My Second Brain – Zettelkasten
scottspence.com
scottspence.com
Here's my rather brief summary of the process:
* You create fleeting notes to capture ideas as they happen. They should be short lived notes that don't become the main store of your knowledge.
* You create literature notes as your read material. These should include your own thoughts on highlighted passages, not just quotes and highlights on their own.
* You organise your fleeting notes as permanent notes into your 'Slip Box' (taken from the original use index cards). Each note should contain a single idea and should be understandable when reading in isolation.
* You want to avoid burying knowledge in large notes as it makes it hard to glance at and link to other notes in a concise way.
* Notes are linked to other notes which support your ideas. This also help the discovery of new ideas.
* You use your slip box to help you do your thinking. You want to ask it questions, find the related notes that support/oppose the arguments and find gaps or newly related information.
* You can create index notes that help you find your way around.
* Part of the process is to help your understanding by writing. With a well maintained slip box, you'll never be starting from a blank sheet. You decide what insight/question/knowledge you want to explore, and pull together the notes that give you the body of research to get you started. You shouldn't need to start a new blog post by researching, that happens prior by taking smart notes as you naturally read what you're interested in.
Hope that's somewhat helpful. I'm still experimenting with it to find out what I understand correctly and what I don't.
I think it works with Dropbox too but you’d have to use something other than 1writer but not sure what
By organised, this means to filter, rewrite, conclude, etc into permanent notes. You don’t want to put in book extracts, or disjointed ideas without putting it in your own words.
I think ZK works very well in team environment. I make "issue logs" - when something goes wrong, or there is a particular detail to understand, I force myself to write out the issue and resolution into ZK. (Force, because it is so much easier to just solve and move on without documenting)
If you can maintain some discipline in writing summaries and updating for yourself, this solves "we discussed x issue 2 months ago, but don't remember {some, all} details now"
I keep sublime text open to my Kasten at all times so using full text search in a database of curated notes is so much quicker than trying to find the right project folder, or searching outlook.
I don't worry about one file - one idea, I just break things up how they make sense to me. The "rules" for ZK don't entirely make sense for engineering work and must be bent to your needs.
My only lament with plain text is difficulty in using figures / images in discussions with myself.
1. Code guidelines, which anyone can contribute to and we can discuss/review them via PRs. As each note is a small rule or idea, it’s easy to reason about. It aids discovery for new team members, but early days to test that bit out.
2. Application implementation patterns. I’ve added small notes in the codebase to outline how concept work, state management, styling, types, etc. Relevant to the codebase, but not specific to any individual code file.
3. DevOps playbook. I’ve been working on this in the past week and having it in a Zettelkasten format has been very beneficial to fill it out with the most important aspects. The set of markdown files covers how to do deployment, how we do security, backups, etc. Would love to turn that into a blog/video series when I’m done - which probably shows the power of having knowledge that is very reusable.
This is the part that I can't get my head around. How do you make such short notes, particularly with enough context to make it understandable in isolation?
> Each note should contain a single idea and should be understandable when reading in isolation.
It's really a guiding principle. Unlike index cards in filing drawers, markdown with wikilinks allows for infinitely long notes and a low cost for context switching. you can choose whatever balance of note size/quantity you want.
In my markdown-based second brain built originally in Obsidian and now in Dendron, I have around 350 notes with what is probably a Poisson distribution of sizes. The important thing is that jumping from note to note costs very little.
further, if a note warrants splitting, now you can do it as fast as you can think it. The founder of Dendron aptly describes this as a part of the amoeba pattern: https://dendron.so/notes/e780000d-c784-4945-8e42-35218a3ecf1...
I think it’s important to bear in mind the context this system grew out of (academic research) and ask yourself whether you really need such a heavyweight note-taking system. It’s also important to realize this system is not a substitute for accepted patterns of knowledge generation, such as summarizing extemporaneous notes and making time to review your notes frequently enough that the knowledge captured actually sticks.
Zettlekasten is a system layered on top of that and basically serves as a database for things you already know. If you’re not already doing the former behaviors then it’s likely you’re just going to be wasting your time hacking on a system that won’t provide you enough value to justify the maintenance cost. In fact, it’s likely that the overhead will be so high that you’ll lose time to practice the foundational behaviors that really pay off in the long run.
The remaining parts of it is the unexpected results for someone who hasn't used this sort of note taking system. The book "How to Take Smart Notes" explains this. You use the notes as a starting point to come up with new ideas. It's useful if you do regular writing, especially as part of work rather than for personal private stuff.
