Tag management

In my different vaults I use tags liberally. Sometimes I will embed tags that I don’t think twice about, in the fury of writing, like “dataScience” and “dataAnalysis”. In my conceptual world the two tags are fungible. They both start with the word ´data´so it’s easy for me to see them in the tag pane and go back and replace “dataAnalysis” with “dataScience” so I don’t have two tags for the same concept.

However, the situation can be worse. “Investor” and “Funder” are also fungible in my conceptual world so unless I actively think about it or come across something in Obsidian that makes it evident that I’m using both tags, I’ll have an inefficiency in the system. In this case, I’ll go to the tag pane and generally click on “investor” and forget that I also have “funder”.

My question, then, is what do users here do to manage tags to avoid redundancy? Is there a way to extract all the tags from a vault into a text file to make the cleanup exercise easier?

2 Likes

Welcome to tag hell. This has been a problem as long as tags have been around.

It is helpful to think about the tagging system to be set up before tagging to avoid exactly this situation.

Some reading for you (and others): The Difference Between Good and Bad Tags • Zettelkasten Method

What you are doing when you make topic-based tags is actually a form of qualitative coding called descriptive coding. This is a method used in social science research to identify themes and whatnot in texts.

It turns out descriptive coding is the weakest method of labeling content: How To Do Descriptive Coding — Delve

There are far better coding methods but they require significant effort: Essential Guide to Coding Qualitative Data | Delve — Delve

The solution depends on your use case. If your use case is collecting information with minimal writing of your own then you are applying a taxonomy to other people’s ideas and I don’t have a sound strategy for that. I’ve been burned by that myself which is why I no longer just collect notes. cf: Collector’s Fallacy: The Collector’s Fallacy • Zettelkasten Method

If your use case is largely to build your own notes that happen to incorporate the thoughts of others, but you want to write notes on your own ideas as a result, then my opinion is that it is far better to build outlines of links for topics of interest (rather than topic-based tags) and to use tags in the following two general situations:

  • to identify the object or “type” of note (e.g. #project or #literature-note or whatever)
  • to identify our own ideas and themes that emerge naturally from our writings (e.g. David Kadavy has a tag #IcebergPrinciple to denote writings about the large invisible efforts people put in before “suddenly” becoming successful)

The second one has the benefit that it shapes our thinking when we read/watch other source material for ingestion into our notes, because we are now actively looking for those materials.

Hope this helps.

4 Likes

Thanks, @davecan. There’s a lot of good material there. I will be looking into them.

I’m concerned more with tag management than tagging method. In both solo and multi-party exercises I’ve learnt that organic input produces the greatest insight and accuracy but it comes at the cost of higher noise. A good management tool would help with the noise.

If Obsidian allowed me to export a list of all listed tags (an export of the tag panel on the right hand side) that would be a good step forward. Even better would be an interface where I could remap tags, but a simple export would be delightful.

Question is what you intend to use tags for. Tags are a poor way of identifying exclusive categories. Even if you have a list, using many tags imposes a cognitive overhead and time cost. Remembering a whole cataloging system won’t solve all the issues because there will always be border cases and further possibilities.

Once you accept that tags may vary over time, depend on the tagger’s perception at the point of tagging and may overlap, you can consider them as adjectives rather than nouns. That makes them available as an aid to searching in the same way that links can be an aid to searching.

You’re right about potential cognitive overhead but periodic pruning of tags helps out. In my first post I referred to the possibility of using the tag “dataAnalysis” or “dataScience”. Depending on my purposes, it could make sense to merge them and just keep “dataScience”.

If I can easily prune, I don’t have to worry about what tag is right. I can just write first, then prune later.