Ask HN: How do you manage your personal documents?
As I get older, I'm finding it increasingly difficult to manage the seemingly constant influx of 'important' documents.
Fortunately, most documents are sent to me digitally, where I can easily file them on my computer and back them up.
For non-digital documents like auto lease paperwork, home-related legal documents, major bills that I want to keep around, etc, my filing system is incredibly poor. By 'poor', I mean: it takes far too long to find the document I'm looking for, it's difficult to decide where to file something, and it's too difficult to find things.
Every few years, we take a stab at reorganizing our files, but the lack of searchability and the other affordances of digital documents leaves me wanting more.
How do you organize your personal documents? Do you digitze them somehow? Do you have a great filing system? I want to know.
185 comments
[ 5.0 ms ] story [ 216 ms ] threadI have an inbox for mail. I throw out junk and envelopes and put things that are important-but-not-actionable in a plastic tote in chronological order. Bills go in there after they are paid. (By me)
Every six months or so or when tax time comes around my wife files the papers. She gets depressed doing it, but at least bills get paid on time because I don’t get depressed about filing paper in real time.
By familial entity (family, mum, dad, kids)
Then By year/month/day (bills) Or time independent (diploma, ...)
Then if I need to organize a document for my family auto
- the contract goes to family > auto
- the bills go to family > yearmonth > yearmonth_auto_provider-name.pdf
A search on the tree on file name is enough for us.
https://wiki.nikitavoloboev.xyz/unix/my-file-system
Those folders are grouped together by broader subject - loan documents, tax documents, etc. - in alphabetical order.
However, the most important documents I've digitized and keep in a secure digital datastore.
In the case of documents that have a reliable internet service source (e.g. Amazon order receipts) I don't even bother filing or scanning. Instead I go get the document again from its source if I need it. Amazon keeps orders going back to the beginning, afaik.
There is a custom program running in a container that will feed any new doc to pdfsandwich (or similar OCR tooling), move the doc into a digital inbox prefixed with the current date. From there I can either leave it as it is (the date is usually good enough to find things) or - if I get to it - give it a proper name and put it into a folder structure (like some of the other responders).
The custom tooling also indexes all docs and provides my family a web interface to search.
Simple and effective.
The date and comprehensive filename is often enough to find a document, but additionally I can do a "full text search" on all documents with the text layer that is put onto scans by the OCR.
I have a Samba share that I added as a destination on my network scanner. I then tag them, add a correspondent, and never think about them again. PDFs that are sent to me are just uploaded and tagged the same way.
The paper copies are then thrown into a box in hopes I never need the originals.
I back up the document storage regularly.
You may have just convinced me to ditch Evernote.
Thanks to you next weekend will likely be a write-off :-)
I’ve been wanting to get off Evernote for years. I can see myself using this in combination with Obsidian.
I'm speaking from the point of view of having looked at using all the different flashy to-do lists, all the different 'revolutionary' note taking tools (including Evernote quite extensively), and now using Obsidian, which is basically an IDE for markdown with ability to link files to one another.
The simpler tools force you to come up with systems of organisation / linking if you want to do something more complex, but that's the great thing - you get to build your own system incrementally to a point where it's something that actually works really well for you. It's also Electron, so you can write your own plugins!
Especially for a company like Evernote who are no strangers to security issues.
I also dislike the “AI based” content suggestion crap or whatever that is. I don’t want that in my note taking tool. Things like that cross a line for me. It feels yucky.
And in general Evernote has totally stagnated as a product in my opinion.
I’d even happily pay them more money if the product fitted more in alignment with my use-case.
It’s really been great over the years but there is so much they could have done to keep it relevant and for me it just isn’t any more.
DropBox has gone a little in the same direction. An “originally great” disruptive product that sadly has bloated and stagnated without adding much more actual value on top of the original concept.
What's border line criminal is the new version is WAY slower than the old one.
That way, I don't need to have a date for every document.
I’ve not heard of this before but it’s cool!
As opposed to paperless, Mayan provides fine grained access control via ACLs and also allows 'directories' in addition to tags. Dropped it after a while though, since it was too enterprise-y and for in-depth configuration, the documentation was insufficient and I would have to buy the advertised book. Paperless-ng is sufficient for my personal use, though I still miss having directories as an additional level of hierarchical organization alongside tags.
Since I don't have a scanner, I just use the Microsoft Lens app to scan documents on my phone (Android). Paired with Syncthing (https://syncthing.net/), my documents are automatically synced to my desktop from where paperless-ng picks it up from the watched folder and automatically adds it. Tags and correspondents can be automatically added based on keywords in the text.
If you want features such as facial recognition, you might try taking a closer look at PhotoPrism/LibrePhotos. I did try them about a year ago, but in the end just went with the desktop app digiKam (https://www.digikam.org/)
From the About on GitHub:
> Damselfly is a server-based Photograph Management app. The goal of Damselfly is to index an extremely large collection of images, and allow easy search and retrieval of those images, using metadata such as the IPTC keyword tags, as well as the folder and file names. Damselfly includes support for object/face detection, and face-recognition.
