Its a beautiful system but where my head explodes (and has been exploding for 4 decades) is over the following scenario.
So in Johnny's system, I assign 21 to automobiles. My VW van gets 21.1, my Citron is 21.2, etc. and the insurance for each car gets a .8 so 21.1.8, 21.2.8, etc.
And I assign 13 to Money. Insurance belongs under money so 13.5 is insurance and life insurance gets 13.5.1, E&O insurance gets 13.5.2, etc.
I also need a top folder for Medical for doc visits, vaxes, ER visits, Surgeries, the kids' allergies and stuff.
So where all this is going is two months later, where is the health insurance policy? Is it under medical or under money? Is the car insurance under Automobiles or Insurance under Money?
Back to my head exploding - this is my issue - I can never remember which branch of the tree to find a specific leaf? Does my annual car tax belong with the Money or with the Auto branch? If I want to see the tax for all the cars at the same time, I put it under Money - Taxes - Auto but when I need to know the last time I paid the tax on the VW, I will assume its filed under Auto-VW-Car Tax.
This is why I can never find anything. All due respect to Johnny but I'm too retarded to use it properly.
You’ve hit the nail on its head here. Almost every piece of information I save has more than one type of contextual relevance, this is not handled by any hierarchical organization system no matter how clever the addressing is. At a certain scale and complexity, I simply cannot remember the magic incantation URL for whatever it is I need. Even search falls apart frequently, because I saved some reference using an abbreviation or synonym to what I think I need.
It doesnt take so much “scale” if one has deficient short term memory/recall/adhd or is (as youve elegantly put it) “too retarded”. (hey - samesies)
Tags/content classifiers/ontologies are I think the solution here, but require continuously grooming your data to ensure it’s classified correctly - a time investment.
My opinion is that modern ML classifiers are helping,l - Ive found some help with tools that recently added auto-tagging - and I think the real magic bullet will be augmenting this capability with relevant personal/activity context. An algorithm can infer much of the contextual relevances that are missing from the current tools if it can match some incoming information to any or all of the areas/topics/projects/horizons/decimal-things that users of organizational tools have decided are important to them.
Yea I don’t often remember which folder I stuck something in.
As a result I’m a heavy search user. I find 90% of the time I can quickly find what I’m looking for with a few keywords.
For the OP, health insurance policy, I’d type “health insurance policy” into my google drive. Then it pops up, wherever it was.
This of course means you name things with something that makes sense to you. Don’t leave it with whatever name when you downloaded it by default.
This also is a huge help with links. I need to find the latest xyz ticket (which I previously viewed) and I’ll type in xyz… and boom, there it is. This is why I don’t clear my browser history very often!
Johnny here. This is the canonical example, and I quote it myself: is it `Insurance > Car` or `Car > Insurance`?
In reality you just decide. One feels better to your brain. And you tend to remember that.
It helps of course if you remain consistent. In the systems we design we’ve realised that most people want the insurance close to the thing being insured.
So in our life admin system we have health, pet, home, motor, and travel insurance as IDs alongside your records for those things. Seems to suit most people.
And don’t forget you’ve got your index as a fallback. I don’t remember most of these numbers but I just launched Bear, typed `insurance` in the search field, and there they are. Now in three clicks I can get to my home insurance which, turns out, is at `12.12`.
This!
i prefer tags over folders for this reason. All notes go into single folder , no sub directories . Because a note can have multiple classifications a tree structure is not natural way to organize them. Add tags , if you have note taking program will show you all possible existing tags you it makes this easier.
I love tags until I actually use them, I always wind up using them inconsistently, or not at all for a specific file, and them bam, I can't find anything at all.
The benefit of file structures is that things have to have a place, you can't not put something in a folder, so for car insurance, it might be in "insurance" or "cars" but it's definitely one or the other. With tags, it could be "insurance", "finance", "cars", "automobiles", "vehicles", "veihcles", etc.
Any tips of how to funnel some strictness into tags so that they're actually usable?
