Just because I advocate for neatness doesn't mean I think you should keep everything forever.
I'm just reconfiguring my mailing list and came across this screen in the excellent (and FOSS) Listmonk mailing list software.
These are the templates that you use when you send an outbound email. I tinkered with this back in July and I know that I always want to use one of these templates.
Hmm. But which one was it?
Less choice is less friction
It's a common theme around here: less choice is to be embraced. Design fewer areas and categories, not more.
Is this the first time I've mentioned Barry Schwartz' Paradox of Choice?! It can't be but it seems so. That link was the TED talk; it's also a book and is worth a read.
This isn't quite the paradox of choice, but the idea is the same: why have all of these choices? Are they good? Helpful? No! In this case they're just confusing. I don't remember which one I want. They slow me down.
I think we keep things around just in case. What if I wanted to use one of these templates later? Is there some setting in there that I couldn't recreate? Something I might look at and learn from?
Eh, there might be. But probably not. Life would be better -- a tiny bit, but surely -- if I deleted the four templates I didn't use. So I just did.
It might be my fault, but there can be confusion between what's your Johnny.Decimal system and, for those of you who are into 'personal knowledge management', what's your PKM system.
What's PKM?
If you're not into PKM, here's a quick primer. People enjoy storing 'personal knowledge': facts and thoughts and notes about whatever it is that they're doing or learning or anything.
Then they like thinking about exactly how to store this personal knowledge, because if you just chuck a bunch of notes in a folder you're unlikely to have a good time. That just became personal knowledge management, or PKM.
It's massive. Some significant percentage of YouTube is taken up by PKM. Discords and forums abound. Creators-a-plenty worship the god that is PKM.
What's not PKM?
I've never framed Johnny.Decimal as PKM; ref. that post I linked at the start. This is a conscious decision: for those who aren't into all that, and just want a way to neatly organise their files, I don't want it to feel like you have to join a cult just to be organised. You don't. You can keep it simple, be organised, and play outside.
But of course there isn't a clean break between these things. The Venn diagram of JD×PKM would look like a solar eclipse. So how do the two intersect?
JD is your PKM
The misconception is that you have this Johnny.Decimal system -- say, Life Admin -- and you have a PKM system. That might look like this.
All I mean when I say that JD is your PKM is that every piece of personal knowledge that you would like to manage must be linked to an ID in your system. In our highly contrived, too-neat-to-be-real example, it looks like this.
22.00.0137B. Right-click and open in new tab to embiggen.
Nothing doesn't have an ID
That's all it is. If you're a good Decimal there should be no notion of a thing in your life that you can't fit into one of your IDs. That applies to personal knowledge in exactly the same what that it applies to your electricity bill or the confirmation of the flight you just booked.
And if you don't have an ID? Create one. You might need to create a new area or category to hold it. Do that.
'How do I PKM?' is answered
So the question of how to organise one's personal knowledge is that you organise it just like anything else. In a neat system, where nothing is buried too deep, where like concepts live near each other.
I'm recording a new video course for JDU, working title Task & Project Management using Johnny.Decimal. I'm tackling (again!) the age-old problem of 'getting stuff done'.
(No, I don't think it's a solved problem. Yes, I have my own ideas. They're really helping me.)
In doing so, I've realised that there's a class of task which seems to go un-done. It's the stuff that isn't truly important, otherwise you'd do it, because you just have to. You just have to pay your credit card, so you find the time.
It's the next level down. Stuff that you want to do, and that if you do manage to do, would make you feel like a better human. So let's just allocate some time to it, and decide: be a better human.
I do this by blocking out an hour of my calendar every day.
This morning: boring documentation for my business
I swear the point of this post wasn't to promote these links, but it's too good an example not to use.
I have a Ko-fi and a Patreon that are scarcely used -- I'd rather you buy a useful product -- but that do generate some income. (If that's you, thanks!)
Each of those connects to one of two PayPal accounts, and/or one of two Stripe accounts! And each of those accounts has settings for when and if it automatically pays out to a bank account, and if so, which one.
So when Lucy, who keeps an eye on our finances, asks whether we have any spare Patreon money and how she might get it, and yet again I scrunch up my face and say … ¯\_(ツ)_/¯ … then, yeah, it's time to trawl those sites and just write it down.
Oh my god that sounds boring!
RIGHT? Riiiiiight. Really really tedious. And so I look at that task and I think, ha!, like I'm doing you today. And so months go by, and every time Lucy asks me, I still just dunno.
It's kinda pathetic.
So. That's what I did this morning. I took about 45 minutes and I went through all of those sites and at 13.22 Accounts that deliver income and 32.13 Stripe in my SBS I wrote a bunch of notes that link to each other that explain the situation.
And I feel like a better human for having done so.
Maybe it's because 'I'm a visual learner' (warning: not actual science), but I find this sort of thing much easier to approach if I crack open a mindmap. My love for MindNode is well known, but use whatever you have.
Turn words in to nodes
My first step is to just read through the question and pull out the key parts. While you're here you might as well loosely organise in to concepts.
It's important to realise that you're not creating some work of art here. Often enough I don't even save these files: I just use them as a scratchpad for my mind. So on the left, I've extracted what I thought were the key parts of the original question.
22.00.0135A. Right-click and 'open in new tab' for full size.
Create links
The point of this is to link these concepts in my mind, and to convey them to others. I was trying to type out a response on Discord and wished I was on camera, waving my arms around. 'This links to that', I'd be saying.
That's when I realised the mindmap was the answer. So on the right I've extracted key parts of the SBS, and I've linked the two maps together (Cmd+Shift+L in MindNode) and added some comments to the links.
Brings an idea to life; breaks it up
For me, this immediately brings the idea to life. And now we have a simple diagram that we can talk to, and specific parts that we can focus on. Rather than one mega-question, now we have a bunch of sub-questions. Much more manageable.
Extend-the-end (EtE) is a new pattern that is still finding its feet in the real world. Here's a deep-dive and discussion of the use cases that I see on the forum & Discord.
Pattern 1: 'lifting up' a header
You have some ID, and in that ID you can hold a whole bunch of information. Let's say you keep one long JDex entry with neat headers.
The simplest way to think about EtE is to imagine 'lifting up' each of those headers into its own document. Because eventually that single note is going to get really long.
So this:
# 11.11 Some note
## Header about thing A
Yada yada blah blah for 10,000 words…
## Header about thing B
Blah blah yada yada for 15,000 words…
– is going to become 3 notes. We keep the parent
11.11 Some note, and introduce
11.11+ Header about thing A and
11.11+ Header about thing B.
Format: AC.ID+▁Title of the thing
Note the specific format here, which is in contrast to pattern 2. The EtE'd note's title is:
The ID of its parent 11.11, then
The +, then
A space , then
The note's title.
This recognises the general broadening of IDs
A broad pattern (see YouTube 'What is an ID?') is that IDs are getting broader, and holding more.
I'm not really sure why this is, to be honest. I can't tell you the last time I got anywhere near filling up a category with IDs (our pre-built systems excepted; they were designed to be full).
So exercise caution: if you do think about 'lifting up' a header like this, ask yourself, should I instead just be creating a new ID?
Concerns
It can be easy to forget that you've created these 'sub-notes'. Think carefully before creating one: prefer to 'fold' the header in the existing note.1
Leave yourself a breadcrumb in the parent note. I always do this at the top.
# 11.11 Some ID
- Sub-notes:
- 11.11+ Header about thing A
- 11.11+ Header about thing B
---
Main note contents here.
If your notes app supports it, make these clickable [[wiki-links]].
Example
At the Small Business System's 14.32 External software & services I've created a note each for the services that I use. In this case I was experimenting with linking to these notes from other notes; easier achieved when the item being linked to is itself a note vs. a header in another note.
Now when I link to one of these items, it reads almost like a sentence. Here's an extract from a note:
curium does a full nightly backup to [[14.32+ Backblaze]].
(Curium in this case being the name of one of my servers.)
Folders in your filesystem
I'm not totally sold on the necessity of this, but there doesn't seem to be a downside.2 Try it and let me know what you think?
If I have an associated filesystem subfolder for an EtE item, I name it with a +. This serves as a reminder that this item is an EtE.
Pattern 2: use it as a short code
I use Life Admin (LAS) as my personal system. So it already has all of the IDs that I need, day-to-day.
I manage my own data storage & backups at 14.14 Data storage & backups. But I also manage this for my bestie Michelle, a photographer. She has her own setup and I don't want that mixed in with my own.
In a 'normal' system, i.e. not the pre-built LAS, I'd probably just create her own ID:
14.83 Michelle's data storage & backups
– but I am in LAS, and I already have a place, and the system is pretty full. Also, this isn't the only piece of data I'm storing for her: I also manage her domains, and who knows what else in the future.
So here's what I've created:
14.14+MGH Data storage & backups
14.25+MGH Domains & hosting
+MGH = a reusable short code
Being her initials, in this case. And note the subtly different naming pattern here: no space after the +. Because this enables searching your JDex for this specific string, revealing all of the stuff that is exclusively related to Mishy.
