JT's Weblog

Note App Design And Manageable Metadata

published: January 12, 2026 estimate: 4 min read view-cnt: 2 views

The first post of 2026! Two goals for this year

1. Make at least two side projects actively develop through the year
2. Keep writing blog posts on a weekly basis

Now let’s talk about the latest note-taking app in my head.

I started the concept of “hierarchical tag system” back in 2020

The reason was that the two common organization strategies — hierarchy and tags — both have their pros & cons

Hierarchy expresses data in great context, but fails easily with interdisciplinary topics

Tags, on the other hand, excel at interdisciplinary topics, but fail to express the context

Thus, the goal is to mix both strengths into one place

I failed several times to implement such a system.

There are even a few mature products (e.g. LogSeq) that have already implemented a “hierarchical tag” feature, but I still struggled to maintain it.

The main reason is that maintaining metadata is not a trivial task.
The cognitive load is real. Let’s break things down a little bit:

• first thing first, a good personal knowledge management (PKM) system should have consistent quality input
  • i.e. keeping notes consistently is already hard
• secondly, after finishing all the typing, you need to spend extra time adding quality tags
  • it is hard to measure what a “quality” tag is (at least, try not to spam tags)
  • tags are named according to the state of mind at the moment (thus, it is prone to change)
  • remember what are the two hardest things in computer science? cache invalidation and “naming things”!
• lastly, after being tortured by the writing and tagging, you question the value of your work
  • you want to do some quantitative research, and evaluate your notes with metrics
  • now you have three problems to solve instead of one
  • none of them is trivial, each one is a tough subject on its own

My solution to this dilemma is to split the workflows into two: top-down and bottom-up

Top-down means a predefined tag-system with built-in metrics.
An example is to set a few goals that you want to achieve.
Organize these goals into groups to express a simple hierarchical context

Bottom-up means your actual notes, which you can assign to multiple goals if it makes sense

This solution effectively:

1. keeps the hierarchical part where goals may be organized into groups
2. keeps the tagging part where a note can be assigned to multiple goals
3. reduces the cognitive load by:
   1. separating writing from tagging — the old workflow required users to tag immediately after writing; the new workflow lets you finish writing first, then drag and drop content onto predefined goals later
   2. automating quantitative evaluation — since goals are now embedded with metrics, evaluation happens automatically as long as you tag content regularly

And finally, I would also like to embed the mandala chart into the application to visualize the top-down goals

The mandala chart itself is essentially a rectangular grid, which doesn’t bring much value on its own

However, it helps users visualize their goal system spatially, enabling quick note-to-goal assignment. Users can map their keyboard layout to the cells of the grid, creating positional shortcuts

This aims to improve the speed of assigning notes to existing goals.

I haven’t yet built a prototype to verify my assumption, however, let’s just imagine the different approaches for adding tags:

1. drag and drop onto a visualized tag object
2. type and search for an existing tag
3. trigger a shortcut that maps to a visualized tag object

The second approach is commonly seen in modern note-apps; however, it doesn’t scale well when multiple tag systems coexist. A text-box is essentially a one-dimensional tool that makes it hard to distinguish between different tag systems or versions

The first approach introduces visual objects that help users understand the existing tags better. With a mandala chart layout, you can visualize each tag system separately

The third approach builds on the first one — the mandala chart with keyboard shortcuts — to improve assignment speed

Closing Thoughts

That’s all for now. It is just the basic outline of what I imagine a note-app can help in the area of personal knowledge management

I’ve been a note-app user for almost 10 years. I’ve found that developing a long-lasting PKM system is a slow-paced process that requires extra care to avoid obsolescence

I basically made my old knowledge base obsolete every time I explored a new method

Single source of truth (SSOT) is definitely another unsolved problem in my roadmap which will be explored after this side-project.

This is one of the side projects I would like to develop throughout the year.

I will try to implement this via the approach I used for developing a snake game

Hopefully, a playable web-based prototype would be available in the next week 🤓

Catch y’all in the next one!