This is why I like Obsidian (or some other linked-documents wiki type of system), because it makes linking things easy, so you can take multiple routes to find a thing. I have a health note and a finances note. Which one does health insurance go under? I pick whichever one seems to make the most sense at the time. Then, in the future, if I'm looking for health insurance and look in the wrong place first, I can easily make a link there to the "health insurance" note/section. Now, I will find health insurance whether I look under health or finances.
The "Obsidian way" that many people recommend is notes that are as small as possible to maximize this kind of effect, but that's not how I like to do it. I prefer bigger notes with lots of headings (that can be nested up to 6 levels), and lots of links within a note and between notes to specific headings. I find this to be a nice blend of hierarchical navigation and link navigation.
Non-text files (like receipts or pictures) get linked from the relevant note or section, and many types of media can be viewed inline in the WYSIWYG editor.
How do you link non text files? Is it like a link to the file on your computer directly? And if so, do you the. Loose the ability to see those files when you are not on your computer but instead using the mobile app? Or is there a way to include those non-text files in the notes syncing apparatus?
I've had this problem for a long time. My solution was to keep my organization as flat as possible. This means everything insurance-related would go to 13.
A flat structure seems less organized, since you are “mixing” stuff, but as long as there isn't too much stuff inside, going through stuff one-by-one is faster than you think. If I do have a lot of stuff in a section, I either split into several sections in the top structure (so 13 is life insurance, 14 is other...), or go one level deeper (not preferred, but I do it when it's very clear and there is too much stuff, like photos, which btw sorting chronologically works best for me).
It is really not much of an issue having 50 top sections. It makes the organization transparent, and indexing, sorting and going one-by-one remains easy.
In today's world, it would be under Ctrl+F -> "insurance". If it were on physical paper, it would be on the folder containing the important documents.
Over-reliance on hierarchy has been identified as a double edged sword of these organizational systems. Hierarchy is only useful when the hierarchy itself matters, e.g. when "grab all documents about this" is a use case, keeping such documents together matters. If the use case is finding one particular document, we have better tools than hierarchy for that, and any time spent on organization is more or less wasted when a good file name and simple storage rules do the work better, and are irrelevant for modern retrieval methods.
Hierarchy is where taxonomy fights break out; tagging is the way, the truth, and the light. That's one of the many, many reasons I love 1Password over Bitwarden. I don't need folders I need tags. Same for Firefox's infinitely better bookmarking system than Chrome's, for the same reason
I recently learned that you can also access the mac spotlight search index from your terminal using mdfind[1].
It's very nice to use when looking for misplaced pdfs.
I had exactly this issue before, an I blame overthinking things. Trying to put in place a system where none is needed.
I ended up with a box, in the box there are large plastic envelopes, and each envelope is labelled.
I have:
- "assets" (cars, warrantees, service records, purchase invoices etc)
- "health" (all medical related things)
- "insurance" (everything insurance related)
- "guns" (I like guns... so licenses, legal paperwork, etc etc)
The best thing is, this is a box. So worst case, even if I misfiled something, all I need to do is rifle through a box. The box is portable and universal, and if my wife needs something, I can easily guide her to where to find it.
I have set up a Zettelkasten, which alternates between letters and numbers.21a5 is for my Alcatel 3085 mobile. So it's a similar idea, I think.
Classification is a vexing problem I've tried to grapple with.
The Dewey Decimal system is a really good example of people trying to get a handle on it. It's not easy. "Arduino Cookbook" is in the Electronics section (621.3810285536 to be precise, although the decimal system doesn't usually get that crazy in its specificity), whilst "Getting Started with Arduino" is in 005.133, more that half a library apart.
As one commentator put it, a book is rarely about one thing. People have criticized the Dewey system, so they throw the baby out with the bathwater and declare their bright shiny new wheel to be the solution.
Except, of course , they skip over the same fundamental problem that Dewey had: books don't really fit into a taxonomy.
One solution which may work for small personal systems is to not bother to use a hierarchy. Put car insurance in one folder, then file all folders alphabetically.
