A question I am regularly asked is how to structure Capacities. Here are my best and current thoughts. I also add some reference materials throughout to download if you’d like.

Disclaimer: this is what works for me. If you read and think this doesn’t work for you, that’s ok! You will have a reason for disagreeing and that will help you define the vision of your perfect set up.

Also, you’ll hear or see me refer to objects, content types, object types and databases. They’re all referring to the same thing.

The advice

1. Have the lowest viable number of objects/content types

2. Have the lowest viable number of tags

Achieving both of these means not neglecting collections, which can be thought of as an object’s subgroups.

Lowest viable number of objects/content types

Why?

Less set up, less maintenance, less confusion, less friction.

What to use instead of objects?

Collections, where possible. Collections are subgroups of the object type.

Setting objects up is not difficult, by why set up ones that might be surplus to requirements? So how do you work out the requirements?

I’d suggest that every system needs objects that fit under three umbrellas: input, processing, output.

You want a place to put what you have consumed (input), which has sparked interest in a variety of things (processing) and that might inspire you to create something original (output).

These are very broad groups, allowing you to really customise Capacities to your liking.

Input

Capacities give you some basic, or automatic, content types for captured content: tweets, images, pdfs, audio etc. These are automatically updated whenever you paste an example of one into your space.

However you might also want another object. I recommend a media object, encompassing any and all media you have consumed: books, podcasts, articles, films. Many people might prefer to separate these, but I encourage you to take a different line here.

There will be a very large amount of crossover in the things you want to track across all media type, or at least there is for me.

For every media type, I want to track the fields seen below. There are some properties (author, director for example) that will clearly depend on the type of media being discussed. But I advocate for a ‘hide when empty’ property here to save setting up 3+ content types where there is only 1 or 2 different properties between them.

To separate different subgroups (books, podcasts, films etc) you can use collections. Collections allow you to sort, filter and view your information in as many ways as a whole object would. They are quicker to set up and very flexible.

Once you have decided on your input objects, you can plan your processing objects.

Processing

I see Capacities as an app that helps you break down what you learn from your inputs into their component parts, to develop in the way they need. I speak about this here:

What you learn, or what stands out to you from your inputs will likely be related to your interests or your studies.

What are the types of things that come up a lot in your reading? Or in anything that can be considered an input in your system. What of those things are you interested in taking notes on?

For me, an international relations student, I want to know about people, events, countries, theories etc. I have objects for most of these.

So what defines when you need an object?

You first need to think about the properties you want. This is because properties are, in my opinion, the key differentiator of object types.

Start by making a list of what you want to track per object. Then compare this to the set up of your existing objects (or to each other if you’re starting out).

Here are some things you could consider with properties

I believe that if there is enough crossover between the planned properties of two (or more) objects, the object should actually be given a name that encompasses all of the functions you’ve mapped out, and then you can create collections to represent the differences between them, and you can add ‘hide when empty properties’.

We saw this with my media example above. If author/director are the only differentiator between a proposed book or film object, I would rather unite all books and films under a name that encompasses both of those things: media.

Another example is in my topics object. I have a collection of international relations theories, amongst many other collections. I could have separated each collection by creating their own object, but I’m tracking the same properties. I essentially want to know which people, events, definitions etc that each entry in the object is related to. I don’t need to create that set up multiple times. I just set up a collection per subgroup.

I learned this when I had a content type called people (indicating ‘famous’ people) and one called academics. I was only regularly tracking two more properties in academics than I was in people. Yet this rigid separation of people and academics is actually quite short-sighted. What about people who are both? Where do I put them?

I realised my error, added the two missing properties from the academics object to the people one, and made them hide when empty. I transferred all my academics over to the people object, and now I have the flexibility of the collections to support me. It’s actually more functional than with two separate objects because people can belong to more than one collection.

This is advice, if you’d rather split them up then go for it. But I do really recommend planning things out beforehand, just to save you time.

Output

You might want to create your own work following your note-taking- be it assignments like me, or blog posts, or anything else. You need to decide how many objects this needs to split into. I recommend following the same advice above and planning out properties.

Lowest Viable Number of Tags

Why

Easier to maintain, keeps the use case of tags and collections separate thus reducing confusion or time taken to make decisions.

What to use before tags

Collections!

Capacities tags are not like Obsidian or Logseq tags. They are not page references, they are their own entity.

If you tag a selection of entities that can be classed as a subgroup of your object type, there is no functional difference in your interaction with that group of entities when you view the collection, and when you view the tag.

For example, tagging a group of French presidents as #French Presidents, and also putting them in a French President collection within your people database, gives you the exact same view. I can sort, filter and view the content in the same ways. Given French presidents are a subgroup of people, I should have just stuck with a collection.

Here are some more examples of subgroups of an object, for your reference.

So when do you use a tag?

The moment at which your want to unite content from different objects under the same heading, is when you need to create a tag.

I should not have created a French Presidents tag until I had consumed another piece of content about French presidents, that was not entered into my people database. This happened recently when I read a tweet about Franco-American relations. My French Presidents tag now looks like this:

Tags, therefore are like better, more functional folders.

The question becomes “is what I want to tag a subgroup of an object type?”. If yes, collection. If no, tag.

Here are some tags I think everyone should have though. These can be used across your space and show you everything on your todo list across your content types, or everything you’ve rated 5 stars.

Summary

Getting to grips with Capacities structure will help you unlock what I believe to be one of the most exciting, joyous and rewarding note-taking experiences around at the moment. I hope this advice is helpful, but let me know if you have more questions and I shall do my best to help!

If you’d prefer to watch a youtube video: