# An illustration of the ZK method on YouTube

edited March 20

This [YouTube video] illustrates one way of implementing a digital Zettelkasten with Obsidian. My comments follow.

1. Note identifiers are avoided altogether in favor of descriptive filenames. This obviates the timestamp versus Folgezettel question. A tradeoff is that note filenames should be unique and should be considered immutable: a title cannot be changed without changing each reference to the note with that title. This seems like a minor inconvenience, in view of the use of backlinks to create indexes and hubs, below. UPDATE: see the thirteen point comment on the video by JET MECH, especially point 9) cautioning against naming notes or prematurely linking them.

2. Bibliographic entries (literature notes) are separate notes that link to a note titled Bibliography. The choice of writing bibliographic entries within Obsidian (the slip box component of the Zettelkasten) seems less efficient than working with a reference manager such as Zotero. How well Obsidian supports reference managers might have motivated this choice. However, switching from Obsidian (The Archive, Zettlr) to a reference manager is an interruption. A compromise (in Zettlr or Obsidian) would be to add within the note a citation (e.g., an ISBN or DOI reference), which could be added to the reference manager later. The note could then be updated with a pandoc reference.

3. Within the note titled "Bibliography," the bibliography is viewable in Obsidian (this will work in Zettlr also) as the list of notes that link to the note Bibliography. That's a clever design, and it's a way to implement structure notes and hubs without writing much more that the title in the structure note. An index of top-level categories could be implemented as a collection of links to notes containing the top-level category names. Visiting the top-level category note and listing the notes that link to it will display that part of the index. Obviously this could be extended indefinitely. For mathematicians and computer scientists: that's a coalgebra for you, otherwise known as a (vertex) coloring.

4. Zettlr supports backlinks by displaying the list of files that link to the current file. However, as in Obsidian, this feature is most useful if the filename is the note title. One way to find out is to experiment.

I might as well reorganize my ZK along these lines.

Post edited by ZettelDistraction on

Erdős #2. ZK software components. “If you’re thinking without writing, you only think you’re thinking.” -- Leslie Lamport. Replies sometimes delayed since life is short.

• edited March 22

Note identifiers are avoided altogether in favor of descriptive filenames.

This reminds me of the way Dan Sheffler is using his Zettelkasten:

Each file name describes the topic in the note. For example: Plato.md, Akrasia.md, Principle of Alternate Possibilities.md. When I want to split up a note that is getting too big, I create new notes that branch off the parent note by adding a hyphen to the original name. For instance Plato - Psyche.md is a child of Plato.md. This could, in principle, go as many layers deep as I like, but I prefer to keep the hierarchy as flat as possible. It is often better to add a new top-level note with a more specific title like Frankfurt-Style Counterexamples.md.

But I think his naming scheme is more elegant, because it also deals with the idea of Folgezettel in a really clever way. You can see neighbouring/branching notes at a glance to a certain level. And by limiting the level of branching you limit the Folgezettels to the most relevant ones, which imho can be more effective, than branching of indefinetly.

However, this kind of system can be limiting. As @Sascha pointed out it might only work, if your Zettelkasten serves a very specific purpose; respectively is monothematic. I am experimenting with this approach right now. Maybe I share my findings, if it works out.

• edited March 23

# # 0.2a3.0.22.0322 Further considerations

[[0000.0000.0000.0]] Workflow

Technical constraints and design choices weigh in favor of the current system of Folgezettel+Timestamp IDs. Some of the IDs can be made shorter: an ID with 14 characters is sufficient not to confuse Zettlr with the ID regex and ID generation patterns I have. The use of dashes in filenames led me to add dashes to hashtags. For example, instead of writing #simplicialset I now write simplicial-set. (Should have thought of that. Now I have to update my tags.) I'm used to Zettlr + Pandoc + MikTeX + LaTeX + Zotero +BetterBibTeX. I already have Pandoc export templates configured for LaTeX and pdfLaTeX in Zettlr, and I have adapted Pandoc's LaTeX template to work with them. And for writing outside of Zettlr, I have WinEDT.

I'm not sure about a comparable workflow in Obsidian.

As for eliminating IDs in favor named notes, an alphabetic ordering of notes could be misleading. I prefer viewing sequences of related notes at a glance. One disadvantage of Zettlr's linking system is that it duplicates the entire note title after the WikiLink including the ID. Previously the ID was omitted after the WikiLink. Now it has to be erased. That could be the fault of my ID regex, or it could be a change in Zettlr.

Viewing "Related files" in the right-hand pane in Zettlr is preferable to maintaining an index. Obsidian has a prettier display, but this is minor. Again, I expect the choice of ten top-level categories I settled on to break down (it's already starting to creak), and I look forward to the collapse. This is one place where internal ramification begins...

Erdős #2. ZK software components. “If you’re thinking without writing, you only think you’re thinking.” -- Leslie Lamport. Replies sometimes delayed since life is short.

