Zettelkasten Forum


Why not just use Zettels in place of tags?

I am working on my own note taking tool and try to keep the amount of primitives in the tool to a minimum to make the primitives it does have as broadly applicable and reusable as possible. So when thinking about the structure of my Zettels I wondered why is there a need for tags? I get their purpose as grouping multiple Zettels of the same subject together, but are they not just a reverse index, where the Zettels of the index point to the index Zettel rather than the index Zettel listing them all? Replacing tags with Zettels also naturally gives you a place to describe what the "tag" is about. I get that [[202004152130]] might be less readable than say #zettelkasten, but if that is the only reason, in my opinion, that then would apply to all links in general and could be improved by the tooling.

Have others thought of doing this? Would there be any problems with this approach?

Comments

  • edited April 15

    I‘m also developing my own Mac app for knowledge management and, some years ago, I‘ve come to a similar conclusion. All knowledge elements (and tags are in fact knowledge elements!) should be fully featured plaintext notes.

    This gives you the advantage that you can add comments to a "tag note". Additionally, and maybe even more importantly, you can add tags to a tag! This allows you to build tag hierarchies where a tag can have multiple parents. This, in turn, allows for powerful tagging features.

    Roam does something similar in that tags are also notes. But in fact everything should be a proper note. This doesn’t mean that the tag couldn’t be displayed in the note as #tag (or, for my app, which uses MultiMarkdown: [@tag]). The software could still resolve this to the corresponding "tag note".

  • edited April 15

    @Vinho I thought that discussion ended up being mostly between when to use tags and the use of structure Zettels as an alternative for using tags, which differs from what I am proposing, which is to use Zettels as tags. Interesting read though!

    @msteffens Nice to see you came to a similar conclusion! The point you make about its advantages are exactly what made me think of it, and what I meant with my first sentence, i.e. by making use of just a few general primitives that are broadly applicable, they generally bring with them added flexibility. Also, all effort put in improving links will also benefit tags and vice versa. For example the tag hierarchies are also applicable to Zettels in general, where you might want to retrieve certain hierarchies of Zettels you've constructed.

  • The example you provide, #zettelkasten vs [[202004152130]], makes sense in the context of existing tools that are used around here, but is very loaded.

    As literal phrases, [[202004152130]] and #zettelkasten are both just a string of characters in text files; the latter is a weird variant of a German word, so I can read it, while the former is a bunch of numbers, that I interpret as a date-time stamp thanks to many years of practice only. But on their own, they "are" nothing special. You seem to conflate what is click-able in apps with the underlying representation: text.

    So is the question "which parts of a text document should be click-able"?

    E.g. we could've decided to not make hashtags click-able in The Archive. People who just love to denote tags with the hashtag convention thanks to years of Twitter would then have to use the wiki-link feature and write [[#zettelkasten]] to make tags click-able. This ends up doing the exact same thing in The Archive, by the way, and that's by design. You can also create a note #zettelkasten.txt and maintain your annotated tag index. The fact that hashtags are click-able is just an affordance. In that regard, they are nothing special. It all boils down to plain text search expressions. If I have a plain text folder of @msteffens notes, I'd simply search in the directory for @tag to effectively get the set of all "tagged" notes, no matter if there's also a @tag.txt file next to it.

    I think we're on the same page here about making tags nothing special: not some kind of magic meta-data to files that is only useful in one particular app. If all boils down to text, you can do a lot more with it.

    Author at Zettelkasten.de • https://christiantietze.de/

  • I'm not entirely sure if the question is about functionality of your tool or a workflow? I try to answer both, I guess.

    If your tool has a search function, then you have tags as well. Hence, functional amount of your primitives is not increased.

    There is an obvious benefit to using tags, at least in my workflow. I can create a note, immediately assign a tag to it, and be done with it. I will sleep well, knowing the note won't be lost.

    This method is especially useful when you're creating notes outside of your main application.
    E.g. I use The Archive on my computer, and Drafts on my phone. I can quickly jot an idea in Drafts and send it to my Zettelkasten archive. Alternatively, I would first have to create a note, and then enter this note to the Index note - one step too many to complete it on a go.

    Another benefit is that you can generate the index of tags automatically (via script in my case). Without tags you could only create you index notes manually.

    And the last point. Tags and index Zettels are not mutually exclusive. I use both. My tags describe more general overarching themes; while index Zettels contain smaller localised subjects and "projects".

    I hope this helps.

  • @ctietze Yes, the examples I used were tailored for the use in this forum, they were only meant as illustrations. I have to admit that I was not thinking in terms of just text. After seeing some of the informative videos on your Youtube channel, I know now that you talk about the software agnostic approach. And in that context I see how they are nothing special, even though software might make them easier to use, they still remain just text like anything else and I see the value in that.

