. Other tags are kept untouched because they are less usefull in Obsidian.
Updates 202204071503: Again, for obsidian, changed from _rate_? to _r/? to make use of the nested tags.

Assign all todo-tags with a same color, red for example here, to keep track of them visually in the item list.

Not updated to be compatible with Obsidian, i.e., : is still used in Zotero tags. My reason is that, these todo-tags should not propagate into Obsidian system. They should be kept in Zotero only. And ideally, they should be cleaned up promptly.

Tags are portable, but collections are not.

Use collection to organize items by purpose, not about specific content. For example, if I creat a collection named Gaussian Process to collect all related papers, it turns out to be a bad idea because I have to creat more sub-collections to distinguish GP papers used in different fields, and the number of collections will soon grow to very big. In such circumstance, a better choice is to use tags, which is not a hierarchical tree-structure.

For a good illusration, I use parent collections named by projects and sub-collections named by specific goals or tasks in the project. For example:

My Library          // zotero default
|- Project XXX 
|   |- AAA          // One specific task
|   |- BBB          // Another specific task
|   |- ...
|- Writing
|   |- Journal CCC
|   |- Conference DDD
|   |- ...
|- File Types (can also be organized using tags)
|   |- Books
|   |- Thesis
|   |- PPTs
|   |- ...
|- ...

Song Wang uses ZotFile, Markdown Here, Org-mode.

Obisidian workflow that connects obisidian. My current trial is heavily based on this post.

Here has a lot of discussion about using Zotero with Obisidian. Didn't follow them yet.

Christian in this post uses Obisidian plugins QuickAdd, Dataview, obsidian-citation-plugin to automatically creat notes from Zotero notes. Feels too complicated for me. Too much variability.

Regina Martínez Ponciano has a very comprehensive workflow that combines Zotero and Obsidian.

First, install all useful plugins listed below and setup accordingly.

Second, some advanced configurations cannot be synced or set easily, so I set them up via the user.js file. Particularly, I pick out corresponding preferences from (see Zotero knowledge base here):

C:\Users\YOUR-USER-NAME\AppData\Roaming\Zotero\Zotero\Profiles\***\prefs.js
/Users/<username>/Library/Application Support/Zotero/Profiles/<randomstring>/prefs.js
~/.zotero/zotero/<randomstring>/prefs.js

and put into a user config file:

C:\Users\YOUR-USER-NAME\AppData\Roaming\Zotero\Zotero\Profiles\***\user.js

Then, this file is backed in the Zotero folder in my Dropbox. Mac and Windows have different key combinations, so I keep two version prefs-win.js and prefs-mac.js.

Downloadable from official plugin collection.

Set extensions.zotfile.pdfExtraction.formatNoteTitle in config editor to empty value, to remove the useless title of the extracted annoataion. In this way, I can see a brief note in the item panel, rather than just the title. The added date of the extraction is still available from the particular panel column.

Under ubuntu 20.04, this line must be runned from terminal to let Zotero handle links like zotero://select/library/items/JXXCQQGR (Source link):

xdg-mime default zotero.desktop x-scheme-handler/zotero

TBD.

https://martinezponciano.es/2021/04/05/from-references-to-fleeting-notes-zotero-and-zotero-plug-ins/

1. From references to fleeting notes: Zotero and Zotero plug-ins
2. From fleeting notes to literature notes: Obsidian
3. From literature notes to permanent notes: Obsidian
4. From permanent notes to a manuscript: Pandoc