For example, I use this method for jokes. Much of the value is simply having a list in one place. If I want to come up with a new joke, I look at all my material as a branching tree. I can extend a branch by going deeper. I can combine branches to come up with something completely different. I can start green branches. I can write new material all day long by doing this, rather than by waiting for inspiration to nail me.
I think people get too caught up in the technical parts. It's just a branching system of ideas.
A lot of people seem to get caught up in LARPing by following strict rules before they see this usefulness and that rational approach holds back the creativity. It's like getting so caught up in the process of how to have sex that you never figure out how to actually enjoy it.
And yes, the card system would require these rules to make it work because otherwise you might not be able to find anything.
Another issue is that people get carried away with some of the better software for using this system and they end up creating their own rigid system. For example, the power in RoamResearch for me is the most basic features it shipped with. But it could also be a sort of "no-code" platform which people use to "no-code" themselves into a corner. I feel like you should have one "database" in Roam which you use only for this purpose and largely without structure. Make a mess and resist temptation to make it super organized. The more you get caught up in structure, the less you're thinking about exploring your ideas.
Strictly inferior to a spaced repetition system.
> Zettlekasten is a system layered on top of that and basically serves as a database for things you already know.
Things you already understand. If you already knew it in a deep sense it would be redundant. In reality zettelkasten enable you to think better by giving you a lightweight prosthetic expansion to working memory. You can get the same benefits by elaborative expansion but writing a ton of loosely related notecards is a lot less demanding than writing an article or book.
To-do lists are for things that aren't a real priority, otherwise you would be giving them your full attention and the tracking would be unnecessary. At most, it's helpful to track administrative minutiae. If you give something your full attention, tracking will seem like a non-sequitur.
Despite all of this, I have been finding the card system outlined here to be quite useful. The key for me however was to keep to a physical pen and paper version. Keeping notes exclusively online used to prevent me from being able to address them regularly with a clear mind. Forcing myself to synthesize ideas and refer to them again has helped me move forward in projects important to me. It's not the act of writing the note that is important, but the mindset towards having ideas that you want to develop.
Ultimately, most important aspects of life are beyond any productivity hack. You can't hack meaningful relationships other than by growing as a person, nor can you hack the strength to work on things that matter every day.
In a sense, the ultimate hack is to come to terms with your mortality and use this as an impulsion to focus on what really matters to you.
I disagree with this bit, but perhaps it’s because “real priority” is vague. To-do lists offload tasks from memory to ensure we don’t forget them. It’s totally possible for something to be a “real priority” (whatever definition) and someone forgetting about it.
Don’t confuse priorities with memorization.
I really suggest looking at the Eisenhower Matrix. Urgent and important tasks are the only types of tasks that really go into the category of "real priority." Todo lists are best used to manage tasks that are important, but either aren't as urgent as the most urgent tasks, or are blocked by some outside factor that you can't control.
It's important to finalize insurance plans for the upcoming year, but waiting until it's "urgent" may result in not having enough time to find the best plan.
Granted, maybe you don't have a problem keeping track of all important tasks -- in which case, that's awesome!
There's only so many things you can handle yourself. Todo lists are essentially the archetype of your second brain. They not only let you remember the non-urgent (but likely important) tasks, they also help you manage multiple urgent-and-important tasks, and in particular large urgent-and-important tasks. When such task is too big to fit in your head, you break it down into pieces digestible by your brain. A todo list lets you tackle bigger tasks, that break into more pieces, and/or tackle more of such tasks.
> You can't hack meaningful relationships other than by growing as a person, nor can you hack the strength to work on things that matter every day.
Hold my beer while I cross off the task of buying tickets to an event my wife wants to go to, because as much as I want to remember it, it's one of those things that are both important for a meaningful relationship and too easy to slip out of one's head under barrage of everyday churn.
> In a sense, the ultimate hack is to come to terms with your mortality and use this as an impulsion to focus on what really matters to you.
I wish this worked, but the brain adjusts and compensates even for memento mori.
>I wish this worked, but the brain adjusts and compensates even for memento mori.
Wouldn't this be a self-fulfilling prophecy? Even if you aren't feeling the understanding of mortality in the moment, you can still draw from the memory and awareness of its existence and your relation to it. You still have choices available to you that can't simply be ignored by neurological hand-waving. And if it turns out we don't have such choices, then the entire thing never mattered to begin with.
