Tag: taxonomy

Practically Paperless with Obsidian, Episode 21: Tags in Theory and Tags in Practice (And Never the Twain Shall Meet?)

blank tags in close up photography
Photo by Eva Elijas on Pexels.com

Welcome to my blog series, “Practically Paperless with Obsidian.” For an overview of this series, please see Episode 0: Series Overview.

This episode is less of a “how-do-I” and more of a “how-I-struggle with…” post. I have been working with tags for more than a decade and still struggle with them. I wanted to give fair warning at the top in case anyone is expecting brilliant revelations about tagging notes in Obsidian here.

I can’t remember when I first learned of the concept of tags. As someone who grew up during the personal computer revolution, my approach to organizing information was naturally hierarchical. From my earliest days using computers, operating systems organized files in a hierarchical structure, and it seemed natural to follow that model.

Moreover, my schooling, which from 5th grade on paralleled my experience with personal computers–seemed to encourage a hierarchical form of information organization. Textbooks had tables of contents that were organized by Part / Chapter / Section. We used outlines to plan arguments. School notebooks were organized by tabs (remember Trapper-Keepers?). Even something as unbiquious as TV Guide was organized hierarchcally: by Date / Time / Station. Indeed. the closest thing to “tags” that I can think of were the indexes that accompanied many nonfiction books. Entries in these indexes (which themselves were sometimes hierarchical) spanned the hierarchy of the book itself. That is, an entry for “atmosphere, of Earth” might have references in multiple parts, sections, and chapters of the book in question.

First page of the index of the book The Nine Planets by Franklyn M. Branley, the first book I ever checked out of a library.
The first page of the index of The Nine Planets by Franklyn M. Branley, the first book I ever checked out of the library, and the book that introduced me to science and astronomy,

The best I can say is that my awareness of tags probably came out of the various photo-storage apps and services that began to appear around 2004 or so, a decade after the Internet as we know it began to take shape.

Tags: In theory

These photo apps (Flickr was an early one I recall using) allowed users to “tag” photos with arbitrary keywords. This was a revelation to me. After living in a hierachical box for so long, tags introduced two ways that freed me from that box.

  1. The keywords were completely arbitrary. There was no set list. I could make them up, or choose from words that had already been used.
  2. One was not limited to a single keyword. I could tag an image with multiple keywords, placing it in multiple categories at once.

The latter point especially was important. Hierarchies, by their nature, limit where you can place something. A photo, for instance, can be placed in a folder path called /Family/Trips/Italy, or in /Buildings/Cathedrals. To put the photo in both places requires duplicating the photo file and then maintaining it in both places. Tags compliment the hierarchy, allowing one to tag a photo as “Italy” and “Cathedral” simulatenously, without having to duplicate the file.

There was another dimension to tags that I found interesting: crowd-sourcing. Because I first discovered through Internet services like Flickr, tags were often crowd-sourced, meaning anyone could apply tags to photos, thereby creating a kind of public melange of the categorization of images.

At the time, I remember thinking how useful it would be to have the ability to artitrarily tag files in a file system. MacOS introduced such a feature in OS X 10.9 in 2013. When that happened, however, I didn’t jump on it and start tagging my files.

The problem for me was that tags seem like a great idea in theory, but are much more difficult to implement in practice. I discovered this in my first concerted attempt to use tags — back when I was writing my Going Paperless series for Evernote between 2012-2016. Evernote allows you to tag notes with an arbitrary number of user-generated tags. When I began using Evernote (around 2010) I did what I suspect many users did: I started tagging everything without a whole lot of thought about it.

Tags: In practice, a.k.a., by intuition

In practice, tagging has been tricky for me: a classic case of having enough rope to hang myself. It turns out that, for me, at least, tagging comes with the same problems as hierarchical organization: if you just start arbitrarily creating tags, you create a mess of confusion that it is difficult to escape from. Escape is difficult because the process is self-perpetuating, and once started, I find I’d rather go with the flow than to start over from scratch.

What is needed for tags is some kind of taxonomy: a set of rules to follow for when to apply tags, when to create a new tag, and how tags relate to the information they are organizing. Here, however, I come up against a limitation of personal knowledge and experience: I am not a researcher, nor do I have expertise in subjects like library sciences that might provide some kind of guidance in the development of a useful taxonomy for tagging. Ultimately, I am tagging by intuition, which is probably not the best approach.

So back in my Evernote days, I began to think about practical ways in which tagging fits into my note organization, and from that thinking.

1. Tags can span hierarchies

Hiearchical organization can be useful because its very structure provides a map for locating something. I discussed the practical utility (for me) back in Episode 18, when I talked about how I use folders to quickly locate Maps of Content notes. The main limitation to hierarchies from my vantage point is that a note falls into only one path.

Tags can span hierarchies. If you imagine folders and tags as two dimensions of a grid, then the organizational model of notes would look something like this, where folder hierarchies cross horizontally, and tags (in this case, notes tagged “#jamie”) slice vertically across all folders.

an illustration of hierarchical folders show on a horizontal axis and tags shows on a vertical axis

Tags, therefore, compliment hierarchies by allowing you to create categories of notes that span multiple hierachies. In a practical sense, this allows two quick ways of finding a note or collection of notes: by either following a path in a single hierarchy, or by tag for specific collection of notes that span multiple paths in a hierarchy.

