Live data from Hacker News

Ask HN: What is your approach for managing personal digital assets?

news.ycombinator.com

21–27 of 27 posts

Re: Ask HN: What is your approach for managing personal digital assets?

#21
For around 15+ years I have maintained a super-curated deep taxonomy of files and folders, symlinking stuff, and so on. I've given up around 6 years ago switching to Emacs, actually almost all my files are org-mode managed, the current approach is:

- time-based notes, like

  ~/org/yyyy/
             notes/mm-dd.org
             binders/{various,topic-based,notes}.org
             work/{work,related,notes}.org

  ~/org/lib/{long,lasting,topic,based,notes}.org

  ~/data/ 
essentially files get attached to headings in daily notes `~/org/yyyy/notes/mm-dd.org` so they get moved to a cache-dir style tree org-attach managed under `~/data/` and I can access files easily with org-attach-reveal or linking them in the org-mode note. Things that have something in common, let's say electricity bills, have a daily note heading relevant note with linked emails (ol-notmuch), transaction (beancount), attached pdf bill etc. Since they are regular I have a binder note, so `~/org/yyyy/binders/electrictyBillNote.org` that org-transclude all relevant daily headings and since it's a topic that goes beyond the current year I also have a `~/org/lib/electrictyBillNote.org` note that transclude all the years.

At first might sound complicated but it allow both an overview of anything, just looking at years binders notes, the maximum level of details going down to daily notes, and the maximum high level zoom via `lib` notes. Essentially I have a personal history, I can slowly change my notes having only to keep stuff under `lib` consistent, and having all things relevant to a certain timeframe immediately at hands.

That's to org-roam-node-find I can directly access any attached file just hitting a key and start typing something, it's a graph not a hierarchy, but when I'm on the desired results or around something I look for I can also "explore the surroundings" thanks to the time-based structure and org-mode links, org-roam backrefs and so on.

Anything is linked or attached in notes: my mails reside also on my iron indexed via notmuch, my files org-attached, some configs are created from notes (org-babel, tangled), something is attached and linked per file, some linked to the parent dir some files reside in dired etc. Basically I have a near-semantic, full-text searchable, metadata rich, tool to manage anything.

I've tried to do more with org-ql but honestly I'm not disciplined enough, I've tried the LLM way (khoj) and find very little useful results, so far direct access using org-mode heading as titles/searchable stuff with tags and optional org-roam aliases prove to be powerful enough. Sometimes when I fails to find something this way I going the counsel-rg way (full text ripgrep) and than I add a roam_alias matching the keywords I've tried before. I've also tried briefly recoll/counsel-recoll (xapian) but I've never have to use it except for testing.

I do miss phone/sms integration but having daily notes and org-capture at hand if something relevant happen I'll note in a daily note. If it became or I know it's relevant it will also be in some year-binder, if it's useful or became useful beyond an year it goes in lib, a kind of long-term archive. That's my current best compromise between "noise" and "order". Various automation do help noting anything, org-store-link/org-insert-link do help do link an email, so org-capture initiated from a mail, similarly yasnippet and hydra do help for recurrent transactions (all babel blocks get tangled to a tree where beancount import * anything), some helpers allow to link things to open with specific apps via elisp: links and so on.

It was a journey, I've hesitated much, I still have a gazillion of things to sort and automate properly, some broken/old automation etc BUT anything works smooth enough and I never lost a single bit of information nor I've failed to find something with nearly no effort so...

Re: Ask HN: What is your approach for managing personal digital assets?

#22
post #16

My approach is a flat date-based file structure with a markdown-based Zettelkasten on top: File are mostly in a flat structure. In ~/, I have directories like documents, photos, videos that directly contain all the files. For certain topic, I create separate directories (e.g., PhD, major projects). The filenames always start with an ISO date (yyyy-mm-dd). For documents, it could be like 2024-03-24_inv_google.pdf. Her…

What do you use for scanning and OCRing? I haven't found a solution that _doesn't_ make me want to throw it all out and just retype the doc by hand. I too am moving towards a flat hierarchy with dates after attempting Zettlekasten and PARA. I like your "Zettlekasten layer on top of the fs" idea. Def going to do something similar.

Not OP but for some stuff I only get on paper, so I need to scan I've simply have a rough "import" semi-automated workflow:

- scan to an image/as many as needed

- mogrify -deskew 90%

- for i in .png; do; convert $i ${i/ng/df}; rm $i; done

- pdftk cat output aname.pdf

- ocrmypdf --force-ocr -c -i (wrap tesseract and others)

I normally get a good enough scan. If it's not the case I might manually edit images than feed to the rest of the pipeline. Gimp G'Mic QT Repair Scanned Document filter works IME far better than unpaper and others, but it's not that automatic, need to be tuned case by case. However that's rare.

