What is your workflow to understand and learn whatever you are studying?
Hi everyone, I want to know what your workflow is when you are studying something you want learn and understand. I'm not talking about memorizing but really understanding the knowledge so you create your own concepts of things.
I usually read slowly, create notes in my own words and try to connect the knowledge. But what do you guys do? I want to take from this conversation some insights to incorporate in my workflow.
It looks like you're new here. If you want to get involved, click one of these buttons!
Good question. FWIW, I have been asking this myself lately. Here's my current experiment https://srid.github.io/learning-fsharp/
Hi! From your website it's hard to come to a conclusion about your workflow. How would you describe it?
@Jvet what you are describing is essentially exactly what the ZK workflow is designed for. Perhaps even moreso the adoption of the evergreen notes concept by Andy Matuschak, which is essentially zettelkasten with an extra layer (the evergreen notes).
Are you familiar with the ZK process at all? If not I recommend you start by learning that. Don't get caught up in the terminology confusion, focus on the principles. Start simple: write atomic notes focusing on individual standalone ideas, add links to other notes when it adds value, iterate and refine. That's it. Focus on learning the fundamental principles and patterns over specific workflow mechanics – let your workflow emerge naturally based on your specific needs.
To give you an idea of a workflow (not to tell you to use this one in particular) here is how I generally work: (which is not necessarily the same as how others work!)
Other example workflows can be seen in @ctietze's live stream recordings (https://christiantietze.de/posts/2019/09/book-processing-video/) as well as Andy Matuschak's recorded live stream ()
Thank you for the detailed answer! Have you noticed improvements in your level of understanding with your workflow?. I am familiar with the zettelkasten method! I have my own zettles in Zettlr, maybe I migrate to org-roam at some point, maybe not.
In my case I have found that using outlines, I use org-mode (emacs) and it helps me in visualizing a deconstructed version of what I want to learn or tackle; helps immensely to create a mental map and understand how things are related.
@Jvet Yes I've found my thinking and my analysis of source material has improved significantly, as has my writing. I also use outlines quite frequently. In particular I've spent a lot of time learning more about how to examine a source and identify the ideas, themes, etc, which was not something I needed to "do" before since I didn't have a system in place that actually encouraged me to do that, which is exactly what the zk does. When I process a source I try to build my own idea-focused outline from the work wherever possible, whereas previously I would take notes by chapter/etc which was less effective. Also my notes previously were source-specific with little by way of synthesis notes, while now I focus on the synthesis wherever possible as this is where we build our ideas and theories. And once we have our own ideas and theories well ordered in our minds, with meaningful labels/titles and strong links between them, it becomes easier to read new sources because we are subconsciously looking for how the new material fits into our internal schema of ideas.
I guess a short way of describing it is that previously my mental schema was largely fragmented with a few key areas built up around major needs (like career, family, etc), while now my schema is becoming much richer and much more integrated overall.
I am a Zettler
This is a great point. Something that I've found very powerful emerging in my own notes is to focus on building towards executable strategies (Matuschak term) for various domains because these act as a mechanism for transforming the ideas into action. In addition to making explicit what we may hold implicitly (in our minds, or winding through our notes but not made explicit as a standalone note) they also serve as a core reason to continue growing the body of notes – an end (output, which is itself just another note in the ZK) in and of themselves that provides at least a minimal justification for the means (daily work in the ZK).
My current approach
Gather information -> decompose to base principles -> incorporate principles -> design strategies -> test and evolve strategies
So to me the ZK can be more than just navel-gazing meditations on intellectual interests (not that there's anything wrong with that!); it can become a vehicle through which we learn to become better across domains and disciplines.
Edit: I should add, the ZK/evergreen writing method itself is an executable strategy for building executable strategies – a meta strategy then.
I like how you describe that you link what other people is saying about the subject with what you believe and from there you generate knowledge.
Most of the time I don’t do that, I just take notes about a subject, writing down them in my own words but hardly ever do I compare things in the way you suggest making a synthesis of everything. Also it seems interesting how you point out is important to notice what the author is trying to communicate, like what is the broader topic here?
Oh yes, I agree with you that practicing is key. Last day I was reading and learning about variables in statistics and started to see how everything I have ever read falls into the different categories of variables. To practice it I am trying to ask myself what kind of variables are these? What is the research question here? every time I read a science article or see a data set of information.