2. Tags can be used to quickly identify an arbitrary collection of notes that have a practical everyday utility

I frequently have to find a note related to a specific person. Kelly might ask, “Do you have Grace’s 4th grade report card from the spring?” Or maybe I need my son’s “School Health Entrance Form” for a camp application. One practical arbritrary collection of notes, therefore, is to tag notes by person. I do this way many notes, and these notes often span multiple paths in my folder structure.

This kind of tagging allows me to quickly locate all notes tagged “jamie”, for instance no matter where they are located, and then quickly whittle those notes down to the specific note I am looking for by adding additional search criteria. In the above illustration, the tag #jamie serves as that arbitrary collection of notes.

Similarly, I use tags to identify document types: forms, statements, correspondence, receipts, confirmations, etc. Again, the documents themselves may be spread throughout a file structure, but if I am looking for all school-related forms for my daughter, Grace, then my search would begin with notes tagged “grace” and also tagged “form” — from that result set, I can quickly locate the specific forms in question, regardless of where they are in the file system.

3. Tags can be placeholders for future ideas and concepts

I’ve also found tags useful as placeholders for future ideas and concepts. I described one of these uses in Episode 20, where I illustrated how I tag task lines in my daily notes files with the “post-idea” tag for ideas that I want to write about here on the blog. I then use the dataview plug-in to collect all of the “incomplete” tasks with the tag “post-ideas” in a query that lists the open ideas in a single place.

image showing a checkbox list of blog post ideas, rendered using the dataview plugin and filtered by the tag "post-idea"

There are also subjects that I have a broad interest in, and I’ve taken to using a tag/sub-tag model for captuing notes related to these interests. A few examples include:

  • theme/theory-of-notes
  • theme/theory-of-work
  • theme/value-of-reading
  • sports/baseball

This provides a quick way for me to collect notes together around a theme or concept that I am interested in.

Tagging problems I still struggle with

Despite these practical uses, I still feel like an amateur when it comes to tagging, and no doubt this is reflected in my tag structure in Obsidian today. Indeed, just looking at my list of tags in Obsidian makes me cringe a little. It still feels too arbitrary.

There is a time investment required to tag a note, but at the time I tag the note, it is not clear whether or not there is a clear return on that investment of time. Adding a bad tag is worse than adding no tag at all because time invested to add the tag is either wasted or, worse, does not help in locating the tagged document later on.

Another problem is that, intuitively, I feel that I should be using the fewest tags required to find what I am looking for. And yet tags, like rabbits, seem to proliferate faster than I can wrangle them in.

All of this points to a lack of taxonomy. A clearly defined taxonomy provides a scope for tagging, and more importantly, removes any abiguity from the tagging process. In other words, there is no confusion between two tags, no flitting about wondering, should I tag this note using tag A or tag B. It is always clear from the taxonomy how best to tag a note. But a taxonomy requires knowledge and experience that I don’t have, and this comes to the crux of the tagging problem that I struggle with today:

I don’t know enough about what I need to know to define a useful tagging taxonomy.

I have started to tackle this problem using sub-tags to try to identify those areas of interest I want to collect, but even this seems tenuous at best. It is an ongoing process, and the better handle I have on what goes into Obsidian, the better chance I have of coming up with a reasonably useful taxonomy in the long-run.

For now, my biggest take-aways when it comes to tagging are to ask myself the following questions when tagging notes:

  1. Does this tag help to locate the note or collection of notes that it takes?
  2. Is this tag clear enough that I will remember it in the future without much thought?
  3. Is tagging this note absolutely necessary? Can I find this note easily without a tag?

Outside of some simple, practical use cases that I’ve outlined above, I’m still wary of tags, and yet, ironically, I use them more than I should. I’ve been doing this for more than a decade with only small hints toward a useful taxonomy. Finding the taxonomy that works for me will be the information-theory equivalent of the pot of gold at the end of the rainbow.

Prev: Episode 20: Experimenting with the Dataview Plug-In.
Next: Episode 22: Daily Notes Revisited: The Best of Both Worlds?

Written on March 7, 2022.

Did you enjoy this post?
If so, consider subscribing to the blog using the form below or clicking on the button below to follow the blog. And consider telling a friend about it. Already a reader or subscriber to the blog? Thanks for reading!

Follow Jamie Todd Rubin on WordPress.com

The quantum(like) effects of e-book shelving

So this guy walks into a bookstore and goes looking for Isaac Asimov’s Caves of Steel. He doesn’t know who Asimov is, but he’s been assured by a good friend that the book is a wonderful mystery story and, being a fan of mysteries, he’ll eat it up.  He makes his way toward the  back of the store where the genre shelves are normally hidden, and finds the mystery/crime section. Once there, it shouldn’t be that difficult. After all, the author’s last name begins with an A, and the books are sorted, with passable accuracy, alphabetically.  However, he doesn’t find the book. Not completely dejected, he goes off in search of someone who can help him. The young man he finds has never heard of this Asimov fellow, but he knows how to use a computer so he looks it up. “We do have the book,” he tells our hero. “But it is not shelved under mysteries. You can find it under science fiction.”

This has been the problem with the way books have been shelves for as long as I can remember. There is no easy way to shelve a book in more than one place. Certainly, if you have five copies of a book, you might put three in science fiction and the other two under mysteries, but I have yet to see a bookstore that does this. And it makes sense that they don’t. Bookstores have trained us all that books fall under one and only one category when more intuitively, we know that is not the case.

Read more