For notes, personally I go for time-based notes, similar to ZK but with the time-constraint, I've described the system in this page.

Re: Ask HN: What is your approach for managing personal digital assets?

#23

Has anyone done a (hopefully) systematic survey of the processes and software people use to store their stuff, sort of like a usesthis.com but just for storing assets, and how well that's worked over time? My guess is the successful strategies would look a lot like Brajeshwar's comment, a thoughtful plan that uses simple software and formats, some planning for the future, and, probably critically, regularly doing "di…

Not much a survey, but in the past I've tried recoll (xapian wrapper, CLI and GUI, very simple to setup) only to realize that for my needs and the way I take notes (described in this page, another comment) I do not need full-text search except for emails (that are handled by notmuch, locally so indexed by xapian anyway).

I've tried for the web part archivy (pythonic web-app) and zotero, but they do not works like I want so I've dropped them and so far I've just pdf-render contents if I really want to save it beside a link/archive.ph and co.

Re: Ask HN: What is your approach for managing personal digital assets?

#24

Not your drive, not your data. For working storage, I use a local server with a Raid 1 array. The drive manufacture dates are offset by a year or more. I keep a spare drive on hand in case of failure. Every 6 months or so I do a full backup to an external, portable Raid 1 array (again offset manufacture dates) that I keep at grandma's house. For routine backup, I have a pair of microSD cards that I update every Frida…

Is the microsd card holder like this one? https://www.printables.com/model/691366-apple-watch-wristban...

Re: Ask HN: What is your approach for managing personal digital assets?

#25
post #24

Not your drive, not your data. For working storage, I use a local server with a Raid 1 array. The drive manufacture dates are offset by a year or more. I keep a spare drive on hand in case of failure. Every 6 months or so I do a full backup to an external, portable Raid 1 array (again offset manufacture dates) that I keep at grandma's house. For routine backup, I have a pair of microSD cards that I update every Frida…

Is the microsd card holder like this one? https://www.printables.com/model/691366-apple-watch-wristban...

Similar idea.

This looks somewhat heavier/bulkier than mine and I designed a latch/lever to grab and securely hold the card using the notch that is provided on one side. I'm pretty sure this is the intended purpose of the notch.

I'm also pretty sure that most ordinary microSD cards are reasonably waterproof. On the outside, they're solid molded plastic with contacts exposed. I've used different brands in the shower over long periods of time and gone swimming with them without any problem. I just wouldn't recommend powering them up wet --- might short the contacts.

Re: Ask HN: What is your approach for managing personal digital assets?

#26
Step 1. Get all of your shit off the cloud.

You don't own anything on the cloud. Yes, you can put stuff on there, but recognise this one fact. Out there, it doesn't belong to you. Nobody cares whether it remains safe or it disappears. especially don't have stuff on the cloud that doesn't exist anywhere else

To have it remain safe and under your control, it needs to be on your own server(s). And those servers need to be backed up daily.

Step 2. Have multiple backups.

Backups have been known to fail or die. If one backup goes, there should be at least one more.

Step 3. (And this might be sacrilege to some) Don't use encrypted drives.

If there is a problem with the encryption, the whole drive is irrevocably lost. If all files are stored individually, one or more files might get corrupted, but losing the entire drive would be very unlikely.

Step 4. Use hard copy. You know, that old stuff: paper.

Periodically, print out all your accounts, logins and passwords. Store that list in a readable documents package which will be read after your death, labelled such as 'Legal Documents'.

Step 5. Where have you copied all that stuff on your phone?

People too often forget to back up their phones. And it needs to be a backup that is not restricted to a particular company. So you've backed up your iPhone to Apple's Cloud. Are you certain that your next phone will also be an Apple? If it isn't can you replace all that stuff into your new Android phone?

Re: Ask HN: What is your approach for managing personal digital assets?

#27
post #16

My approach is a flat date-based file structure with a markdown-based Zettelkasten on top: File are mostly in a flat structure. In ~/, I have directories like documents, photos, videos that directly contain all the files. For certain topic, I create separate directories (e.g., PhD, major projects). The filenames always start with an ISO date (yyyy-mm-dd). For documents, it could be like 2024-03-24_inv_google.pdf. Her…

What do you use for scanning and OCRing? I haven't found a solution that _doesn't_ make me want to throw it all out and just retype the doc by hand. I too am moving towards a flat hierarchy with dates after attempting Zettlekasten and PARA. I like your "Zettlekasten layer on top of the fs" idea. Def going to do something similar.

I have some high end scanner (because I sometimes scan negatives from film photography). It has a built-in OCR that works pretty well. That's basically the only reason I need to boot MacOS because the Linux drivers are rather bland.
Post reply on HN