This note summarizes my knowledge management system both for myself and others. It uses a combination of types of notes, processes, and rules to function in a way that is robust, flexible, and easy for me to keep up with. It may seem intimidating at first glance, but this is the result of years of tweaking and iterating on my process. Try things out incrementally, and go slowly over time. As Tim ([@doubtshrine](https://doubtshrine.uk/)) has dubbed it, this is my SADALKASTEN. ## Needs and Solutions What needs do I have and how does my system solve them for me? I have the following needs: - **Recollection and details** - I need to be able to keep track of details in my life for future retrospection, problem-solving, and general biographical record keeping. - **Timestamping and tracking** - I would like to have timestamped metadata on nearly everything I touch for astrological purposes, which leads to some unique patterns in how I work. - **Knowledge management** - I need to be able to manage my ideas and knowledge. I need to be able to explore, connect, and rework ideas. - **Creation with knowledge** - I need to be able to build on top of old ideas and produce writing, content, etc. which is based on old knowledge. This may be a red herring. I'm thinking of essays, for example, where I need to go back over my old notes. I think at most I just have a list of links to them and that's it. - **Task and project management** - I need to be able to manage tasks and projects. I need to receive or find reminders. I need to be able to view my goals in a prioritized way. - **Inspiration management** - TBD, this is coming to me now, but I need a way to manage the random ideas I constantly have and capture them instead of letting them fly in and out. They're not quite projects, and often not quite just ideas, and want to be worked on a bit. These needs are solved in the following ways: - **Recollection and details** - Solved by my Journal folder, which is just a folder of Daily Notes, subdivided into years, so I have Journals/2019, Journals/2020, and so on. - I use a [[Daily Note]] template to create new Journal pages. Journals should contain a chronological log of the day, including thoughts (or links to thoughts for Zettelkasten). - **Timestamping and tracking** - Solved mostly by my Daily Note by including timestamps and faithfully linking to a note any time I create it. It's too much work to rigorously timestamp every single link I make across my knowledge management, and adding timestamps everywhere will make it tedious to traverse. Keeping timestamps and a log manually in the daily note is a decent compromise, because dates can be found in the future via backlinks. - A different solution such as a daily LLM-generated diff summary may be a better way to go in the future, stored as raw text somewhere to be looked at later. - **Knowledge management** - I am learning how to think in a modified [Zettelkasten method](https://zettelkasten.de/posts/overview/) focusing on capturing atomic thoughts which can be reused, linked, and contextualized from different levels of abstraction of thought or different needs. This is accomplished by learning how to use links, tags, full-text search, and other organizational tools. - **Task and project management** - Tasks plugin and Projects folder with a dataview - I use the [[Project]] template for this, the Tasks plugin dataview for tasks within the project folder or project tag, and Obsidian's nested tag to define project-specific tags. ## Zettelkasten principles Here are some ways I handle my graph. I don't strictly follow Zettelkasten, as I find some of its commandments to be overly rigid, and the drawbacks they were designed to solve can be handled with digital interfaces that have full text search and tag support. - **Hierarchies of abstractions, not hierarchies of ideas** - [[Obsidian]] - [[Zettelkasten]] -> [[Using Zettelkasten with Obsidian]] **Folders are useful** While Zettelkasten advocates that you don't use folders for category, some groups of ideas need clear edges, such as project materials. For me, this is things like content for books, written text for essays, or notes for software projects which may link liberally to the nodes in my main zettelkasten folder. Another good example is transcripts I've created from lectures which I want to keep as separate content for quick lookup and reference. Of course, this isn't strictly necessary, as I could manage this with tags or spines but I think having some partitioning helps a lot. For files that are only useful in the context of a certain window of time, such as a note containing hotel information for a vacation, I might just throw it into the Journals folder, the only area of my vault graph with a temporally ordered structure. To help me find it, I'll link it to the appropriate day or trip planning node and I might tag it \#document or similar. It doesn't really matter where it is in terms of the folder structure, because I can always find it via links, tags, or full text search. **Self vs Knowledge (Folders are useful continued)** I have two main folders, Self and Knowledge. Knowledge lives Out There. It's ideas which exist independent of my relating to them, and can be stored as information, structured, presented, and so on. "Self" is all my personal notes, thoughts, ideas rooted in time. It's for anything I might want to work on or create, or use as scratchpads. It's anything I might want to keep around for reference but which I'm not likely going to want to refactor into a polished note of cleanly structured and connected ideas like I would in Knowledge. **Timestamps sometimes** I use timestamps for one-off ideas I'm not likely to build on or have many links to, and don't use them for more singular ideas or central nodes which will have many links or that I will turn into structures/more detailed ideas. If a one-off turns into a bigger/core node, I just remove the timestamp. I also use timestamps when an idea seems like its moment in time is important, like a note about a personal life event. **Dates are more useful than I thought** - They help tie together related thoughts in memory for me. Thoughts existing in space without the time component drift and are intangible, and even create a form