Getting more sophisticated, if you have a hierarchical system, consider indexing. I found an old address book that I had lying around. Add entries into that. Sure, entries won't be in strict alphabetical order, but hopefully you won't have a system so big that you can't find anything.
Indexing is "the" solution, because you don't have to try to figure out where in the taxonomy something is. You just look it up. Indexing also allows you to construct different "views" of a subject, thereby allowing you bypass taxonomic choices.
Being a computer geek, you could keep an index file on computer. You can then simply grep it.
Using your example, you could describe the insurance for your Citron as 21.2.8. Or, you could call it automobiles.citron.insurance. Which looks a lot like automobiles/citron/insurance. How is this any different from just using a folder structure?
I guess the key difference is the amount of branches that you can make at any given level.
Calling a folder citron instead of 2, gives room to have N amount of folders at that level instead of just 10, which I think is the constrain that made this system manageable.
I haven’t used this yet, but I imagine that it’s easier to look in two places than a dozen places. So maybe it doesn’t matter if it’s perfect. Pick one of the two possibly correct categories at random and you’ve still reduced your search space. I’ll probably give that a shot and report back ;)
symlinks or hardlinks might help with that, depending on your needs. With hardlinks you will see the same file in both locations, and if you change or remove the file it will be removed in the other directory as well
> or remove the file it will be removed in the other directory as well
Your operating system must behave way different from mine. Hardlinks in both Linux and macOS are merely "named inodes" and only when all references to the inode are removed is the storage actually subject to reuse
Off the top of my head, all PKMs make trade-offs on discoverability, portability, maintainability, and ease of recall. Broadly, "discoverability" is how likely you are to stumble on something you'd forgotten (just recently, I found a file in my "taxes" directory listing all the documents I needed last year, which was a big help, and which I did not remember writing), "portability" is how resistant the system is to a company shutting down/project being abandoned, "maintainability" is how easy to keep your system consistent with its principles (including inserting a new note), and "ease of recall" is how easy it is to find something if you know you're looking for it.
When thinking about a lifelong PKM, I feel like I value portability more than most; something highly tied to a particular company like Notion is right out for me, and I'm leery of stuff like Obsidian or even org-roam, since even if the entries in those systems are just text, I just know that someday the logic that ties them together will stop being developed/maintained and I'll have to migrate.
I feel confident in directory structures and text files as long-term mediums though, and so JD is appealing to me, but its maintainability (specifically the cognitive load around inserting a new note) is such a stumbling block for actually creating content for it. Not to mention the primary thing it trades maintainability off for (ease of recall) is almost entirely solved by search functionality, leaving discoverability as the only benefit over just chucking everything in a flat "notes" directory.
I do something PARA-adjacent now, and I might just commit to that, although denote is interesting as an Emacs user for a slightly more portable tagging- and search-based option.
I keep everything in a single folder, as plaintext Markdown files.
Even if my own software breaks someday, I can always ingest these into a flavor-of-the-month indexer (though I think sqlite + fts plugin goes a long way) and carry on.
I use paperless-ngx with about 10 document types. Most documents don’t have a type. I scan everything we get in paper form (mail, prescriptions, anything) with a document scanner and dump it all in there. For email pdf I forward it to an account that ingests the documents.
It has a classifier so after a while it picks up the date, correspondant (if you set it), document type automatically.
Daily backups to rsync.net and the export function is a dump of all the files + metadata in json format.
The search (via OCR) is pretty good and it’s the easiest system to manage I’ve found. I used mayanedms before but it’s too cumbersome and upgrading major versions was very complicated.
> trade-offs on discoverability, portability, maintainability, and ease of recall
Good insight. Thx.
I feel fairly confident in Obsidian as a portable system, since its basic concept (markdown docs with tags and links between) can be found in many other apps [^1].
I'm currently using Bear, but I've made experiments migrating to Obsidian without issues.
[^1]: Bear Notes, IA Writer, Ulysses, Craft (kinda), NotePlan, ...