This is not true for humans with deficits in memory or attention, which is a common problem in a world of constant distraction. Even when the majority of the articles and techniques is geared towards people with underlying problems that should be fixed first, it doesn't mean the gained knowledge and functionality is useless to everyone.
I've been using TiddlyWiki on and off for many years, but 2-3 years ago I moved heavily into it (the Drift [1] distribution), and I haven't looked back. To me, it has become less about the tool and more about the information, and ensuring I have complete access to it, even 20 years from now. That includes data, metadata and even the software itself, regardless of the platform or OS.
I love my two columns UI and seeing many notes at once. Why does Drift only show one note at once?
This just applies to me, when my efforts are not directed towards making things and adapting on the fly, I tend to get sucked into thinking that I need a ton of knowledge to make things which leads me to find out systems which can help me manage that volume of knowledge, ironically and laughably before gaining that knowledge. Putting the cart before the horse !!!
I then go back to making things and I think these systems may be useful for certain styles of learning.
Does anyone here have similar experiences/thoughts?
edit: fixed idiom typo!!!
Nowadays, with all the SEO going on, Google is useless for showing you things like the author's site. To combat that, I was thinking of making an open community of notes, sort of a Wikipedia/WikiHow where every user gets their personal site and publishes notes, but with search being global. You could go on the site, search for, say, "fpv airplanes" and find people's notes on the topic.
It probably wouldn't have the high-level overview of Wikipedia, but it would instead be a sort of "StackOverflow for knowledge", where you could find solutions to minor annoyances like this: https://notes.stavros.io/software/monero-gui-syncing-stuck-w...
Instead of going through the effort of writing a whole new UI, I was thinking of making the server be a synchronization backend for Joplin [0] instead, and publishing the notes every time there's a change.
What does everyone think? Would you be interested in participating in an early alpha? Any other feedback?
I use my personal blog as my note taking location ([link redacted]). It uses blogger because I just wanted to get content published and available without worrying about theming.
https://twitter.com/lostintangent/status/1282047676377231360...
https://twitter.com/lostintangent/status/1334225746751983618...
https://twitter.com/lostintangent/status/1282850227695652865...
I built this so that I could manage my knowledge in GitHub, but edit my notes like I would with Notion/OneNote (i.e. commit/push on save). It also supports managing code snippets via gists, since I’ve found that my personal knowledge is composed of notes and code snippets, and I wanted a single, editor-integrated solution for managing them both.
Foam + GistPad recipe: https://foambubble.github.io/foam/recipes/write-your-notes-i...
Anyway, so I created something over the weekend called 'notenox' [0]. It creates a a JSON file of relevant information, one JSON file per note, with keywords and a "special" keyword prefix called a 'title' that mimics how I've actually been taking notes (email, so the 'title' mimics an email thread). For display, I consolidate all JSON files into a single JSON file and then have it loaded into the browser with some Javascript to group by title or keyword, along with doing all cross referencing and counting on the client end.
Creating notes is done through the command line, because that's a common way I interact with my computer, with different options to create titles, links, keywords, etc. I'm sure there are many different Zettelkasten implementations out there but they always seem so clunky and cumbersome. It's not hard, so the simple use case should be simple, nor should it proprietary or locked behind a SaaS.
You can see my personal notes in action, if you like [1] (sorry, not mobile friendly!).
[0] https://github.com/abetusk/www.mechaelephant.com/tree/releas...
The Zettelkasten thing is really nice but it keeps coming up as a 'new discovery' every few weeks.
Personally I don't have the attention to detail for it so I'll never do it.
I’d be interested to hear more about how your notes are organized and what they help you accomplish.
The idea of moving notes across 4 platforms as described in this article makes me wince, which is why I’m on OneNote for now - maybe not quite as nice for linking stuff but included in O365 which my work uses, cross platform, etc.
I always got frustrated with the implementation of Markdown and no being able to manipulate the text as I would in a .md document.
I have lost some functionality that Notion had, like for Kanban boards and such but being able to manipulate the plain text is where I want to be.
The only downside I have really is that I'm tied to using a VS Code now for writing notes, if I want to go mobile I can use something like GitJournal on android and make changes there, and there's now code spaces in GitHub so I could in theory use a code space in a tablet on the move, but none of us are out and about that much these days
I've started using GitJournal a bit with my tablet, though it's a really old one. I would like to make GitJournal more tablet friendly in the future. If you happen to try to it, please feel free to to tell me about the issues. I'll be happy to fix them.
PS: For markdown + kanban, imdone.io looks interesting. I haven't tried it out though.
