Articles Are Not That Pure

This was not the first time I had run into the problem.

Using OBS with VTube Studio: Be a Vtuber has Vtuber in its title, yet the article itself is plainly a software configuration tutorial. Starting from a NEKOPARA Review begins with a Steam review and eventually turns toward life, change, and meaning. When My Cloud Provider Disappeared: Rethinking the Shape of a Personal Blog and What Is Worth Recording begins with a vanished server and lost archives, then arrives at the question of what deserves to be recorded at all.

Some articles are half technical practice and half a question directed at myself. In the LLM Era, What Exactly Is My Ability? is one example: AI coding provides the background, but the real question is what I can still call my own ability after losing much of my capacity to code unaided.

If categories are based only on what an article mentions, these posts could belong simultaneously to technology, books and screen culture, games, and philosophy. If I force each one into a single category, some important part of it always seems to disappear.

But perhaps the problem is not that the articles are “impure.” A personal blog is not a standardized library catalog. An experience can lead to a technical experiment, and a technical experiment can turn back toward an understanding of oneself. The real problem is that I have been asking one category field to answer several different questions.

One Field Doing Three Jobs

The blog currently uses five main categories:

• Resources
• Reviews
• Tutorials
• Reflections
• Learning as I Build

They all look like categories, but they do not live on the same conceptual level.

Tutorials and Reviews are closer to forms: they describe how a reader will encounter the article. Resources describes what the article provides. Learning as I Build describes the writing process. Reflections describes the depth the article tries to reach. At the same time, I also expect Category to tell readers whether a post is about technology, a work of culture, everyday life, or philosophy.

In other words, one field has coupled together three questions:

1. What is the article mainly about?
2. How is the article written?
3. Which specific objects and concepts does it involve?

When all three answers happen to align, the old structure appears to work perfectly. An OBS setup guide is a Tutorial; a response to a film is a Review. Nothing seems wrong.

The answers begin to diverge when an article starts from a game and moves toward the meaning of life, or when affection for a fictional character leads to an experiment in image upscaling. Adding categories such as Random Thoughts only creates a new drawer for every possible mixture.

The Taxonomy from the Cloud Provider Article Still Matters

In When My Cloud Provider Disappeared: Rethinking the Shape of a Personal Blog and What Is Worth Recording, I divided the things worth recording into Resources, Reviews, Tutorials, Reflections, and Learning as I Build.

That distinction still matters to me.

It was not trying to describe which subjects exist in the world. It was asking why I write things down. Resources can help someone else. Reviews can become both a cabinet of memories and a lure for kindred spirits. Tutorials preserve reproducible procedures. Reflections let me arrive at a provisionally coherent conclusion. Learning as I Build preserves the process of exploring, testing, and correcting an idea.

That is precisely where the old taxonomy remains valuable. It allows an article to deserve a place before it has become a definitive answer. It also avoids the trap of refusing to record anything merely because an early understanding may be naive or wrong.

The new structure should therefore not replace that logic. It should return it to a more suitable role: describing an article’s form and promise to the reader, rather than its thematic home.

Separating the Three Questions Again

The structure I now have in mind consists of category, kind, and tags. They are not three parallel sets of labels, but three axes with different responsibilities.

The Other Fields Keep Their Own Roles

The blog already has organizational fields more specific than Tags, and the new structure should not absorb them:

• shelf identifies the medium of a work, such as books, films, animation, or games;
• subCategory identifies a subdivision within that medium, such as galgame;
• series groups articles into a continuing line of writing deliberately organized by the author, such as Blogging, LLMs, or Reading.

Their relationship with Category, Kind, and Tags can be summarized as follows:

Each field must answer one kind of question, or they will eventually stick together again.

How Several Mixed Articles Would Be Classified

The new structure does not eliminate mixed articles. It gives the different parts of that mixture somewhere to go.

1
category: culture
2
kind: note
3
tags:
4
  - Yandere
5
  - AI Manhua Drama
6
  - SeedVR2
7
  - Image Upscaling

1
category: technology
2
kind: tutorial
3
tags:
4
  - OBS
5
  - VTube Studio
6
  - Vtuber

1
category: thought
2
kind: reflection
3
shelf: Games
4
subCategory:
5
  - galgame
6
tags:
7
  - NEKOPARA
8
  - Meaning of Life
9
  - Existentialism

1
category: thought
2
kind: reflection
3
tags:
4
  - Blogging
5
  - Writing
6
  - Archives
7
series:
8
  - Blogging

What Becomes Possible After the Separation

The most immediate benefit is that I no longer need to invent a Category for every article that is not “pure” enough.

For readers, “Technology · Tutorial” and “Arts & Culture · Note” communicate both subject and reading expectation. They say more than an isolated Reflection or a flat sequence of labels. For archives and statistics, content domains and writing forms can be observed separately instead of mixing technical tutorials, responses to works, and personal reflections in the same category chart.

Stable internal keys also separate data from display copy. Chinese can display 格物集 while English uses a name natural to English readers; URLs and article metadata do not have to change every time the visible wording does.

More importantly, the structure accepts that articles on a personal blog will naturally mix. Classification is not supposed to assign an article a single identity. It only needs to help a reader find a suitable way in.

Writing Down the Principle First

This article describes a structure that is ready to be implemented, not a feature the site has already completed.

The schema, interface, archive filters, compatibility for old links, and migration of existing articles still need to be handled separately. How Category and Kind should appear on cards and article pages deserves its own design discussion. The English wording can also be refined alongside that interface.

As for the briefly invented Random Thoughts Category, I am folding it back into the existing Reflections category for now. It was never a durable category so much as a note reminding me that the drawers needed to be reorganized.