I really appreciate what Johnny Decimal is trying to solve - we're all struggling with digital organization and the appeal of a clean, simple system is undeniable.
Having implemented similar approaches across several teams, I can say it works beautifully for personal projects or well-defined small team efforts. But here's the challenge: most real-world information refuses to fit into single categories. A technical spec might be simultaneously system architecture, compliance documentation, etc. While the Johnny Decimal strength is its rigid simplicity, that's also its weakness when facing actual organizational complexity.
Rather than fighting these natural interconnections, I've found more success embracing them - using approaches that allow documents to exist in multiple contexts while maintaining the Johnny Decimal core goal of findability/searcability. The solution to chaos might not be enforcing a decimal hierarchy, but rather building systems that match how information actually flows in modern organizations.
For me, that is the value of tags. No need to have duplicates to have items represented in multiple categories, yet each appropriate category gets a nod about the particular item.
I think having a system is more important than which system it is. I don't see much benefit to limiting your hierarchy to 3 levels. Putting metadata like creation time in filenames is probably the wrong thing to do, since it's redundant, although it's mighty tempting-and I do it all the time.
I have found that after multiple migrations from one computer to another, some of my file creation dates are incorrect. I don't use JD but I do have a lot of stuff in yearly folders and some of it's clearly wrong. Like I know that I started one of my graphic novels in 2012 but some of the first few pages have dates in 2014 and 2019. Did a computer migration change the dates? Did some edit I did later on save it as a new file? I don't know. I just know the date's way off.
I agree that the choice to have any system is important.
Totally agree. Consider the simple act of copying a file. Will it retain the original date or start fresh? There is a correct answer, but maybe it depends on the operating system, or the program you're using to do the copy. But I don't care. "Ain't nobody got time for that." When I want to know the creation date or if I just want a unique name I add a the date as a suffix; 022125. It also helps that it's much easier to see at a glance.
Allowing the filesystem to track creation time means you have to worry about how you move the data around and whether the tools you're using preserve it properly. A folder named 20250221-nyc-trip is a coarse but very durable way to store that.
Agree. The benefit of posts like these is that someone has documented their system and iterated on it. You can then steal ideas that work for you.
As a not very organized person, and having struggled with getting personal systems running, guides like this help quite a bit. I've only improved by taking bits that stick for me (https://www.hanselman.com/blog/one-email-rule-have-a-separat...). Anytime I tried a whole system, it failed to get going at all causing me more stress.
I will say that I have a few hierarchies I use regularly that go beyond 3 levels, and they are annoying to work with. There are times where I will copy the entire sub-directory to my desktop just to reduced how many levels I'm working from. Then once I'm done, I'll copy the files back into their little "box" and delete the desktop version.
That sort of thing really makes me miss the Miller Column Filebrowser on my NeXT Cube (and wish that Apple's implementation were more like to it --- it just doesn't "feel" right to me when I use it on my MacBook).
I tried many organization systems, including Johnny Decimal like PARA. And none of them worked for me.
As an ADHD person, I've found the best way for me is not put effort in organizing at all.
For that reason I've found tools like Logseq/Tana/Reflect does a great job. I just write in the journal and tag items accordingly if required, then if I need to write some long form document, I create specific pages for it.
Then search and backlinks are everything I need. My brain works better searching than browsing.
After years of searching for an organizational solution myself, switching between countless applications, numerous applications, and a concoufany of feedback, insights and ideas from xyz influencer, this is exactly the same path i've settled on, despite not being diagnosed with ADHD myself -- though, the signs are all there.
A structure loosely connected to past notes via a weekly 'cleaning/review' process in my "PKS", where I'll /search/ for tags, filenames, file contents and loosely link things together.
It's saved me countless hours, but more importantly its drastically reduced analysis paralysis and kept me focused on the most important thing -- writing.
Exactly. Intricate systems are pure noise for me, a simple MD file opened in a texteditor like Sublime is enough, or as you say just a simple taggable system or really just a bunch of files in folder - as long as you have great search you'll find stuff in no time and forget most as you should.