of mental stress to just have floating, as if being asked to be dealt with or encountered. But put a date on something and you can say: I'm not worried about that, that's in the past. I might be annoyed with it encountering it later in a rawer form, but the key is that I can know what the *context* was that I can recover, and often the context is that this is so old I don't care anymore. **References are knowledge in progress** I've struggled a lot with reference material versus "knowledge" or personally worked concepts. Some people advocate for reference to be completely separate from their zettelkasten, but I want my references close by. References also create useful central points to jump off into old ideas, but as this happens they also tend to accumulate I don't want to have my reference material separated in a different application, and in my graph reference notes that simply refer to links quickly get overrun with the need for personal thoughts and TODO: figure out reference system vs types - a reference is like, a single link/summary for instance - or a larger collected page of links - I don't care which is which, since So a Link is a thing. A PDF is a thing. A URL is a thing. In Obsidian I have this crappy workaround to make a note PER file. Which I don't like so much. But I think that I *should* be able to append something with metadata in my own files. **Leave yourself a red string wherever you go** I try to design my links in such a way that I'll run back into TODO: come up with some concrete examples - I know that there are some projects I want to tackle, such as an essay on astrology. I create a node for the essay, and then go into my old notes and tweets with that essay in mind, looking for ideas to relate to it. I'll find lots of interesting stuff, and as I go I throw these into my graph as nodes. Chances are a lot if it won't even relate to the essay, but keeping the essay in mind as I go gives me the motivation and framing to keep going. - Another example is, I have tons of astrology essay ideas. I'm not sure which essay ideas I actually want to make. I *think* I need a spine node to collect a bunch of my ideas scattered all over. So I go searching for "essay" "astrology" in my notes and other similar keywords, and as I find them I link them to the spine node. I might not take the time to do any organizing or turning those unorganized notes into atomic thoughts; just having the link is enough for me to come by later the next time I'm exploring this spine. If it turns out that I don't actually think this node is important enough to be a spine, I can just demote it to a regular old list/topic note I can come across again whenever. Chances are, though, that this "List of astrology essays" idea might also be linked to a "Goals - Writing" or similar. (<-- tangent) Note on a way to mark things as undigested #wip/stub, #zk/todo, and so on **If it's not linked, it probably doesn't have meaning** - it will be harder to contextualize or may get lost An example of this is when creating new notes but I need to link those back to my daily note so I can remember where the hell I was--and I have a chronological record of some kind. Ideally you'd be able to actually have a live query of Nodes embedded, but instead in Obsidian I must link, and then copy them to "commit" the query and actually have them show up in the Graph. **Avoid busywork** Avoid tasks that feel nice and tidy but don't actually solve the problem or contribute to clear thinking. For me, one task is creating an index of every year and month. I wanted to do this because I thought it would look really cool in my graph, but truthfully, it wouldn't actually help me find new information because I don't recall most information from a date alone. It would look very pretty, but it wouldn't help me find anything or make new connections any easier. **Links should encode non-redundant meaning** You might link 2023-01-01 to 2023-01-02 and so on, but does that actually help you? You already *know* that the 2nd follows the 1st. If you want to get to the next one, it's just Ctrl + O and then type the next date. ... more examples This is turning into a blog post. I think I need a condensed version for myself at some point. - Good tutorial to attach for using Obsidian: [How to use Obsidian Notes: a step-by-step guide](https://elizabethbutlermd.com/obsidian-notes/) ## Project tracking - Use tasks and then filters to scope down by folder - Use top-level Project pages and Worklogs in each to keep track of what I did so I can catch up on my own notes later ## Key Plugins - Dataview - Folder Note - Paste image rename - Tasks - Templater - Omnisearch / Text Extractor? need to evaluate ## Current Gaps What am I missing? What doesn't work well? - Projects fall through the cracks. I don't check [[Projects and Tasks]] because it's overwhelming to look at. Need to figure out better filters or how to segment it. - [[202309051454 How do I handle tasks]] - Massive amount of old notes that need to be reorganized in general. - Filesize - There are images, PNGs, that I should convert to JPGs. - Internalizing links, references, studies, and so on. I need to ask ChatGPT about this. My current strategy is to just add #archive to something - Todo: look into [GitHub - ArchiveBox/ArchiveBox: 🗃 Open source self-hosted web archiving. Takes URLs/browser history/bookmarks/Pocket/Pinboard/etc., saves HTML, JS, PDFs, media, and more...](https://github.com/ArchiveBox/ArchiveBox) - It's probably fine to just keep an archive folder myself and know that I'll be ZKing it a bunch anyway. Old articles aren't useful to me as pristine, untouched documents; I need to link! ## Things to Try - Can I give a whole messy note to ChatGPT? Will it restructure it according to the principles of ZK? - It sure can! [[202309071722 LLM prompt to refactor notes according to Zettelkasten principles]] ## Misc. thoughts on benefits of this system - Because everything is directly timestamped and atomized, I can speak in present tense without worrying how future me will interpret my tenses in terms of what I was referring to or at what level of development of knowledge I had at the time ## Other Sources - [[Hypertext, Personal Knowledge Management, Thinking, Second Brain Software]]