More importantly it conveys to you the fact that this is a reusable short code.
Using this to manage the family
A very common question in the context of LAS is how do I use the system to manage my family?
The simplest pattern is to use EtE as just described. Assign your husband and each of the kids a short code. Obviously this code has to be the same everywhere. Initials are great, but use whatever works for you. I'd probably keep it to <5 characters, but really that's arbitrary. There's nothing fundamentally wrong with +BARTHOLOMAEUS but I think you'll get sick of typing it.
(Actually there is a sound logic to shorter codes: they're harder to get wrong. It'd be easy enough to type +BARTHOLOMEAUS one time and now you're at risk of losing that note.)
Think of this like a 'shadow system'
In this situation, you have one primary system. All IDs are created in this primary system.
But underneath that, you have any number of 'shadow systems'. They're the same system; same structure, so you don't have to remember a whole bunch of other IDs. And you reveal it by searching for your short code.
Alternative (not EtE)
For the record, the other way you might do this is by completely duplicating the LAS structure for each of the kids, or whomever.
10-19 Life admin ← your stuff
…
20-29 Michelle's life admin
24 Her online life
24.14 Data storage & backups
…
30-39 Bartholomaeus' life admin
…
This feels like you'll have a lot of empty space. Personally I prefer the EtE pattern.
Date as short code
As usual, if the date is at all relevant you should probably use it.
We're currently on a trip. I'm tracking that at my personal system's 15.53 Jucy's World Tour.
For each new segment of the trip, I create a sub-note starting with the date.
15.53+2025-09-19
Double-chain short codes
Actually, I go further. The 19th September trip was to Japan.
15.53+2025-09-19+JP
I've used the date first, as it'll force the note to sort. Then I've added +JP.
This assumes that I might find +JP a useful filter on my system. Otherwise I might as well have named it 15.53+2025-09-19 Japan. And let's look at that carefully: the ID of this note is now 15.53+2025-09-19, and the title of the note is Japan.
In this case that probably would have worked; but countries already have 2-letter code that I can use. And that feels neater to me, personally. But I'm kinda in to ISO codes which might be a niche thing?
Ac-tually before even finishing this update I realised that I do find this short code useful. I stored a QR code for Japanese immigration there and knew that I'd want to find it real fast when I was in the immigration line.
Knowing that +JP will surface just this note, and no others, was a comfort. And that's what I did, finding the QR code while everyone else was scrambling for theirs.
Pattern 3: use it to store notes 'outside' your main system
Credit for this one to @salutalice on Discord, who wanted a way to keep their EtE notes 'out of the way' of their primary JDex notes.
This leaves your JDex with its AC.ID notes clean. Every note there is a primary, ID-defining JDex entry.
So that's what they do. In their case, using a Bear tag that places their EtE'd notes out of the primary system. Similarly but kinda opposite, @aviskase tags their primary JDex notes with an Obsidian property, allowing them to show only these notes. Same idea, different mechanism.
Exactly how you do this will depend on the app you use for your JDex. Ask if you need help -- channel #13-system-expansion is your best bet.
I don't do this myself yet but I can see myself migrating to this pattern. I really like it.
The concern above re: forgetting that these notes exist is doubly relevant here.
Retraction: the 'blog pattern' was wrong
The EtE page lists as an example this blog. E.g. the page you're reading would be 22.00+0133.
I'm going to remove this example and instead number blog posts using the 5-digit creative pattern. That's a future post.
100% human. 0% AI. Always.
Footnotes
Any good notes app will let you 'fold' or 'collapse' a header. Typically hovering over the header will show a reveal/collapse triangle ▶ ▼. ↩
This is a common question that I just received by email. The author has a personal and a work system -- let's say they're Life Admin and Small Business, but that isn't important -- and wants to know how the two should interact.
The key question is whether these systems should be merged and/or interact in some way, or if they should remain divided. For example should you move Life Admin's 10-19 under your business system and re-number it? Should these systems share a JDex?
(An assumption is that they're on the same computer, like they are for me. I don't have a separate work machine; if you do, this is simpler.)
I split my systems
I have P76 Johnny's personal life and D25 Johnny.Decimal. To each, it is like the other does not exist.
They're different parts of my life and I find the separation helpful. Also, for example, I share the business with Lucy, but not my personal stuff. As soon as you merge them, you lose the ability to do that sort of thing.
I keep my personal JDex in Bear and work in Obsidian. Again, because Obsidian is shared and Bear can't do that. This works for me: when I'm 'at work', I'm at work. It's rare that I find myself wishing I could, say, link a note from my work system over to my personal system.
