Comment by kens | Hacker News Reader

Comment by kens | Hacker News Reader

I've been using Obsidian for a few weeks. It fits my note-taking style (Markdown and wiki) and I want to like it, but honestly it's a struggle.

My main issue with Obsidian is the steep learning curve because Obsidian is not self-contained; it needs many plug-ins, templates, and customization. I tried to keep things simple but it seems that to be useful you need Better BibTeX, Zotero, Zotero Integration, Pandoc, DataView, and so forth. All of these have their own learning curves and configuration, so it's a slog to get started. Obsidian seems designed for people who enjoy extensive customization and tweaking.

My second issue is that Obsidian is very free-form, so you need to decide on an "ideology" of how to use it. Do I want to go with Zettelkasken or Johnny Decimal or organize in folders or organize with tags or use Maps of Content? So far I feel like I'm ending up with a mess of notes, which isn't really helping me.

I started using Zotero at the same time. I highly recommend Zotero. It's essentially a PDF organizer and document citation system. Before Zotero, I had piles of PDFs in various directories and I couldn't find the information I wanted. Now I click in my browser and Zotero downloads the PDF, does OCR, auto-generates a detailed citation, makes everything searchable, and lets me annotate PDFs and take notes. Zotero is straightforward and has been a big help to me.

Zotero and Obsidian are commonly used together, but I haven't got the hang of the integration, so I'm open to suggestions. Specifically, I have notes and annotations in Zotero. Then I import everything into Obsidian and have a copy of the record in Obsidian. Now I have duplicated data. How do I keep things in sync? Should I be doing all my note-taking in Obsidian, or use Zotero's notes for a first pass through the document, or keep some notes in each?

I'll try.

1. What are you trying to do? If it's simple research off the web, simply copy some files and paste, or use MarkDownload and get fully formed markdown from a page to put in your vault.

2. Can you code? You don't need to code, but things make a lot more sense in Obsidian if you know markdown, CSS, and maybe a bit of javascript/html. I didn't use anything but markdown productively for a long time. Even now there's very few plugins I would recommend. ADD: When I started, I went through all the plugins and added everything under the sun. This is because I was stupid and thinking I was at some kind of nerd buffet. Once I realized I was looking at complexity upside down, Obsidian kicked ass.

Your second issue is a feature for me, so I can't help there.

As to your specific question, if you're stuck on Zotero, then add stuff there and make a one-way pipeline to Obsidian. Obsidian becomes your source of truth. If not, ditch it. I don't think it does anything Obsidian doesn't, but you should check with some users.

Remember KISS, Keep It Simple, Silly. One tool. If you're stuck using two tools, one-way data pipelines. Syncing two knowledge tools together sounds like a nightmare and I'd stop doing that right away.

One more thing: I decided on folders for long-term categorization, things I think will mostly never change. Then tags, nested tags specifically, for trying to freeform WTF I'm doing. Tag Wrangler is awesome for this. There are only about five extensions I'd always use. Tag Wrangler is one of them.

Thanks for your reply. For background, I'm doing technical/historical research. I have 100 documents on a topic (mostly PDFs, also some scans) and I want to keep track of them, pull ideas out of them, and write a coherent article. Cut-and-paste isn't enough at this point. Zotero is the perfect solution for me to keep source documents organized and I would pay money to keep using it. Researchers who use Obsidian mostly use Zotero too (it seems) since Obsidian doesn't handle citation management, tracking documents, downloading, OCR, PDF annotation, and other important tasks.

For your second point, sure, I can write tons of code. That's why I'm cautious with Obsidian; I don't want to get distracted by writing templates and JavaScript customization rather than useful work.

By the way, the basic Obsidian process I'm using is "Notetaking for historians", which is described at https://publish.obsidian.md/history-notes/01+Notetaking+for+... This process is very similar to the process described in the topic article.

My first Obsidian project was a research project for an upcoming non-fiction book. It's not academic, but I believe the synthesis/citing issue is generally the same.

I just looked at it. It has a little over 1,000 files in it, so I've been filing a lot.

I only have dataview, PDF to markdown, templater, and a couple other community plugins enabled. My decision was to dump everything, kit-and-kaboodle, into a Reference folder. No processing, no tagging, just dump it in there. Then I created a Topic Summaries and a Synthesis folder. The idea here is not to start categorizing and synthesizing until you're absolutely forced to.

Between Topic Summaries and Synthesis I create a book outline. Note that the folder names are things that are set in concrete. Then I created a tag system. As I begin teasing out ways to organize my research, simultaneously I create a tagging system. So one research item may eventually get referenced in several Synthesis Notes. I may go back and create a Topic Summary based on the syntheses or just ad-hoc. The tagging system links to that layer, not the data dump layer.

This allows me to continue dumping data, then join it up in various creative ways later. I don't even hook my synthesis into an outline directly. That's a separate and also creative step. Once joined and I've committed to an outline, I can chase down the original research and provide whatever footnotes or links I feel are appropriate.

Hope that helps. My big lesson was not to categorize my research as I gathered it. Usually my first guess is wrong. Instead use an abstraction layer and continue to churn through synthesizing many docs into new ideas. Then join those ideas into an outline.

Best of luck.

I agree that the learning curve can be steep, but it's worth it. The key is to keep things simple for as long as possible. There isn't that much you REALLY NEED to get started and to grow up until you reach a critical mass of notes (>=500 IMHO). Avoid falling into the trap of focusing too much on the system and too little on get value out of it

There are vault templates you can start with and build upon, instead of trying to build your own system. Some are very basic, but others are more elaborate and better documented. I have to mention mine since I've been working quite a lot on it, based on my own usage as an author/blogger. It's called the Obsidian Starter Kit [0]. It includes a clear structure, templates, plugins, automation rules for easily creating and filing notes, and a detailed user guide going from theory to practice.

The key parts it includes and that I recommend are: a simple folder structure (I chose to combine Johnny Decimal and the PARA method of Tiago Forte out of habit), tags (to ease search), clear naming conventions (for consistency), templates that define the base structure for the different note types and include the base tags to recognize those.

The combination of links/backlinks, tags and Maps of Content (even simple indexes based on tags) is key to finding whatever you need whenever you need it.

Regarding the integration between Zotero and Obsidian, having looked into it a while ago, I would recommend importing what you need in Obsidian, and accepting that as duplicate information, but enabling you to centralize as much as you can (reference information) inside of Obsidian. It's all right if you capture highlights in Zotero (I do that with Readwise), but I consider those as "transient" knowledge inboxes. You can find a copy of my Zotero template in [2].

In my own system, described in [1], Obsidian is my single source of truth. The other apps (eg Zotero, Readwise, etc) are just secondary and limited in scope. The important point for me here is the ability to connect everything within a single tool, which is where KM really shines.

PS, I've described my system a while ago in this blog post [3].

Hope this helps!

[0]: https://obsidianstarterkit.com

[1]: https://www.dsebastien.net/overview-of-my-personal-knowledge...

[2]: https://gist.githubusercontent.com/dsebastien/30c5d5e85a3dd2...

[3]: https://www.dsebastien.net/2021-12-03-personal-knowledge-man...

Reply on news.ycombinator.com