I personnaly just have a huge file with various notes, text, todos or whatever for each year divided into days, then i can just scroll up through days, or search to find out what i did and what day - some days have nothing, some have lots. Some topics / projects get their own file.
I don’t have ADHD (that I know of) and still love Logseq. For me, it’s the perfect mix of notetaking, journaling, outlining, task tracking, and lightweight hierarchy/linking.
I find that if I have to organize or categorize entries in a system, entries just don’t get logged at all.
most stuff don't work, and don't stand the test of time.
anyway, here is what's has been working for me:
for physical stuff (documents, printouts etc): a dumb file organizer box, one of those where you can hang those hanging manila folders. and of course a few such folders. I bought fifty such folders some years ago, have used about half so far?
for digital stuff: a simple mediawiki installation. it's hosted at home and it's not accessible from the public internet. the visual editor makes it low-friction to edit, the categories system works well enough, a page can belong to more than one category and there's always a search function that works well enough.
the nice thing about mediawiki is that you can upload and embed images, you can link to other systems (like files in nextcloud) and you can upload whole files and link to them from various pages.
ok, it's been a few days now using it, and I'm starting to love it. Specially their API access, so I can send my voice note transcriptions to my daily note.
The custom object types is great, I missed that feature from Anytype in other tools.
The only drawback is that they didn't implemented an import from other platforms tool. And I have all my history and notes on Logseq now.
I just tried it, looks good. But TBH, I miss outlining. I would like they offer a way to have an outlining mode (with collapsing ability). Thanks for the recommendation.
> We believe that everybody should have access to tools for building knowledge. Therefore, the core product of Capacities is and will remain free. Read our promise
I think this is definitely a cool system so not meaning to knock it, but I used to hyper-optimize all parts of my life and it was exhausting. So one day I just stopped. I started focusing on just being present, prioritizing, and trying to remember things that were important. I still take notes and have todo lists and stuff but they are similarly for being in the moment- just for the time right when I'm using them. I may have lost some things over the years but the removing the stress has made me better at all the things I was working on in general.
This exactly. Losing a few things pales in comparison to less mental overhead in every single task I do.
All I do now is keep a notebook with a rolling to do list where I make a new page every day and write what I have to do. Meeting notes also go there if they don't belong in some project files.
The only thing I wish I could do is keep notes from books somehow in an easily referenceable format. Kindle sucks, Obsidian is 180 from TikTok but still brain rot.
I originally started with Johnny.Decimal for my life and after giving it a big try, switched to PARA.
J.D is fine (maybe even great) if your categories are relatively static, such as a small business, but as an individual, I found it very restrictive and challenging to remember. Moreover, while the decimals are cool, I found them somewhat irrelevant if I was the only one referencing them.
J.D is optimized for retrieval, where what I needed was optimized storage, and then occasional retrial.
To each their own of course, and using any system is better than none.
Readit News
So in Johnny's system, I assign 21 to automobiles. My VW van gets 21.1, my Citron is 21.2, etc. and the insurance for each car gets a .8 so 21.1.8, 21.2.8, etc.
And I assign 13 to Money. Insurance belongs under money so 13.5 is insurance and life insurance gets 13.5.1, E&O insurance gets 13.5.2, etc.
I also need a top folder for Medical for doc visits, vaxes, ER visits, Surgeries, the kids' allergies and stuff.
So where all this is going is two months later, where is the health insurance policy? Is it under medical or under money? Is the car insurance under Automobiles or Insurance under Money?
Back to my head exploding - this is my issue - I can never remember which branch of the tree to find a specific leaf? Does my annual car tax belong with the Money or with the Auto branch? If I want to see the tax for all the cars at the same time, I put it under Money - Taxes - Auto but when I need to know the last time I paid the tax on the VW, I will assume its filed under Auto-VW-Car Tax.
This is why I can never find anything. All due respect to Johnny but I'm too retarded to use it properly.
