Can someone help me understand what the point of Hashtags are if the search function is so strong? If I search a keyword, all of those instances show up as well as the hashtagged keywords. In Roam, the hashtag becomes it’s own page. It seems that in Obsidian, most people just turn topics into their own page by default.
Hashtags are a form of controlled indexing. Yes you can search for keywords, but you’ll also get false matches that you didn’t want. If you say, “I structure my notes so that won’t happen,” then all you’ve done is develop your own form of hashtags. I come from many years of using my own methods for taking notes. I was using hashtags before the term existed, but my “hashtags” are hierarchical so I can easily find subcategories if necessary. When I came to Obsidian (a few days ago) I realized that when I import my notes, I could continue to use my existing “hashtags” and their hierarchy, by simply leaving them as-is and using regex searching. So, if I want to see all notes related to my medical records, I’m looking for the “medical” hashtag, but if I want those notes specific to skin problems, then I’m looking for “medical.dermatology” hashtags. The point is, even if I ignore the built-in hashtag feature, I’m still doing “hashtags” of my own form. My method or Obsidian’s, it doesn’t matter, they do the same thing; define sets of notes and set intersections (more than one hashtag). Powerful stuff.
4 Likes
Some relevant thoughts here:
This article argues that tags/keywords are a better tool for Zettelkasten than hierarchical categories. Might be not as relevant for you if you are not into Zettelkasten.
1 Like
Hi laughedelic, thanks for posting the article. Having just read it, I have to say that it just doesn’t resonate with me; not because I’m “stuck in my ways,” but because it takes the position that the issue is black and white, and it is not. Tags / categories, hierarchical or not, are just another form of information representation and are not mutually exclusive with other data structures. Using them does not tie one’s hands in any way.
I have observed that some systems which draw associations to data in the form of a graph espouse the notion that this is the way our brains work. I first saw this perspective with a note tool that referred to a note file as a “Brain” that is managed graphically. Ultimately it was just an acyclic graph, aka tree, aka outline by another name in another shape, with it’s own limitations, being acyclic. I draw a finer association regarding this notion that our brains work in the form of a graph. I believe our subconscious makes these associations, while our conscious experiences reality as an ordered series of events; one thing happens after the next. Therefore I do agree with this article’s one point about timestamps and the usefulness of a chronology, as one view of data. We do tend to think about our experiences consciously in a temporal-geographic way: “I don’t remember exactly when event X happened, but it was after event Y.”
When I’m setting up a new data organization, I’ll start by asking myself what I know for sure that I have to be able to get out of it. Individual methods are neither good nor bad; they simple serve different purposes. For example, some of my notes absolutely must be in the form of a log. They must be an ordered chronological list, such as visits to a doctor or a log of symptoms that represent the progression of an illness. Having come to investigate Obsidian just a few days ago, I’ve been puzzling out the best way to represent a (potentially long) chronology, and it looks like prefixing each note with a timestamp is the best method because any search will return results chronologically, in case that’s useful. Still, hashtags, hierarchical or not, can provide a faster way to get to subsets of that chronology with results returned chronologically, akin to a database view or specialized index. It’s just another way to look at the data. In terms of a general association graph of notes (non-chronological), hashtags continue to serve this same purpose of a fast index to topic-specific data.
Being (very) new to Obsidian I have been puzzling over when to use hashtags vs backlinks. Similar to the OP’s point, are these redundant with respect to each other? For example, let’s say I want to have certain categories for notes, two of them being college and finance. Now let’s say I have a billing issue with my son’s tuition. I need a dialogue with the school’s Student Accounts department, and I’m going to pay a bill. Both of these categories apply. If I were using hashtags, I might have this at the top of my note:
#college
#finance
If I were using backlinks, it might look like this:
[[college]]
[[finance]]
The only functional difference is that if I’m looking for notes associated to one of these categories, the first method would have the list in the search results window by selecting the desired hashtag, whereas the second method would have the results in the backlinks window by selecting the desired category as a note. What are the advantages and pitfalls of each method? If my goal is to always be able to work with the notes outside of Obsidian (for other types of analysis) by being able to identify these same notes by category, is one method better than the other?
2 Likes
I’ll try to answer with examples from my own usage:
| Tags | Pages/entries |
|---|---|
#recipe |
[[Spicy New Orleans Shrimp]] |
#time |
[[16th Century]] |
#place |
[[Italy]] |
#person |
[[Jorge Luis Borges]] |
#quick-capture |
Whatever I was watching on mobile #quick-capture |
#bookmark |
[[On-line color palette generator]] |
#song |
[[Górecki - Symphony No. 3]] |
#article |
[[How I learned to stop worrying and love tags]] |
#follow-through |
There's a bug in the [[Awesome Script]]. #follow-through |
| … | … |
So, as you can see, I use tags to identify different kinds of information (either full, single-file notes or disperse little snippets), and pages for specific subjects, concepts and entities.
Let’s say I want to search my recipes for something with Marmite. Easy: tag:recipe Marmite. For me, there are two main advantages in using this over [[recipe]] Marmite:
- You don’t get an empty
recipes.mdfloating around. - In graph view, you can turn off tags and see how your notes relate to each other, regardless of their kind or format. (Is it relevant to see all the
#persons clumped together?)
Does this make sense to you?
3 Likes
Yes, it does make sense, and thanks for the detailed examples. Given my desire to be able to identify and work with the note files outside of Obsidian, I am inclined to use tags because search results can be copied, so there is an easy way to get a list of note files out of Obsidian based on a search query. Strangely, Backlinks doesn’t have this feature, even though it’s functionally similar to search results: a list of notes.
About this item:
You don’t get an empty recipes.md floating around.
It doesn’t necessarily have to be empty. These “backlink anchors” could contain some notes about what they really mean and when they should be linked. I admit, “recipes” is fairly straightforward, but I could imagine far more ambiguous terms. I could also imagine “recipe” being ambiguous in the right context. As a whole, then, the collection of these “backlink anchors” form a sort of data dictionary for link meanings.