Thanks for suggesting how to start selecting my sources of information and the diverse concepts related to knowledge work.
I agree a lot with what has already been written. Especially with regard to applying knowledge and reflecting upon what is read/learned.
One technique for reflecting, which I have recently (last six months) started using is to mainly take notes in the form of questions. Instead of writing out explicitly what is in the source text (f.x.), I will try to formulate one or more questions about the text for me to review.
As an example, I could have written the following note:
To formulate it as questions, I could write:
Note that I only write the question. I just think of the answer. When I am done reading I quickly review that the questions make sense and then I leave it on the table for the next day.
This last gap in time helps with practicing just a bit of recall, instead of having everything fresh in the mind.
I've found a few benefits from this approach: 1. writing questions can be really difficult (that is good!) 2. I sometimes can't answer my own questions, which is a good check to see if I remembered the text well enough. 3. I sometimes think of questions beyond the text, which prompts interesting new directions.
@henrikenggaard This is interesting because you are (intentionally or unintentionally) practicing one of Wozniak's 20 rules of knowledge formulation: Redundancy.
What this does is "attack" the particular piece of knowledge you want to retain from multiple "angles." This helps ensure multiple strong neural connections, and multiple strong connections makes recall easier.
The active recall bit is important for memorization since it forms much stronger connections.
I'm very interested in the intersection of zettelkasten and spaced repetition, since I used SR extremely successfully for a while but was overloaded due to the Collector's Fallacy. I have a long-term goal of integrating SR into my ZK system at some point to try it again the right way.
TLDR of the below: Spaced repetition and ZK should be mutually supportive and create a virtuous cycle of insight if implemented correctly, with ZK acting as the filter to reduce Collector's Fallacy pollution of flashcards and ensure understanding before flashcards are created, and insights from the interleaving effect of SR feeding back into the ZK.
What's interesting is spaced repetition provides a slightly different model for learning and gaining insight than zettelkasten. I'm not convinced they are mutually exclusive but instead can be mutually supportive: use the ZK to synthesize and develop ideas and insights, then encode those few golden nuggets into flashcards (executable strategies developed in the ZK are potentially good candidates for this). As the flashcard sessions are executed in random order new insights may appear which are then fed back into the ZK notes. A virtuous cycle.
Once I have a good base of information in my zk with a strong set of my own integrated ideas I'm considering expanding to include spaced repetition memorization of a few key points from the notes. Obsidian has a plugin that round-trips with Anki enabling easy Q&A flashcard formulation directly in the notes themselves, but I haven't tried it yet. I may build flashcards manually instead. When I used SuperMemo the memory effects were incredible, far better than with Anki.
But SuperMemo as it is often described and demonstrated (even by the creator in his screencasts) encourages the Collector's Fallacy and memorizing the thoughts and ideas of others by making it extremely easy to create flashcards from pasted text during incremental reading. This is seductive and leads to memorizing words/phrases/concepts/ideas as written by others. Of course you can create your own flashcards and modify the text that is memorized instead of just using their words verbatim, but its easy to just highlight a word or phrase pasted from e.g. Wikipedia or whatever and create a cloze deletion. This is directly counter to the ZK philosophy, although again SM can operate like the ZK but it makes it seductively easy not to.
My hope is to eventually begin using Anki to get daily reps in for memorization of the most important bits of info from the ZK, which already be filtered to only contain the information most important to me. So that should cut down on the flashcard overload. Anki's algorithm is not nearly as refined as the SM algo (Anki requires much higher mental load than SM with higher burnout and less effectiveness) but it's an open platform vs SM's highly brittle proprietary system.
This approach would also hopefully ensure the flashcards that would be generated are based on synthesized material, which satisfies the first two rules from the above link: (which again, SM makes seductively easy to ignore)
@davecan I am a quite avid flashcard user myself (with Anki) and it is probably right that I got drawn to the question-method from Wozniak's writing, since I have read it, but I didn't explicitly think of it when I decided to try it.
Flashcards are still part of my repertoire of tools, but I use it much less than I used to. During my master studies it was my primary method of note taking. It was intense, but also very fruitful if I should say so. However, my current work as a Ph.D. needs a different kind of interaction with knowledge, so I spend much more time on Zettelkasten.