It doesnt take so much “scale” if one has deficient short term memory/recall/adhd or is (as youve elegantly put it) “too retarded”. (hey - samesies)
Tags/content classifiers/ontologies are I think the solution here, but require continuously grooming your data to ensure it’s classified correctly - a time investment.
My opinion is that modern ML classifiers are helping,l - Ive found some help with tools that recently added auto-tagging - and I think the real magic bullet will be augmenting this capability with relevant personal/activity context. An algorithm can infer much of the contextual relevances that are missing from the current tools if it can match some incoming information to any or all of the areas/topics/projects/horizons/decimal-things that users of organizational tools have decided are important to them.
As a result I’m a heavy search user. I find 90% of the time I can quickly find what I’m looking for with a few keywords.
For the OP, health insurance policy, I’d type “health insurance policy” into my google drive. Then it pops up, wherever it was.
This of course means you name things with something that makes sense to you. Don’t leave it with whatever name when you downloaded it by default.
This also is a huge help with links. I need to find the latest xyz ticket (which I previously viewed) and I’ll type in xyz… and boom, there it is. This is why I don’t clear my browser history very often!
In reality you just decide. One feels better to your brain. And you tend to remember that.
It helps of course if you remain consistent. In the systems we design we’ve realised that most people want the insurance close to the thing being insured.
So in our life admin system we have health, pet, home, motor, and travel insurance as IDs alongside your records for those things. Seems to suit most people.
And don’t forget you’ve got your index as a fallback. I don’t remember most of these numbers but I just launched Bear, typed `insurance` in the search field, and there they are. Now in three clicks I can get to my home insurance which, turns out, is at `12.12`.
https://share.icloud.com/photos/0afQRa-furBCpa9rOIc3r3Q7g
Haha, nope. Different brains work differently. One day I genuinely prefer one, a week later another.
The benefit of file structures is that things have to have a place, you can't not put something in a folder, so for car insurance, it might be in "insurance" or "cars" but it's definitely one or the other. With tags, it could be "insurance", "finance", "cars", "automobiles", "vehicles", "veihcles", etc.
Any tips of how to funnel some strictness into tags so that they're actually usable?
The "Obsidian way" that many people recommend is notes that are as small as possible to maximize this kind of effect, but that's not how I like to do it. I prefer bigger notes with lots of headings (that can be nested up to 6 levels), and lots of links within a note and between notes to specific headings. I find this to be a nice blend of hierarchical navigation and link navigation.
Non-text files (like receipts or pictures) get linked from the relevant note or section, and many types of media can be viewed inline in the WYSIWYG editor.
A flat structure seems less organized, since you are “mixing” stuff, but as long as there isn't too much stuff inside, going through stuff one-by-one is faster than you think. If I do have a lot of stuff in a section, I either split into several sections in the top structure (so 13 is life insurance, 14 is other...), or go one level deeper (not preferred, but I do it when it's very clear and there is too much stuff, like photos, which btw sorting chronologically works best for me).
It is really not much of an issue having 50 top sections. It makes the organization transparent, and indexing, sorting and going one-by-one remains easy.
Over-reliance on hierarchy has been identified as a double edged sword of these organizational systems. Hierarchy is only useful when the hierarchy itself matters, e.g. when "grab all documents about this" is a use case, keeping such documents together matters. If the use case is finding one particular document, we have better tools than hierarchy for that, and any time spent on organization is more or less wasted when a good file name and simple storage rules do the work better, and are irrelevant for modern retrieval methods.
[1]https://ss64.com/mac/mdfind.html
I ended up with a box, in the box there are large plastic envelopes, and each envelope is labelled.
I have:
- "assets" (cars, warrantees, service records, purchase invoices etc)
- "health" (all medical related things)
- "insurance" (everything insurance related)
- "guns" (I like guns... so licenses, legal paperwork, etc etc)
The best thing is, this is a box. So worst case, even if I misfiled something, all I need to do is rifle through a box. The box is portable and universal, and if my wife needs something, I can easily guide her to where to find it.