But for me it'll never work, I know I'll start on it and then I have a rush project where I don't have time to keep the docs up to date and then I'll abandon it altogether :) I'm the kind of guy who fills up their desktop with everything they're working on and when it ends up being full just moves it to a folder "Old Desktop" (which is in fact a chain of "Even Older Desktop" folders :) :) :) ). Somehow it seems to work for me though, it really surprises me how I can find stuff back from 10 years ago just by seeing the icons of stuff I worked on around the time. And it's really zero effort which I like.
And really, I like relying on my memory.. If I don't remember something maybe it wasn't worth remembering. And discovering it once more may lead to other interesting discoveries.
But I'm glad it helps others forward in their goals! I didn't expect it at all though.
-Academics like Andy Matuschak who are trying to create better technology for learning or do work on really hard problems
-Productivity gurus like Nat Eliason and Tiago Forte who make money selling tutorials on how to use these systems.
For “The rest of us” there could probably be some gains from using these systems... but sticky notes, texts, notes app, Slack, wherever else you jot stuff suffice, and there’s a lot of administrative overhead & habit change required. I think I am gradually implementing more “linked notes” into my work as systems like Microsoft Teams make it possible to combine a wiki and a file directory easily. But I’m skeptical that this is going to become mass market in the near term.
1. Keep your notes short. I do 10 sentences max.
2. Link them as much as possible. Find a program that does that (Obsidian, Roam, Foam, etc).
It truly is a wonderful feeling when you do those two things many, many times. Instead of a YouTube rabbit hole, I can fall down a rabbit hole of my own thoughts just by clicking on a random note, following the links, and re-discovering some of my thoughts from months or years ago.
Unique IDs are easy. For example, Unix timestamp or just YYYYMMDDHHMMSS. Here's a commonly shared public database to get you hooked: https://notes.andymatuschak.org/About_these_notes
Would you break them into multiple notes? Group them together somehow? Have an "index" note with just hyperlinks to the rest?
Overall, they don't. Zettelkasten is just one of the folders within my note-taking app (Obsidian), and notes from books go into a separate folder. If there are some titbits that I see being useful outside of the book, I duplicate it into a zettelkasten note.
> Group them together somehow?
I have an "also see" section in my template where I link every other short note that's somehow connected. That's what I meant by "linking them as much as possible". The end goal is to be able to follow the links between similar notes instead of having to search for them. My library isn't big enough at the moment, but once it becomes, I should be able to click on a random note, follow a few links, and have an entire article with sources to back it up with little to no effort, using thoughts I had months or years ago.
> Have an "index" note with just hyperlinks to the rest?
Yeah I also do this. I have a note titled "00 zk" that's an index note for all the zettelkasten notes. I don't fill it out automatically at the moment, but I have a plugin that can do "link all notes within this folder", and I trigger it with a hotkey from time to time. That index note helps me to see similar notes and interlink them, because I never view that folder with a file explorer.
I've installed Obsidian and am going to give Zettelkasten a go.
After lots of trial/error and at the end of the rabbit hole I found Emacs with org-roam [2].
It has a steep learning-curve and often seems outdated, but it is also very powerful, has VIM hotkeys, and allows me to create the academic workflow I want - so far I can automatically create a note from my Zotero .bib library, fill it based on a template and the insert all my annotations from the associated PDF. Afterwards I also semi-automatically extract the references from that PDF, insert them into the annotations and then start to link everything into my Zettelkasten system.
Sometimes I wish I just stayed with VSCode/Markdown, but then I remember that I can now put "elisp" on my resumee :)
Overall I think that the "new" note-taking/Zettelkasten-systems is very cool and useful, but I wish someone would come along and create "the next big thing" which in my opinion is multi-dimensional notes.
Tiddlywiki/Tiddlyroam [4,5], TheBrain[6] and even Scrivener [7] seem like a step in the right direction, but they also make some other things overly complicated (convoluted UI, no plug/play export, bad editors, ...).
I want to be able to freely take notes on my computer the same way I can do on paper, and then be able to "super-charge" them by linking, aggregating, searching them. At the moment notes are "one-dimensional", i.e. I can only write from top to bottom. Compare it to paper where I can freely change my style of writing, add drawing, annotations, change directions, ... Writing on the computer just feels very restricting.
1: https://dendron.so/ 2: https://github.com/org-roam/org-roam 3: https://github.com/inukshuk/anystyle 4: https://tiddlywiki.com/ 5: https://tiddlyroam.org/ 6: https://www.thebrain.com/ 7: https://www.literatureandlatte.com/scrivener/overview
If it works the same way then is it a case of installing the extension with little config after that?