And when this does happen, I just use the full SYS.AC.ID identifier to tell myself this is what I mean. It just happened this morning: in a note in Obsidian I wanted to refer to a personal thing so I just typed out 'blah blah P76.15.54 blah blah' and that's all I need.
No renaming necessary
Our emailer had asked whether it was necessary to rename any of the systems' subfolders to indicate the system that they belong to.
No. I have two top-level folders which mirror the system names, then the folder structure underneath is as-downloaded from JDHQ. No renaming of folders.
I do name files
I find it useful to name files with SYS.AC.ID at the start. This takes no extra time and now I can quickly see to which system they belong.
A century ago, you were an old-timer. You made stuff in your workshop.
It took a while to set up a new job. You had to figure out which tools you need. Maybe sharpen your plane, oil your strop. Clear the bench, sweep. Check that you have enough birch, go to the store, pick up some more two-by-four and a bag of nails. Lay it all out. Have a look at it. Is the plan good? Do we know what we're making?
And now you can start.
Later, the new table finished, this all needs to be tidied away. Ready for the next job. You look over your shoulder as you lock the workshop. Neat. Calm.
You can't wait to work again tomorrow.
Is this how you feel?
Computers allow us to do any number of things 'at once'. Of course you can only ever do one thing at a time, but it's nice to pretend.
When was the last time you prepared the next piece of work? Closed all of your old windows. Opened up a new view on your files and found the place you needed to be. Opened the note you'll be using in a new, clean window. Focused your task manager on the task at hand.
Your computer is your workshop. If you want to work well, you need to slow down. You need to think more, and do less.
What you do will be better, and you will enjoy it more.
Tip: use your task manager
Here's what I've started doing. I use Things but any task manager will work.
Create a new window. In Things, File > New Things Window or Ctrl+Cmd+N. Move it up in to the corner, and make it small. Think of it like a tiny dashboard.
Highlight the single thing you're working on now. Here, I'm using a tag 🫵🏼 focus on the task, then viewing that tag in this window by typing focus.
That's it. Remember to be doing what this window says. Don't do anything else until you've finished, and cleaned up, and prepared the next job. Simple.
More on this in the upcoming JDU course
I'm figuring this out as part of the next JDU course: Task & Project Management using the Johnny.Decimal system. Sign up to the mailing list for updates.
When I'm working on project-level stuff (which is larger than task-level stuff; see upcoming JDU course) I typically (~always) have a task in Things, serving as the reminder, and notes in Obsidian, because it's a notes app and Things isn't.
I always cross-link the two. When you're in one place, you should make it obvious to yourself that there's relevant stuff in another place; otherwise you have notes over there that you'll forget to look at. Links make it trivial to switch between these two places. They're worth the ~minute it takes to set up.
Deep-linking to an Obsidian header
This didn't work with the native Obsidian URL scheme, but the Advanced URI community plugin [Obsidian|GitHub] by Vinzent does the job.
You need to add a block to the header you want to link to. These don't seem very well documented on the Obsidian site (am I missing something?). But they're simple: just add ^xyzabc to the end of your header (or any paragraph).
It must appear as the last text on the line, and xyzabc can be any string of, it seems, arbitrary length. They can start with a number, which is nice. So I have a subheader within JDex entry 21.41:
### +WP1 Map out the course structure ^2141wp1
– and you can see how natural it is to create a block ID to link to that. The block ID represents the header's location within my system.
Creating the link in Things
I make these 'cross-links' look the same every time, using a dash (i.e. Markdown bullet list) and the unicode ⇄ 'rightwards arrow over leftwards arrow'. Raycast's Search Unicode Symbol, bound to Ctrl+Opt+U, enables me to type that easily.1 (Hopefully it does the same for you Windows users?)
If you use the Advanced URI plugin to Copy URI for current block you'll get the full path to the file:
Fuck AI for making use of proper typographic symbols some sort of indicator of slop. On a Mac it's trivial to type these symbols, and you should be using them all the time. An en-dash, for example, is just Opt+-; for the em-, also hold Shift. I've used it 20 times a day for the last decade (though I've never liked the em-dash in my own writing – just feels way heavy. I go with space-en-space.).
(Also FWIW I have Ctrl+Opt+I bound to the emoji palette and Ctrl+Opt+C to the clipboard history. I use all 3 many, many times a day.) ↩
And Things' Markdown support is a bit lazy in that it 'supports' Markdown, but does nothing with it; that is, a long link is always displayed, and never replaced by the [text of the link](uri://…). ↩