Classification is a vexing problem I've tried to grapple with.
The Dewey Decimal system is a really good example of people trying to get a handle on it. It's not easy. "Arduino Cookbook" is in the Electronics section (621.3810285536 to be precise, although the decimal system doesn't usually get that crazy in its specificity), whilst "Getting Started with Arduino" is in 005.133, more that half a library apart.
As one commentator put it, a book is rarely about one thing. People have criticized the Dewey system, so they throw the baby out with the bathwater and declare their bright shiny new wheel to be the solution.
Except, of course , they skip over the same fundamental problem that Dewey had: books don't really fit into a taxonomy.
One solution which may work for small personal systems is to not bother to use a hierarchy. Put car insurance in one folder, then file all folders alphabetically.
Getting more sophisticated, if you have a hierarchical system, consider indexing. I found an old address book that I had lying around. Add entries into that. Sure, entries won't be in strict alphabetical order, but hopefully you won't have a system so big that you can't find anything.
Indexing is "the" solution, because you don't have to try to figure out where in the taxonomy something is. You just look it up. Indexing also allows you to construct different "views" of a subject, thereby allowing you bypass taxonomic choices.
Being a computer geek, you could keep an index file on computer. You can then simply grep it.
https://en.wikipedia.org/wiki/Melvil_Dewey#Controversies
Bear does this really well with its hierarchical tags.
Most filesystems can do this with hardlinks (but the UX mostly sucks).
Using your example, you could describe the insurance for your Citron as 21.2.8. Or, you could call it automobiles.citron.insurance. Which looks a lot like automobiles/citron/insurance. How is this any different from just using a folder structure?
Calling a folder citron instead of 2, gives room to have N amount of folders at that level instead of just 10, which I think is the constrain that made this system manageable.
I use Paperless to catalog all my PDFs, Obsidian for notes and Gmail for email, Todoist for tasks and Cloze for CRM, all of which support tags.
Your operating system must behave way different from mine. Hardlinks in both Linux and macOS are merely "named inodes" and only when all references to the inode are removed is the storage actually subject to reuse
When thinking about a lifelong PKM, I feel like I value portability more than most; something highly tied to a particular company like Notion is right out for me, and I'm leery of stuff like Obsidian or even org-roam, since even if the entries in those systems are just text, I just know that someday the logic that ties them together will stop being developed/maintained and I'll have to migrate.
I feel confident in directory structures and text files as long-term mediums though, and so JD is appealing to me, but its maintainability (specifically the cognitive load around inserting a new note) is such a stumbling block for actually creating content for it. Not to mention the primary thing it trades maintainability off for (ease of recall) is almost entirely solved by search functionality, leaving discoverability as the only benefit over just chucking everything in a flat "notes" directory.
I do something PARA-adjacent now, and I might just commit to that, although denote is interesting as an Emacs user for a slightly more portable tagging- and search-based option.
I keep everything in a single folder, as plaintext Markdown files.
Even if my own software breaks someday, I can always ingest these into a flavor-of-the-month indexer (though I think sqlite + fts plugin goes a long way) and carry on.
It has a classifier so after a while it picks up the date, correspondant (if you set it), document type automatically.
Daily backups to rsync.net and the export function is a dump of all the files + metadata in json format.
The search (via OCR) is pretty good and it’s the easiest system to manage I’ve found. I used mayanedms before but it’s too cumbersome and upgrading major versions was very complicated.
Good insight. Thx.
I feel fairly confident in Obsidian as a portable system, since its basic concept (markdown docs with tags and links between) can be found in many other apps [^1].
I'm currently using Bear, but I've made experiments migrating to Obsidian without issues.
[^1]: Bear Notes, IA Writer, Ulysses, Craft (kinda), NotePlan, ...
Having implemented similar approaches across several teams, I can say it works beautifully for personal projects or well-defined small team efforts. But here's the challenge: most real-world information refuses to fit into single categories. A technical spec might be simultaneously system architecture, compliance documentation, etc. While the Johnny Decimal strength is its rigid simplicity, that's also its weakness when facing actual organizational complexity.