    @karoshiya I guess a bit of both, the tool can help with the workflow after all. I can see the benefit of them when you are outside your Zettelkasten and looking up the IDs for linking would be a pain, good point! I have to disagree on automating the index, there is no reason why I could not do the same for searching all links to the tag note and generate a list of links like you would for tags.

    There are multiple aspects to tags and I think it is useful to make them clear for the discussion, because they are entangled at the moment. Tags can be used as a means for classifications, similar to categories, for which good arguments against them have been made. So even though tags might be plain text, they still denote a concept which is generally used for means of classification. For this aspect of tags I plan to use links, but as discussed here and talked about in various other posts, I do not plan to use them without any additional explanations, as implicit meaning associated with tags might change over time.

    The other aspect of tags mentioned in this post is their use as keywords. They can be useful to increase the ability to find notes through search mechanisms, and, as both of you mention, they are just text that helps with search expressions. There will be times when the contents of a note will not have the best wording for the note to be found through search, so I definitely see their value.

    When talking about tooling, it could be useful to have all the keywords used in the linked notes listed such that the connectivity of links between notes becomes easier to search for with just words.

  • @grayen thanks for this discussion. It triggered the above lingering question I had about the "Use Selection for Find" functionality. Best of luck with your development efforts.

  • @grayen said:
    @Vinho I thought that discussion ended up being mostly between when to use tags and the use of structure Zettels as an alternative for using tags, which differs from what I am proposing, which is to use Zettels as tags. Interesting read though!

    Hm, I thought we were talking about the same thing. When I have a structure note instead of a tag, I would "tag" notes with links to that structure note. Am I missing something here?

  • @Vinho I am new to this, so I might just have misunderstood some of the terms being used on the forum. I thought that structure notes are like index notes in that they show a particular organization of those notes listed in them, i.e. a particular way to structure them. So they are inherently organized, while I thought that one of the things that tags give you is that they are an unstructured grouping of notes.

    So this would be an index or structure note:

    [[note A]]
    [[note C]]
    [[note B]]
    

    So you get them organized in the particular order of A, C, B. While what I suggested was with links as tags, you get for note A, B, and C:

    [[tag note]]
    

    Given no particular structure to the links to the tag note, that might be dependent on how the tool shows the search results, like is the case for normal tags.

    Or have I been wrong to presume index and structure notes to be the same? Are structure notes maybe exactly what I meant with the above? In that case I find the naming not very intuitive, since I consider it to have less structure.

  • @msteffens wrote:

    I‘m also developing my own Mac app for knowledge management....

    NO NO NO NO please not another Mac-only notetaking/ZK app there are already so many well it's your life who cares what I think....darn TheArchiveScrivnerDevonThinkTinderboxeDrafts.app.etc.etc.etc.

    Ahem.

    Re: TAGS - tags can be used in ways that duplicate some of the functions of categories and links. Based on personal experience and significant anecdotal input from people doing many kind of tasks, my suspicion is that for many people tags are best used as a way to indicate stage of workflow/processing (if that's something you track), such as "to process", "waiting for book to be delivered", "notes completed", etc. This would be because:

    • if you track workflow, you must track and update every item in the workflow, and tags are probably the fastest type of indicator to and to notes and revise
    • workflow is something that is conceptually "skew" to note content and placement in the ZK structure (not sure if that's clear - I mean a note's placement in the ZK structure and content have no relationship to workflow stage) and tags seem to work well in capturing this other domain of categorization of notes without getting in the way when you are working with note content and note placement
  • @cobblepot wrote:

    NO NO NO NO please not another Mac-only notetaking/ZK app there are already so many

    I can understand that sentiment. However, app development started a long time ago (before The Archive existed), and its design and feature set is still pretty unique. If you're interested, see this and this post for more info.

  • @msteffens said:
    I can understand that sentiment. However, app development started a long time ago (before The Archive existed), and its design and feature set is still pretty unique. If you're interested, see this and this post for more info.

    The website and app look very nice! Out of curiosity, is it different from Zotero + Zotfile?

  • @cobblepot wrote:

    is it different from Zotero + Zotfile?

    It's hard to compare with these apps since IMO they have different workflows & goals. With my app, you wouldn't export but rather work with the highlight notes inside the app, link them, tag, rate or comment on them, etc. In the app, all elements are hot linked and clickable, link connections can be visualized, and the PDF source location is displayed next to the highlight note (if you prefer).

    The latest status report in the forums has some screenshots and a 1-minute screencast that shows it in action.

