I use Obsidian for research and I ran into a problem with organizing the information.
There are things that are part of something. For example, I have a computer mouse. It is contains wheel, buttons, sensor, etc. I have some information about wheel and part of physics theory that can be applied to wheel (it’s not theory only for mouse wheel). I made the new note with part of physical theory, and I should link this note in the text about wheels. And I can’t find a good way to put muse, wheel and other parts into my database.
What I tried:
I made notes: “mouse”, “mouse wheel”, “mouse button”,…, “physics theory”
I linked “physics theory” in “mouse wheel” and I linked “mouse wheel” and other parts in “mouse”.
Problems: first and second relations are the same in Obsidian, but in fact they are very different and it is very easy to get confused. And this is not all! I had another physics theory, “physics theory 2” that can be applied to all mouse. If I link" physics theory 2" in “mouse”, then in “mouse wheel” and others this link will not be. Linking “physics theory 2” in" mouse wheel" and others I think is a bad idea because it will be hard to understand what I meant.
I would be very grateful if you could help me. Thank you.
Yes this is a tough question. I am using Obsidian for research as well and modeling more complex relationships between notes is something I wish I could do every once in a while.
My first thought is this. It’s not because the relationship is not exact that it is useful to model if precisely. You could work harder to model the relationships accurately, but will these efforts help you realize your projects in the long run?
If you do decide to better model the relationships between these notes, there are a few options.
This post has the seeds of a solution. What we are looking for is a way to label the links we make so that we can identify the relationship types. The first solution is to use the Dataview syntax and make links similar to
link_type:: [[LinkTarget]]
This does some of the work and Dataview will help you craft queries to exploit these annotations.
In the long run I believe what we are looking for are ontologies, and better tools to manipulate and edit them inside Obsidian.
One thing i’ve been experimenting with recently is using the Graphiz extension to create rough schemas of the ontologies I use in my notes. The diagram defines the relationships types I want to use. I note what they mean beneath the diagram. Then, when I use these relations in other notes, I make a link to the ontology page.
So if I try to replicate this experiment with the problem you pose, I would suggest
Make a diagram of all the entity types and their relationships in a note, for instance an [[Ontologies/Mouse]] note. Write down how the ontology works. For instance your ontology could define a “isModel” relationship type between the mouse and the applicable physics models.
Use the dataview syntax to make the relation: `isModel:: [[Physics Theory 2]]
Where you use the relation, you can optionally make a link to the ontology page. If yiu ever forget what isModel means you can always go back to the ontology note to remember what you meant.
Note that the next version of Obsidian will improve the way links are handled in Obsidian and the Dataview syntax might not be necessary anymore.
I hope this helps. I believe ontology tools could be very helpful for some things in obsidian, but parcimony is important because the notes can quickly become very confusing.
It’s not clear from your post how you’re using the notes, and that’s the first thing to consider when choosing how to link and arrange them. “How will I want to access this info?”
But in general a simple answer is to use sentences:
[[Mouse wheel]] is part of [[Mouse]].
[[Physics theory]] is relevant to [[mouse wheel]].
Also there is a community plugin called Breadcrumbs which is meant to help define relations between notes. There may be other plugins that approach the problem differently.
@CawlinTeffid, @davidlandry, Thank you for your responses! Link types are what I was looking for. I understand that there is no one easy and convenient way to make them, so I will try to find the best one for me. I love the Obsidian community for your willingness to help.
It seems to me that until the concept of “Named Link” is introduced, there will be no flexibility. A connection, a link is the same object as all the others.
The link is currently used as an unnamed pointer. For example, Document1 points (refers) to Document2. What does this tell me? Should I throw them both in the trash? It doesn’t tell me anything. Communication, as an entity, sometimes has even more meaning than the objects themselves.
For programmers, a pointer is just a pointer. It does not require the definition of its properties. Therefore, in almost all programs like Obsidian, links, pointers, links are treated as something secondary.
Different plug-ins are “crutches” that only complicate an already deplorable state. Relationships must necessarily have a Type, Kind, Property… or whatever else you want to call it, but formally it will always be the name of the relationship.
I can not fully understand the Anнtype application, but it seems that the naming of Relationship Types was declared there.
I disagree that Obsidian is useless without this feature. As @CawlinTeffid pointed out, natural language is sufficient in a lot of cases. I think ontology tools would be useful in certain circumstances, but I also think that the best way to start a vault is just to type the notes simply and add structure later, as needed.
Also the Excalibrain plugin is a very powerful tool that goes in that direction. It lets you define different relationship types and exploits the structure by making graph visualizations with it.
I don’t classify Obsidian as a useless tool. Moreover, for me, acquaintance with Obsidian was a discovery. Nevertheless, I would like to have a full-fledged system in which, referring from one note to another, I could understand later - why did I refer to another from this document?
Right now I’m intensively “sharpening” Obsidian to control my workflow. And I am very pleased with the almost limitless possibilities of Obsidian! This is more than enough for my work. Well, connections … I just would like to. And in simple tasks, this would simplify the work space.
I hope in the future there will be an understanding of the importance of naming relationships. For example, in the task “Tie two logs”. The programmer put a pointer from one log to another. But what is a lumberjack to do? What to tie? Rope, chain, rope, fishing line, thread? Moreover, as I think, technically it should not be difficult.