Rather than fighting these natural interconnections, I've found more success embracing them - using approaches that allow documents to exist in multiple contexts while maintaining the Johnny Decimal core goal of findability/searcability. The solution to chaos might not be enforcing a decimal hierarchy, but rather building systems that match how information actually flows in modern organizations.
* https://news.ycombinator.com/item?id=36308366
* https://news.ycombinator.com/item?id=37506640
* https://news.ycombinator.com/item?id=25398027
Johnny Decimal: A System to Organize Projects (2015) - https://news.ycombinator.com/item?id=37506640 - Sept 2023 (118 comments)
Johnny Decimal - https://news.ycombinator.com/item?id=36308366 - June 2023 (193 comments)
Johnny.Decimal – A System to Organize Projects - https://news.ycombinator.com/item?id=36300472 - June 2023 (1 comment)
Johnny•Decimal - https://news.ycombinator.com/item?id=33683874 - Nov 2022 (1 comment)
Johnny.Decimal - https://news.ycombinator.com/item?id=25398027 - Dec 2020 (187 comments)
Johnny.Decimal – A system to organise projects - https://news.ycombinator.com/item?id=13770827 - March 2017 (2 comments)
I agree that the choice to have any system is important.
As a not very organized person, and having struggled with getting personal systems running, guides like this help quite a bit. I've only improved by taking bits that stick for me (https://www.hanselman.com/blog/one-email-rule-have-a-separat...). Anytime I tried a whole system, it failed to get going at all causing me more stress.
A structure loosely connected to past notes via a weekly 'cleaning/review' process in my "PKS", where I'll /search/ for tags, filenames, file contents and loosely link things together.
It's saved me countless hours, but more importantly its drastically reduced analysis paralysis and kept me focused on the most important thing -- writing.
cacophany?
https://www.merriam-webster.com/dictionary/cacophony
Deleted Comment
I personnaly just have a huge file with various notes, text, todos or whatever for each year divided into days, then i can just scroll up through days, or search to find out what i did and what day - some days have nothing, some have lots. Some topics / projects get their own file.
Agreed - I looked at the website for a hot second, got overwhelmed and immediately closed it
Consistency is key for a good organization system. Unfortunately, consistency in such manners of life isnt our forte
I find that if I have to organize or categorize entries in a system, entries just don’t get logged at all.
Only bad thing is their mobile app, it's so bad.
anyway, here is what's has been working for me:
for physical stuff (documents, printouts etc): a dumb file organizer box, one of those where you can hang those hanging manila folders. and of course a few such folders. I bought fifty such folders some years ago, have used about half so far?
for digital stuff: a simple mediawiki installation. it's hosted at home and it's not accessible from the public internet. the visual editor makes it low-friction to edit, the categories system works well enough, a page can belong to more than one category and there's always a search function that works well enough.
the nice thing about mediawiki is that you can upload and embed images, you can link to other systems (like files in nextcloud) and you can upload whole files and link to them from various pages.
The custom object types is great, I missed that feature from Anytype in other tools.
The only drawback is that they didn't implemented an import from other platforms tool. And I have all my history and notes on Logseq now.
Wow, I'm sold.
We've received great feedback from ADHD users about how it has helped them throughout their lives
All I do now is keep a notebook with a rolling to do list where I make a new page every day and write what I have to do. Meeting notes also go there if they don't belong in some project files.
The only thing I wish I could do is keep notes from books somehow in an easily referenceable format. Kindle sucks, Obsidian is 180 from TikTok but still brain rot.
J.D is fine (maybe even great) if your categories are relatively static, such as a small business, but as an individual, I found it very restrictive and challenging to remember. Moreover, while the decimals are cool, I found them somewhat irrelevant if I was the only one referencing them.
J.D is optimized for retrieval, where what I needed was optimized storage, and then occasional retrial.
To each their own of course, and using any system is better than none.