  • @cobblepot You make a good point, that would be another aspect of tags, the marking aspect of tags. Although links could be used for that in theory, I think they are a poor fit for that aspect of tags, the link would not be meaningful, its title maybe, but then you are probably better of using a text label as a marker.

    Haha, I totally get your feeling about the many Mac apps out there. I am mostly using Linux (NixOS), so to be clear, when @msteffens said "also" it was no reference to the app I am developing.

    I am a bit ambitious with my app, so I realized I am better of with first going with a lightweight alternative that fits my basic needs, so at the moment I am developing a VS Code extension. It already implements most the features I could find in existing Zettelkasten VS Code extensions and I plan to extend with some nice tree views representing the network of my Zettels later today. The reason why I am going for this lightweight alternative first is that it is not so much a believe as I know this to be the case for me, namely that too much focus on the software and technology gets in the way of achieving the original goal, getting into building up a knowledge base, not just its foundations, like @ctietze mentions in his videos about technology.

    @msteffens Nice app you are building! I could learn a thing or two seeing how you deal with annotating documents. I did get to the point with one of my previous attempts at building my ideal knowledge base app were I too allowed annotating documents, mine were just limited to highlighting text and coupling a note to it (not a simplified one, but one like any other note in my system), and making an easy link between the two.

    My app focuses on combining the capabilities needed for Personal Information Management (PIM) and related concepts (Tasks, Calendar, Contacts, Movies, Books, Bookmarks, etc.) and note taking together, because I firmly believe there is a huge overlap between all these and the ability to interconnect between them and being able to leverage the same tooling is very powerful. It is implemented with a lightweight markup language that is balancing ease of writing while still denoting very generalized data structures. The goal is to make it very easy to interact with the data denoted in the notes, making a lot of automation possible. I will probably write more about this in the future.

  • @grayen said:
    @Vinho I am new to this, so I might just have misunderstood some of the terms being used on the forum. I thought that structure notes are like index notes in that they show a particular organization of those notes listed in them, i.e. a particular way to structure them. So they are inherently organized, while I thought that one of the things that tags give you is that they are an unstructured grouping of notes.

    So this would be an index or structure note:

    [[note A]]
    [[note C]]
    [[note B]]
    

    So you get them organized in the particular order of A, C, B. While what I suggested was with links as tags, you get for note A, B, and C:

    [[tag note]]
    

    Given no particular structure to the links to the tag note, that might be dependent on how the tool shows the search results, like is the case for normal tags.

    Or have I been wrong to presume index and structure notes to be the same? Are structure notes maybe exactly what I meant with the above? In that case I find the naming not very intuitive, since I consider it to have less structure.

    You don't seem to have misunderstood anything – I think I'm the one who has. Still not sure whether I fully understand it, though ;) I guess after clicking on [[tag note]] you would want all the notes A, B and C to be displayed in the search result – and no other notes. That would mean that the text in the link (in your example "tag note") would have to be unique to the three identical links in the three notes. Is that right?

  • edited April 18

    @Vinho Not quite what I meant. The texts note A and tag note were meant to be UIDs like 202004152130, but I thought this to be more descriptive. Index note versus tag note are opposites approaches with the same goal, but useful for different use cases. Index notes give structure by creating outgoing links within it, tag notes give structure by creating incoming links to it from other notes. So yes, for a tag note to give a similar structure as usual tags give, it would require you to be pointing to the same UID everywhere, but that is to be expected.

    I guess after clicking on [[tag note]] you would want all the notes A, B and C to be displayed in the search result – and no other notes.

    I consider this to be a question about software and as such not that important. What and how information is encoded within a note should be the focus, if you ask me. However my expected behavior would be that you click on it like any other note link, and then you can see a list of all the notes by just looking at the back links (i.e. incoming links) of the tag note. No need for any special support.

  • @msteffens said:
    @cobblepot wrote:

    is it different from Zotero + Zotfile?

    It's hard to compare with these apps since IMO they have different workflows & goals. With my app, you wouldn't export but rather work with the highlight notes inside the app, link them, tag, rate or comment on them, etc. In the app, all elements are hot linked and clickable, link connections can be visualized, and the PDF source location is displayed next to the highlight note (if you prefer)

    Just FYI, Zotero + Zotfile do have similar abilities - one of the main reasons I like them. When you highlight and comment a pdf and then use Zotfile to 'extract annotations' the resulting Markdown plain-text note puts your comments and the quotes in different styles (italics/normal) and puts a link next to each extraction that opens the pdf file to the exact location of that highlight or comment using the zotero URL scheme, such as:

    zotero://open-pdf/library/items/QNBU74VN?page=4

    Zotfile does not, however, allow you to visualize link structure. In any case, this info may be useful to Windows users!

Sign In or Register to comment.