Indexing information

I have to admit that I am struggling a bit — well a lot, actually — with the new found emphasis on linking files using applications like Hook and ‘back links’ in the numerous note taking apps that have sprung up recently.

My issue is this: I fully appreciate the utility of linking files in the context of some distinct activity — what @macsparky calls contextual computing. There is utility in being able to step easily from, say, a to do list item to the client email that caused the creation of the task, to a relevant document; indeed I am experimenting with Hook for exactly that purpose. Linking creates a graph of related pieces of information and works because you have direct knowledge of something (in my example the to do task) that provides an entry point into that graph.

What I don’t grasp is how, particularly in the context of a notes ‘database’, building a graph helps absent that contextual entry point; finding out what one already ‘knows’ about some topic, but can’t remember. Surely that’s the whole purpose of a ‘second brain’ or whatever the latest buzzword is.

Tagging can be useful, but isn’t a general solution; tags are essentially sets, but we can never have a priori knowledge of what sets we will, as opposed to presently, find useful.

Search too can be useful, but it doesn’t scale, see most search engine results. Anyway, you can search for documents that contain specific words, but really what you’re looking for is documents that relate to some specified concept and there are many document types — images, diagrams, maps etc. — that are simply not susceptible to search in the first place.

This is certainly not a new problem, a library would be pretty useless unless you could ask the librarian “do you have anything on X?” The trick must surely involve an extensible, structured set of ‘well known’ categories into which each new accession is placed.

Are there any librarians out there who can explain, at a high level, how a library actually works? And whether the approach is replicable at a personal level?



Not a librarian, but I often think about the issue @David_Roper raises. Apps like Hook can answer the question “what is this related to” only if I’ve built link(s) between this and other items. It does not answer the question “what could this be related to” by suggesting possible links.

Off the top of my head there are only two apps that can suggest relations within a corpus of notes: MarginNote 3, and TheBrain 12 (beta). In both cases, if we type a word that is the same as the title of another note in the current database, then the app will highlight the word as a possible link. In the case of MarginNote, the link is automatically active. For TheBrain 12, the user has to explicitly activate the link.

1 Like

Exactly. In your notes app, create an index note, then put links to conceptually similar notes in it. You can also write about the concept itself in that index note, your thinking on what qualifies to be linked there, etc.

Then, you can create a master index that links to your index notes.

These index notes are sometimes called higher order notes, or maps of content.

1 Like

I can tell you that in my case I build them as needed and I use Obsidian as the master nerve center for this stuff. (Although I also will build links back to the master note in obsidian from other sources using their URL scheme.)

It takes just moments to grab a link and set up when you’re already working on the project and it does pay off later.

1 Like

This. Not to bash, but even though I hopped on the Roam train in the early days, I still don’t understand the suggestion that it magically makes all of your stored knowledge/references actively available. I understand that it effectively lowers the pain points of establishing links between items, but you still have to do the linking, right? You still have to consider (whether in the moment or after the fact) what links to what…

I am open to being hearing contrary views from any Roam/like adherents— I’ve never been able to shake the feeling that I’ve missed an epiphany or aha moment, and I’m sure there’s something I’m still not fully understanding about the benefits of such systems.

It’s taken me a while, but this is pretty much how I work now, except my notes are in Drafts, and the master index is an iThoughts map. It’s a system that lacks the automatic updates of a Roam-like knowledge graph, but I like to think I gain a greater level of familiarity through manual upkeep. I’m constantly cultivating/pruning/splicing branches in my master index, and it’s not a chore; rather, it’s a valuable part of the process…


(Hi there! Longtime lurker, dipping my toes into this forum).

I’d love to learn more about how you’re using iThoughts to do this master index. I have iThoughts but haven’t felt like I’ve done the best job of utilizing it to keep track of things. I keep my notes in a massive group in DT but am struggling to keep the big picture in mind. I have been maintaining some growing index notes on key concepts (I’m an academic) but am craving perhaps more visualization? Anyway, would love to hear how you’re doing it.

1 Like

It’s really nothing complicated, and there’s no exact science to it (though I may be compelled to refine further through conversation)! Just a map imaginatively titled “overview” that I use to maintain a sense of that elusive bigger picture. Please forgive if any of the rest of this is really obvious…

I’m guessing you’re more interested in the “how” than the “what”, so in short: iThoughts allows me to author a map of linked concepts, wherein phrases or keywords are rendered as markdown links to workspaces, filters and/or search queries in Drafts.

Three things in particular that make this work for me:

  • deep links: if you’ve got a link for it, you can reference it in the map. While my note-taking tool of choice is Drafts, you could link out to Drafts, Roam, documents in MarginNote or anything that facilitates deep linking.
  • Drafts’ URL scheme, which allows for all kinds of things to be triggered from any actionable URL.
  • iThoughts’ own url scheme, which allows for creating maps from markdown or appending text to existing maps.

Slight tangent:

I’ve constructed a few Drafts actions to push notes to iThoughts maps, which is useful for creating indexes. I also started playing with ways of generating maps based on arrays of linked drafts. I haven’t yet gone further generate maps more than a level or two deep, but with bit of javascript know-how, it’d be possible to script something to parse branches of linked notes deeply. I’m not currently exploring this any further simply because the map that this kind of generative automation creates needs to be updated whenever you need to reflect updates in the source drafts, which means you can’t really make edits to the map itself— it becomes a static visualisation, and I didn’t really find that way of working to be particularly useful. Linking to a keyword search or similar dynamic filter seems to me to be the best way to work with both apps together in this way: a hand-rolled map of subjects and keywords that link to dynamic libraries of notes…


Thank you for this! Definitely not too basic and just what I was looking for.

I think this in particular is also what I’m trying to build too! It doesn’t need to be automated necessarily given that part of the work for me is the mulling over, the adapting, the pruning.

An additional thought: iThoughts’ url schemes seem to work particularly well when you store your mind-maps in its default iCloud app folder. While you can open an iThoughts document from pretty much wherever you might want to store it, urls seem to fail for other locations. Found this while experimenting with using DEVONthink as a storage provider…

Oh, I hadn’t yet noticed that! Thanks for the tip. I’ll definitely keep that in mind as the deep linking is part of what I’m hoping to do!