I wrote a program to generate flashcards (in Anki) from my Zettelkasten (text files). It is very idiosyncratic, but it fits my needs exactly. I've written about it first here, a bit more on the experience here and finally the Github repo.
@davecan I am curious about this:
I like Anki and have never used SM. How is this different? What does SM do?
@henrikenggaard Anki is based on the SM2 algorithm (written by Wozniak) with some tweaks by the Anki author.
Last I checked (couple years ago) SuperMemo (created and maintained by Wozniak, who is very likely the foremost researcher on flashcards and flashcard-related memorization in the world at this point) was on the SM18 algorithm. One significant change made with SM17 is reducing the load through a determination that the act of creating the card itself provides sufficient memory imprint to eliminate essentially the entire first week of reviews. That means you don't get a review until roughly one week after creation.
I've seen Anki vs SM compared as trying to fly a Cessna vs trying to fly the Space Shuttle. After using SM and its utterly massive set of options and stats I agree.
My opinion is that the memory imprint of SM is far better than Anki.
That said, SM still produces a heavy load if you fall into the collector's fallacy, which is too easy to do with its incremental reading functionality. The IR function is amazing – if you want to memorize what someone else said. In principle its possible to use SM as a Zettelkasten, and judging from how Wozniak writes about it his approach to using it seems to have a lot of overlap with the ZK approach in several respects. He literally wrote a small book on memory (on his site) using SM, incrementally, as a series of small notes gathered together – sound familiar?
But most people don't use it that way, and using it that way requires a lot of time and effort and expertise with the program. And its a very proprietary program that only works on Windows because it is written in Delphi which is an old language and framework for building Windows apps.
Imagine it as a PhD research app that has grown organically for 30 years under a single dev. Because that's basically what it is. He uses the money from the sales to cover his living expenses so he can continue updating the app, so people will buy upgrades so his expenses are covered so he can make updates, .....
And its only maintained by him, so if he gets hit by a bus and there are no further releases then in five years Windows could change how it handles something behind the scenes and all that knowledge is no longer accessible.
So as incredibly powerful as the capability is (and he has a lot of stats and research to back it up) the insistence on trapping the algorithm in the software makes it unusable for virtually everyone else.
The sad part of the story is that SM isn’t available for apple products. I used to have a PC and I bought it to support the developer.
Nowadays with my MacBook I think I’ll go with memosyne as I’m not a big Anki fan.
You definitely can run SM on a Mac but it must run in a virtual machine. I actually have an SM VM on my Mac right now, and occasionally turn it on to pull a batch of flashcards into Obsidian to convert them into proper notes. Going from single-fact quantum notes to merely atomic notes.
Which one do you recommend?
I've had a good experience with the Parallels Desktop over the years. I've used it for at least 10, maybe 15 years now, running a lot of engineering software, and never had a problem.
@davecan Thanks for the answer. I read a little more on SuperMemo's scheduling algorithms and it definitely seems like Wozniak put a lot of thought into that topic.
Incremental reading never really sat well with me. Of course, I have not used SuperMemo, so I can't say that I have truly tried it. But from my own experience of writing flashcards the one of the biggest benefits was the need to really digest and grok material before I could write a good flashcard. Also, very little of what I read comes on webpages However, I have only read about it so perhaps I am missing something subtle.
Parallels is also what I have.
@henrikenggaard Agreed. IR is very good if you formulate the material into your own words. The problem is that its too easy to just highlight some material and turn it into a cloze deletion flashcard. I prefer synthesized knowledge which is one reason why I moved away from it, although I didn't know about ZK so I didn't really have it well articulated as a reason at the time. It was more of a gut feeling that memorizing all these facts was interesting but what's the point?
For me I was studying some intense material so it worked well for getting through exams, but it was difficult to turn that into a reason for grinding away at SM every day for the rest of my life.
ZK by comparison makes learning fun – and based on the way Wozniak describes using IR I suspect he uses SM and IR in the same way as we use the ZK.
But its much easier to deal with plain text files in a simple app than it is to constantly wrestle with the researcher's own personal app that meets his needs that he also happens to sell to fund his further research.
I get the general gist of what I'm studying. Then I try it if it's something like programming or hacking, then a day or two later I try to explain it via writing as if I were explaining it to someone who was unfamiliar with it. I just repeat the process and slowly fill in the gaps where I fail.