Building a Knowledge Tree

Obsidian Bookmarks

Bookmarks are the key to effectively using Obsidian as a non-relational database. The virtues of approaching your notes this way are several and I covered some in Freeing Your Thinking. In this article I want to develop the idea of using Bookmarks to construct flexible, interactive Knowledge Trees to organize notes. Just as a reminder, we are organizing information, not files.

The Navigation tab gives at best a static arrangement of information, while the Bookmarks tab provides a dynamic way to interact with, develop and access your information. Here you can create Views of your information which you can organize by domain, subject, activity, etc. Views are saved search filters, or queries. Your information, unlike your files, can happily exist in multiple places in your Knowledge Tree.

When you begin to build your Knowledge Tree, you might start by creating views which map to your current directories using filters. After all, you want to preserve the information you have put into your vault by organizing your notes in a file tree. Most people have a separate directory for fleeting notes, so you might make a view containing path: "fleeting-note-directory". But is such a view terribly useful? Browsing through all your fleeting notes is fun if you've nothing better to do, but it's not focused or efficient use of time if you are working on a particular topic. Wouldn't it be better to just see the fleeting notes related to my topic? So I save the filter My Topic \type:: fleeting\ alongside my views of the topic itself and have a list of fleeting notes only on that topic.

As you start mapping your old directories to views, you might begin to ask yourself at some point why you would ever really want to see all your MOCs in a list? Or Daily Notes? Or Logs? If you want information from these types of notes you typically create a note containing a Dataview query. Bookmarked filters, which I'm calling views, are like virtual Dataview queries. And, since you can do this without referencing file locations, you might question how much useful information was actually embedded in your folder structure. Then you might have a cup of tea.

One last point to stress: the bookmarks, which I will call views, are ephemeral. They can be created, duplicated, renamed and moved around without affecting the information itself. When focusing on a topic, drag it to the top of the tree and put it back when you are done.

The Mechanics

Commands

The following table shows the bookmarks commands with suggested Hotkeys which are most likely unassigned in your vault.

Bookmarks Panel

The panel lists all bookmarks, which, remember, are saved queries, or views. In the panel you can create, rename and delete (virtual) folders to organize your bookmarks. Bookmarks can be dragged around to rearrange them.

Knowledge Trees

When working with a database, we typically create views. Views are subsets of your information based on queries which have been saved. They can be used passively, as a way to look at certain information on a certain topic. They can also be used interactively, as a basis for more fine-grained queries, which, if interesting, can themselves be saved as views.

I have my Knowledge Tree organized by the different domains of my life, Living, Learning and Creating. In addition, I have a section with utilities I call the Toolbox.

Living

This is the structure of my Living section:

Daily Notes
Logs
  Exercise
  Piano
  etc.
CV
People
Leisure

Daily Notes is a view of type:: daily. Exercise and Piano are actual files with interactive Dataview queries. The rest are bookmark groups containing other views.

Learning

I use Learning for study and research activities. This part of the tree is the most complex in terms of structure.

By creating multiple views on a given topic and organizing them in folders I can essentially create an outline of my information using views. This kind of outline, again, is very flexible and can be reshaped easily as my knowledge develops.

This is an excerpt of my section on the normal distribution. Its "home" is in the bookmark group Learning/Data Science/Probability/.

Normal Distribution                 # All information on the topic
Normal Distribution                 # The reference note for the topic
                                      The 2 are distinguished by icons
Central Limit Theorem               # A bookmark group with other views
68-95-99.7 Rule                     # A view of a block in a note
R Functions for Normal Distribution # A view of a section of a note

Creating

This domain is for writing and creation of other content. Creating includes works for external and internal use.

Articles 
  Beta
  Alpha
  Idea
  Published
Stories
Video Scripts
Musings

I actually have more stages for my articles than I have views. For example, I have beta and beta review, but they are combined in one view. My query filter is /Stage:: (beta|beta review)/. (This isn’t actually necessary in this case, since the first term also matches the second, but this demonstrates how to combine multiple possible values in one view.)

Toolbox

In the toolbox I have the following directories, which should be self-explanatory:

Templates
fileClasses
    Class Definitions
    Value Lists
Fleeting Notes
Other
    MOCs
    References
    Permanent
    Atomics

Other than Templates and fileClasses, I rarely use the Toolbox.

Conclusion

I hope I have adequately conveyed some of the advantages of building Knowledge Trees. Maybe you noticed that the word "tree" is actually a misnomer. The branches of real trees don't reconnect, but ours can. What we really have is a Knowledge Network. Bookmarks allow us to leverage Obsidian's NoSQL features to free our Second Brain.

---

If you have found this content useful, please consider a donation to support future work.