[1] https://github.com/Webreaper/Damselfly
- self-hosted on docker, Linux, macOS, or Windows
- cross-platform libraries that can track assets across volumes without duplication [1]
- supports reading and writing to .XMP sidecars and Google Takeout JSON files [2]
- sophisticated image and video deduplication [3]
- quick and novel "samples" UI designed to scale to extremely large libraries (500k+)
https://PhotoStructure.com/why
[1]: https://photostructure.com/faq/what-is-a-volume/
[2]: https://photostructure.com/faq/takeout/
[3]: https://photostructure.com/faq/what-do-you-mean-by-deduplica...
https://github.com/sismics/docs
It stores everything into postgresql... filesystems are ok but it can get out of hand.
I am trying to deploy on microK8s with helm3
I scan everything with ScannerPro on my iPhone (which OCRs text to make it searchable) as items come in via mail etc. or print to PDF if I see myself needing something later. Scanned docs sync to Dropbox, and I move them to their appropriate Dropbox folder (e.g. property, taxes by year, biz, etc.) when I sit down at my laptop. It’s worked pretty well. I think the key is focus on the new docs, and you’ll eventually stop needing to refer to the old paper docs in your old folders (rather than trying to scan literally everything, which seems like a daunting task).
Phone cameras are easy and good, and I'll never need a perfectly cropped scan
The bot names and places the file in the right folder for paperless to ingest.
For digital, I use Google Drive, with rough organization of Taxes/2021, Taxes/2020, ..., Health, Family, Notes, etc. Don't obsess too much, you rarely go back in time so it's OK to keep it lightweight and optimize for the "now" when inserting and spend that little extra time searching later when retrieving.
The exception are things that are too tedious to scan (50+ page mortgage doc), are required to be paper for legal reasons (birth certificates), or sentimental items. Those are surprisingly few, and go into a hanging-folder plastic tub that has a waterproof gasket on the lid.
Most damage from house fires is the water from the firehoses, so it's important to protect from water damage first, then worry about fire damage if you keep your documents at the top of your home (attic). I store my document tub in the bottom of my closet, so it's relatively safe.
And honestly, I'm slowly de-Googling my life and will moving (or at least replicating) my online storage elsewhere.
The main threat model these days is digital, rather than physical. I favor paper records, in a folder by tax year (sorted by month). The iPhone Notes app has a really good document scanner built in, if you need a digital copy you can create it, then delete it. For the house I have a house folder, for the car I have a car folder. I just pay cash for cars, so that really cuts down on documents needed (this kind of thing should be part of your decision-making process!).
I found that such an approach doesn't work for me. I currently own two cars but in the past I have owned over ten cars in my life. I keep purchase and sale documents for each car in a separate folder until no longer needed. I keep maintenance records in their own separate folders. When I sell a car I no longer need, I pass the maintenance records on to the buyer, and I easily locate the title to sign over in the purchase/sale folder from last time. Having these together wouldn't make sense because purchase/sale is a one-time event whereas maintenance happens all the time. For example, to answer the question should I accept my mechanics' recommendation to replace some given part, I review the documents and show them that they already replaced it last year, so they better have a really good reason why it needs replaced again this year.
Finally, there are other records related to cars such as insurance documents. These I just keep for one year only so having them separate in their own folder is a better idea to make it easy to prune out the old/no-longer-necessary documents.
Otherwise everything is digitized and shredded. iOS has a great scanner built into Notes, which can be exported / airdropped as a pdf.
Digital assets can be kept in a managed shared cloud folder like Dropbox or 1pw if containing secrets / identity matters.
If I need to get an invoice from last december, I just lookup around this date in my stack.
Time spent to store information : 0 ; time spent to find something : a few minutes, once every other month.
Every five years, I take the bottom of the stack and file it in the cellar. And I come back from the cellar with 10 year old documents I can either trash (in my office secured bin) or keep in my filing box.
I also keep contracts (insurance, bank, ...) in this filing box.
Last thing : All documents that will be used for my tax returns (at least the equivalent of it in France) go in one folder. I will use it once a year then file this in the "taxes" box.
MRU cache, if you will
It increases the Big O on capturing it, but reduces space and access complexity.
For long running docs like titles, and certifications, passports, and whatnot I stick in an important papers storage Tupperware so it's safe from water damage
I was thinking increasing insert time would increase Big O but they made the point that it's a constant which will cancel out as n approaches infinity.
God I'm a nerd.
The "important papers pile" is a great example of that. Not a second spent wondering where to put something, or even whether it's worth keeping at all.
EDIT: Its date-night bitch - figure it out
Auto-file everything as a time-stack:
- new files go on the desktop
- if I didn't name file right rules automatically rename it per ISO 8601 to "YYYY-MM-DD - title.ext"
- after the file isn't opened for >2 days, it gets auto-sorted into (more or less) "archive/yyyy-mm/yyyy-mm-dd - title.ext"
- note: stick to ISO 8601 across tooling: https://www.iso.org/iso-8601-date-and-time-format.html
With this system, the effect is:
- my desktop has today's and yesterday's working files, everything else is auto-cleaned (you may want 5 or 7 days depending on nature of file work)
- when i need something, I "lookup around that date in my stack" because I find it easier to find things by time than by trying to re-imagine what I'd named it.