• The last principle should be named "thoughts should be connected" since the reason for the principle of atomicity is to make a web of thought possible. But in practice, those principles are more post-hoc abstractions then actual guides.
• Awesome choice of the book.
• I really like her gestures.
• Her note lacks rigor for a scientific context. It is more simlar to what I did as a demonstration with The Godfather by Puzo. It is not appropriate for science to process a claim in a book without explicitely stating the state of justification. But of course: It could be that she leaves such rigor out to focus more on the overall Zettelkastening instead of the actual working with knowledge.
• I disagree with the sample not being a single thought. At a glance, it is one main thought (incentive structure by grades) and some auxiliary thoughts (e.g. the feedback mechanism overrides the "official" goal of grades). My suspicion is that she is using just one lead indicator (length of the note) which is just a rule of thumb to help beginners (not in the Zettelkasten Method but in critical thinking)
• The connection to the second note has great potential!
• The use of showing the graph view as unusable chaos to highlight the benefit of the Zettelkasten Method is exactly my beat. (Enough confirmation bias, Sascha!)

I am a Zettler

• @Sascha said:

The use of showing the graph view as unusable chaos to highlight the benefit of the Zettelkasten Method is exactly my beat. (Enough confirmation bias, Sascha!)

The poor usability (chaos) of Obsidian's default graph view is due mainly, I suspect, to the fact that the node types are not sufficiently visually differentiated and that the automatic layout is not sufficiently adjustable. But as I recently mentioned in another discussion, there is a plugin for Obsidian called Juggl (by Emile van Krieken) that can create more customizable graph views, with features such as a variety of automatic layouts, styling, and workspace mode. And when combined with another plugin called hover editor, one can edit notes from a hovercard that pops up when the cursor is over a node in the graph view. None of this substitutes for a having a good method as a foundation, but customizable automatic graph views are more and more able to provide novel usable visual analytics from the note system.

• @Andy The plugins look lit!

The poor usability (chaos) of Obsidian's default graph view is due mainly, I suspect, to the fact that the node types are not sufficiently visually differentiated and that the automatic layout is not sufficiently adjustable.

I don't think that is the reason. I think that the level of abstraction from the actual knowledge is to high. The foundation of my line of thinking can be found in McGilchrists insanely good book The Master and His Emissary. These are the quotes I am refering to:

The defining features of the human condition can all be traced to our ability to stand back from the world, from our selves and from the immediacy of experience. This enables us to plan, to think flexibly and inventively, and, in brief, to take control of the world around us rather simply respond to it passively. This distance, this ability to rise above the world in which we live, has made possible by the evolution of the frontal lobes.[21][#mcgilchrist2009]

Clearly we have to inhabit the world of immediate bodily experience, the actual terrain in which we live, and where our engagement with the world takes place alongside out fellow human beings, and we need to inhabit it fully. Yet at the same time we need to rise above the the landscape in which we move, so that we can see what one might call the territory. To understand the landscape we need both to go out into the felt, lived world of experience as far as possible, along what one might think of the horizontal axis, but also rise above it, on the vertical axis.[21][#mcgilchrist2009]

To live headlong, at ground level, without being able to pause (stand outside the immediate push of time) and rise (in space) is to be like and animal; yet to float off up into the air is not to live at all -- just to be a detached observing eye. One needs to bring what one has learned from ones's ascent back into the world where life is going on, and incorporate it in such a way that it enriches experience and enables more of whatever it is that 'discloses itself' to us (in Heidegger's phrase) to do just that. But it is only on the ground that will do so, not up in the air.[21/22][#mcgilchrist2009]

I am a Zettler

• @Sascha said:

I don't think that is the reason. I think that the level of abstraction from the actual knowledge is too high.

That is a good point. That is part of what I was trying to say, but I didn't emphasize that part clearly enough. One way of reducing the abstraction of the graph is increasing its visual iconicity. That is why "richer" graphic representations (such as rich pictures) can be helpful when recording knowledge. If it would be possible to summarize a note with an image, and then use that image as the node icon in the graph view, that would help the brain process it in a more experiential way instead of a more abstract way. But even using different standard pictorial symbols in the graph for different node types is a step away from abstraction toward experiential thinking. From abstract to experiential this would be ordered as: graph without different symbols → graph with different symbols → graph with rich pictures.

• @Andy

From Wikipedia:

Visual analytics is "the science of analytical reasoning facilitated by interactive visual interfaces."[2] It can attack certain problems whose size, complexity, and need for closely coupled human and machine analysis may make them otherwise intractable.

• The vast majority of people do not face the problem of complexity with their Zettelkasten yet. It is very premature for a) the overall scene and b) most users to use visual analytics to access their Zettelkasten.
• The Zettelkasten Method is self-scaling by nature. The issue of size and complexity is already solved by the method if applied properly.

However (introducing my positivity):

• Visual approaches to knowledge processing are very powerful, tried and tested (cluster mapping, mind mapping etc). I use it to take notes during discussions and as thinking tools.
• My above criticism aims only to the current thinking of putting yet another layer of abstraction between the actual terrain (knowledge) and the territorium (notes, links, groups). Visual tools are very powerful and we should not miss out.
• I already have some ideas on how to produce an graphical interface for the Zettelkasten without running into the above issues
• There can be learned a lot from thinking about large scale problems.

We (the research field) are running into the issue of building theory with to little application. I think 95% of the notes I see are vastly underdeveloped, sometimes outright useless after a couple of months. My caution is activated by the fact that vast majority of user I aren't producing complete and atomic notes in the first place. I love the enthusiasm of the overall field and really like to geek out. But on the other side, I care for the people who actually want to use the Zettelkasten Method to solve a problem in their lives and I fear that to many people lack the foundation to build on.

The visual interfaces could be very useful but they will dissapoint when the actual foundation the atomic note and the necessary knowledge work is not done.

I am a Zettler