I'm not opposed to switching again
I liked Dendron more, as it seemed a more curated experience - the creator has a plan and is hard at work to make his "customers" happy.
"Dendron, the client, is free and will always remain free. It is also open source so anyone is free to make their own fork of Dendron.
That being said, I'm all in on Dendron and this is my full time gig. I want to make sure that developing Dendron remains sustainable. To that end, I plan on introducing value add server side functionality that folks may pay for."
PS: Why not switch to Emacs? It will only take a measly year to get comfortable! :)
Drift (https://akhater.github.io/drift/): I use this one as it has many nice features without becoming overwhelming
Stroll (https://giffmex.org/stroll/stroll.html): like Drift, but more features, including two column view
TiddlyRoam (https://tiddlyroam.org/): Have a graph showing the relation between your notes
I can keep my notes not only synced, but I also get version controlled markdown files. You can setup VSCode to accomplish the same tasks through extensions, but it's not 100% perfect.
The Archive is great software, but they also have a fantastic community that's researching this subject.
I would think that to collect knowledge about the world you would (ideally) need to have an ontological model of what are the categories of things we can talk about, and how those categories are connected to each other.
Creating such a model even if only in your head means organizing the knowledge we have and collect, and is not a trivial thing to accomplish.
Or does it not matter with Zettelkasten?
Use the same structure across all devices. Dont be afraid to archive whole directories when projects are done.
https://github.com/srid/neuron "neuron is a future-proof command-line app for managing your plain-text Zettelkasten notes."
Here are some sites that use it: https://neuron.zettel.page/examples.html
(ps: I'm not affiliated to neuron, just found it on the web and have been following since)
I'd love to give it a go,but that's a bit of a necessity for me.
Admittedly perhaps notion for fleeting notes, with a foam setup for managing permanent notes wheb at mt desktop may be a good compromise
I'm hoping something good is going to come out of notions backlinks support. $15/m is pretty steep on roam, and I do prefer notion for most things
If you use an iPad and have your notes in a GitHub repository you can use code spaces which is the vscode editor (I'm not sure if the extensions will work though)
Is there more to this that I'm not getting?
What I will say is I am in favor of anything that encourages people to write and create versus just consuming content. If fiddling with a new system encourages this, that is great.
At the end of the day, I find the overhead of highly structured notes gets in my way and distracts me from actually accomplishing anything using the notes. Obviously these systems work for some people! But I’ve found detailed engineering of knowledge doesn’t work for me...
What does work is a stripped down pair of tools: a) an inbox for capturing in-the-moment, which can be anything (a dedicated page in a notes app, my email inbox, a piece of paper); and b) a medium term “filing cabinet” for storing info for later (these days I use notion).
These let me not lose an idea, and reference it later. The key is having one default place to dump everything.
I mostly just dump in ‘facts’—my frequent flyer number, gift ideas, recipes. It’s just an external drive for my memory...
But the active process of reflection and editing and synthesis based on actual notes...I completely miss out on that. It’s not a piece of tech, so much as a process.
It reminds me of the Cornell notes system [1]: take raw notes in the moment, then condense and summarize them later. Reviewing and refining force you to internalize the content, not just record it.
[1] http://lsc.cornell.edu/wp-content/uploads/2016/10/Cornell-No...
And that lack of mass market adoption really annoys me because I believe in the benefits of a Zettelkasten and it would really increase creativity and productivity across many fields if there was mass market adoption.
Maybe Zettelkasten won't be ever 'mainstream friendly'?
I use Obsidian and Notion (before that I've used txt files, Evernote and Joplin) but I don't use Zettelkasten method. And personally I think there is already so many different note taking apps that building another one won't do anything. And there are some good open source apps that would use some extra help. Of course you can do whatever you want.
Notion is a place I keep tidbits of information... for example a database of my videogame collection, a history of all my previous addresses (I've globe trotted my entire life) or a repository of highly detailed professional notes on complex topics. Even then I'm still not sure that Notion is really a good long term solution since their staff could theoretically access all my private information without me knowing (I've had this confirmed to me via a support ticket).
So, while Notion has extremely good problem solution fit for a lot of my use cases, Notion however doesn't actually help me develop my thoughts on philosophy or help structure my thoughts on how to approach questions I have in my fields of enquiry or help me create content from original ideas. My physical Zettelkasten does.