- when working somewhere where i create more content then I consume, I group by weeks not months: "archive/yyyy-ww/file" (given ~520 folders per decade, you may, or may not, want "archive/yyyy/yyyy-ww/file" depending on your file system's speed at iterating dirs)
- turn week numbers on your calendar, and you can directly open any folder for any week you did stuff.
- no brain power needed!
More about usage:
Method works beautifully even if you prefer topical folders like project-name or finances or whatever, and computer search is fast/easy at still showing you everything made in a given month no matter what folder it's in.
Seeing files by when made visually 'bundles' files made around the same time (e.g. several days' or weeks' work on the same project). Update dates for new versions and a find by name for the rest of the file name will show you all the versions you have, sorted.
This is infinitely superior to the standard workplace practice of "Title Whatever v3 (tim edits B) jfk (copy).doc".
Thanks to the date prefixed name, if I email files, backup/restore them, or otherwise round trip them to some other file system (looking at you, most NAS, SANs, and object stores), I don't lose the metadata of when created, meaning I can still sort round tripped files by name reversed and see files by recency or clustered by when made.
If I can't find by looking around the date, I can always fiddle with search to find things.
I now get unhappy any time a file doesn't start with a date.
Pro-tips: Apple Shortcuts supports ISO 8601 by name for dates and time. On MacOS I use Hazel to maintain this w/o touching it: https://www.noodlesoft.com ... On Windows I use powershell, Linux perl.
I wonder if there are any digital document organizers like this. Something you just drag and drop important digital documents you want to keep track of and it automatically adds date and time information so you can browse through them in the order they've been added.
If you wanted something Web-based, you could use Google Drive, Dropbox, OneDrive, etc. and get similar functionality.
It could watch certain directories on your computer or in cloud storage and create a queue to suggest adding new documents that appear there in the same order of their appearance, that way you never forget to add an important document and they still get put in chronologically even if you haven't checked the queue in months. It could make different "vaults" you could use to separate documents by person, or separate work and personal life. Vaults could have long-term upkeep rules to keep them from getting bloated, like maybe you delete everything that's 10 years or older.
Basically if it can help me even if I have sloppy organization and take a ton of the pain out of it for me, I'd definitely be willing to pay a bit to try an app like that. Can't say for sure that I would stick with it, but it's a reoccurring problem for me as I have documents across multiple devices, cloud services, emails accounts, and thumb drives that I just can't make myself organize or create and maintain a system to do it for me.
I might resume that practice, now that I think of it.
You can organize it better, but the truth is that search works well enough and you are probably not going to need it.
I use a Doxie [0] scanner and scan absolutely everything meaningful: documents, receipts, invoices, physical photos, note from friends, Christmas cards, credit cards etc. Having a "scan everything" attitude means I don't have to think about what I do and don't scan. I try and save as many of these files as PDFs with OCR (which the Doxie supports)
Every week or two I dump all of the scans from the Doxie into my "Downloads" folder where I rename any files that have obvious content to something more meaningful (easier to search this way). Anything that needs to be manually filed (company documents for example) I do myself.
Everything else get automatically sorted using Hazel [1] after 2 weeks in the Download folder. They get sent to an "Archive" folder that is split into subfolders: PDFs, videos, images, music, design files, documents, etc. All files are automatically sorted into [year]/[month] folders.
This archive is great as it becomes a location for everything that doesn't have a home. For example I also perform a WhatsApp backup every month or two that adds all new photos and content that I've received from people into the same Archive. Likewise anything I've downloaded that hasn't been deleted or moved somewhere can be found there.
[0]: https://www.getdoxie.com/product/doxie-go
[1]: https://www.noodlesoft.com
For "long-term" "important" files, I have some fire-resistant safe boxes
For "really-important" stuff, I use a gun safe (this list is incredibly tiny)
It may be an upfront effort to get going if you have a morass of documents, but helps later on when searching for things. There's old command-line tools like 'qmv' if you want to easily do bulk renaming within a text editor.
Same thing with PDF statements, invoice, receipts, etc I receive via email or whatever. Goes into the giant folder.
Most of the time, I don't even bother renaming the filename to something useful like "Auto Loan Statement". The handy thing that Scanbot does is OCR the scanned content, so I can use macOS's Spotlight to easily pull up whatever I'm looking for. It works really well and I don't worry about it, anymore.
At some point, I may get an itch to run Spotlight queries and start dumping stuff into organized subfolders, but that itch hasn't come yet.
The one thing I break out into an individual subfolder are my yearly tax docs.
I've never been able to stick to an organizational plan involving renaming files with ISO dates (but in fact, Scanbot renames it with ISO date automatically) and organizing stuff into sub-folders. You're a better (wo)man than me, if you can stick with that, though.