Example of topic vs object tags to make more sense of tagging practices

ctietze

I shared this on my blog earlier this week:
https://christiantietze.de/posts/2024/02/example-object-tag-vs-topic-tag-programming-zettelkasten/

So I talked with a fellow iOS developer about tagging recently, then overhauled a note I found by accident and shared this as an example.

The gist is:

• I increased specificity of tags. The general #image becomes → #uiimage on iOS, #nsimage on macOS, which are actual types in the macOS API.
• I removed the topic tag #appkit which doesn't tell me anything interesting here.

Why is the object tag #appkit useless now? Almost all my programming notes are about macOS and thus #appkit, the macOS UI API framework. So it doesn't help find anything, and it'd produce so many notes that it becomes useless. Thousands, if applied consistently!

There are notes about the AppKit framework as a topic. These stay put, of course!

The thing is: for my very first notes on programmatic image manipulation, it sounded sensible to use a tag like #image. After all, I did not expect to need specificity here at all: "something with image in Appkit" sounded good. But with expertise, I need NSImage-related resources, and I can name contexts better. Like embedding images in text views, as wirh The Archive

All notes are malleable, so no problem.

By the way, new notes on similar topics I tag with the very specific API titles right away, e.g. #nsimage for image stuff on Mac. This changed in the past couple of years and I don't have to think about this, at all. I believe it felt similar back when I used the """bad""" tag. So chances are I could discover even more specific tags in the future and find the current one lacking At the moment, it's the best topic tag I know to describe this.

zk_1000

I don't believe "image" was a bad tag at all. It's more likely that, after refactoring, you came to find a bad tag. So this rule of topics vs objects should be applied on refactoring, but not on tagging in general.

#IDONTTAG

ZettelDistraction

Is there a clear distinction between topics and objects? Perhaps the intention is to use more specific tags. I see nothing wrong with a tag about a specific topic, and I could imagine vague object tags, such as #thing. I use tags to write broken English sentences and summarize my notes with tags. I don't often search on tags. It's more of an obligation to use them.

Abstract Objects, which is freely available online until February 20th, 2024. It's a reminder of the difficulties inherent in classification systems.

Andy

@ZettelDistraction said:

Abstract Objects, which is freely available online until February 20th, 2024. It's a reminder of the difficulties inherent in classification systems.

Thanks for pointing to that publication, which led me to its series Elements in Metaphysics, which led me to another new publication in the series, Ontological Categories: A Methodological Guide, which is also freely available until 22 February 2024. I read its introduction, and the following paragraph struck me as highly relevant to the preceding discussion:

I then move on to the meta-ontological discussion of the relationship between ontological entities, categories, and ontological roles. Following Oliver (1996), I suggest that the term "ontological category" should be reserved to pick out a particular ontological role that a philosopher finds in need of being filled in their theory, as they address a particular philosophical problem. Ontological entities, on the other hand, can fill one or multiple such roles, and thus can fall under one or more distinct ontological categories. This, I argue, tracks nicely what metaphysicians already tend to do – that is, it tracks how they assign distinct roles to distinct entities in their system and how they then employ such entities in different places in their overall metaphysical system. Of course, not everyone approaches ontological categories in this way; I conclude the section by contrasting this kind of bottom-up approach to ontological categories with descriptive and prescriptive top-down approaches.

I'd point out a couple of helpful concepts in this paragraph: first, its term meta-ontological discussion, which is what the present discussion is (@ZettelDistraction asked a meta-ontological question about alternatives to the topic vs object tag distinction); second, its proposal that the term category (or topic?) should be reserved for a role that you think needs to be filled in your conceptual system or note system to address a particular problem—and the author's suggestion that this is a "bottom-up approach to ontological categories".

Now, the notes in @ctietze's example are not about a theoretical system, but they are about a software system, and his refactoring of tags about a software system is analogous to the refactoring that a philosopher would do about a conceptual or theoretical system. The term conceptual engineering has become popular among philosophers to refer to this kind of work, analogous to software engineering.

Sascha

@ZettelDistraction said:
Is there a clear distinction between topics and objects? Perhaps the intention is to use more specific tags. I see nothing wrong with a tag about a specific topic, and I could imagine vague object tags, such as #thing. I use tags to write broken English sentences and summarize my notes with tags. I don't often search on tags. It's more of an obligation to use them.

Abstract Objects, which is freely available online until February 20th, 2024. It's a reminder of the difficulties inherent in classification systems.

I think it is a subcategory-relationship: Object-tags can be seen as topics, since they cluster notes according to a common denominator. But topic tags include way more notes than notes that use a certain concept.

The specificity is surely a by-product. But it is not the whole intent.

It is also about creating reliability and trust: With object tags, you can be sure why they are tagged. With topic tags the relationship between the note and the search results are not so clear.