Hoarding and Herding
Why I Consciously Manage My Digital CollectionsI still remember well how I used to save everything: music, photos, eBooks, documents – just everything. The fear of missing out or losing something was simply too great. But at some point, I realized: I had everything, but I wasn’t using any of it. The shift from hoarding to tending was a real liberation for me.
Hoarding vs. Tending
Back then, I just collected everything without thinking. The result? A huge mess. I no longer knew what I actually had, and when I searched for something, it was often hopeless. In the end, I had a data graveyard instead of a usable archive.
Today, it’s about actively maintaining and using my collections. Every file has its place and purpose. With clear structures and a little appreciation for what I truly need, everything stays organized and useful.
How I Tend to My Collections
Principles
For me, tending means that every file has its fixed place, and I know why I keep it. I regularly spend time tidying up: deleting what I no longer need, archiving what’s important, and updating what’s outdated. What’s especially important to me is maintaining the metadata in the files. This way, I always have all the information at hand, no matter where I use the files. For this, I often write small programs to take the work off my hands. Tools like Org-Mode, Notion, or my own solutions like the Book Sorter project help me with this.
Practical Steps
I prefer working with tree structures – my brain just likes that. Every file has its place in a logical hierarchy. To find everything quickly, I use tags. And to avoid manually sorting all the time, I automate as much as possible. For example, with scripts that move new downloads directly to the right place.
My Collections in Detail
Music
My music collection is currently about 68 GB in size. I chose MP3 because it’s playable almost everywhere. The files are stored in a tree structure: ~/music/Genre/Artist/Album/XX Songtitle.mp3. The XX represents the track number on the album – if available. As a tree, it looks like this:
Pop/ ├── Adele │ ├── 21 │ │ ├── 01 Rolling in the Deep.mp3 │ │ └── cover.jpg │ └── Skyfall │ └── 02 Skyfall (instrumental).mp3 ├── Alanis Morissette │ └── Jagged Little Pill │ ├── 01 All I Really Want.mp3 │ ├── 02 You Oughta Know.mp3 │ ├── 03 Perfect.mp3 │ ├── 04 Hand in My Pocket.mp3 │ ├── 05 Right Through You.mp3 │ ├── 06 Forgiven.mp3 │ ├── 07 You Learn.mp3 │ ├── 08 Head Over Feet.mp3 │ ├── 09 Mary Jane.mp3 │ ├── 10 Ironic.mp3 │ ├── 11 Not the Doctor.mp3 │ ├── 12 Wake Up.mp3 │ ├── 13 You Oughta Know (Jimmy the Saint Blend) _ Your House.mp3 │ └── cover.jpg
I wrote a Python program that runs through my entire directory tree and generates tags from the path names where none exist yet. This means I can simply copy new songs into the directory structure, run the program once, and the new songs are already tagged with meaningful entries.
Photos
My photos take up about 67 GB of space. Here too, I rely on JPEG because it’s the most universal format. The structure is similar ~/img/fotos/Year/Year-Month-Day/ISO-Timestamp.jpg. This way, I can find every photo quickly by date. As a tree, it looks like this:
2025 ├── 2025-04-10 │ ├── 2025-04-10T16:53:35.jpg │ ├── 2025-04-10T16:53:43.jpg │ ├── 2025-04-10T16:53:54.jpg │ └── 2025-04-10T16:54:07.jpg ├── 2025-04-14 │ └── 2025-04-14T12:48:22.jpg ├── 2025-04-17 │ └── 2025-04-17T20:40:33.jpg ├── 2025-04-25 │ └── 2025-04-25T09:30:41.jpg ├── 2025-04-26 │ ├── 2025-04-26T14:29:16.jpg │ ├── 2025-04-26T18:16:10.jpg │ ├── 2025-04-26T18:16:30.jpg │ ├── 2025-04-26T18:16:34.jpg │ ├── 2025-04-26T18:20:30.jpg │ ├── 2025-04-26T18:20:36.jpg │ ├── 2025-04-26T18:22:01.jpg │ ├── 2025-04-26T18:22:37.jpg │ └── 2025-04-26T20:14:18.jpg │ ├── 2025-05-14T19:55:05.jpg
For photo management, I mainly use a small shell script that sorts freshly downloaded images from my camera into a folder structure based on the metadata in the image. In every image, I store the time when taken, and for photos from my phone, I also store the location unless I explicitly disable it. For manual editing and managing the images, I use the program DigiKam, as it can write metadata back into the photos and also recognize faces.
eBooks
My eBooks take up about 60 GB. I mainly use EPUB and PDF because these formats work on most devices. The structure is a bit more complex: ~/ebooks/Genre/SubGenre/SubSubGenre.../Series/Title -- Author.epub. This helps me keep track of my many books. As a tree, it looks like this:
science fiction/ ├── fermi's progress │ ├── dyson's fear -- chris farnell.epub │ ├── descartesmageddon -- chris farnell.epub │ ├── planet of the apiaries -- chris farnell.epub │ ├── the phone job -- chris farnell.epub │ ├── fermi's progress -- chris farnell.epub │ └── fermi user manual annotated -- chris farnell.pdf ├── mind control -- finn, thomas.epub ├── monk and robot │ ├── psalm für die wild schweifenden, ein -- becky chambers.epub ├── the expanse │ ├── abaddons tor -- james corey.epub │ ├── calibans krieg -- james corey.epub │ ├── leviathan erwacht -- james corey.epub ├── the further adventures of sampson 39 -- chris farnell.epub ├── the murderbot diaries │ ├── home_ habitat, range, niche, territory -- martha wells.epub │ ├── rapport_ friendship, solidarity, communion, empathy -- martha wells.epub │ ├── all systems red -- martha wells.epub │ ├── artificial condition -- martha wells.epub │ ├── exit strategy -- martha wells.epub │ ├── fugitive telemetry -- martha wells.epub │ ├── network effect -- martha wells.epub │ ├── rogue protocol -- martha wells.epub │ ├── system collapse -- martha wells.epub ├── wayfarers │ ├── a closed and common orbit -- becky chambers.epub │ ├── galaxie und das licht darin_ ausgezeichnet mit dem kurd laßwitz preis 2023, die -- becky chambers.epub │ ├── record of a spaceborn few_ wayfarers 3 -- becky chambers.epub │ ├── the long way to a small, angry planet_ the most hopeful, charming and cosy novel to curl up with -- becky chambers.epub ├── zero gravity -- schuhmacher, nicole.epub
To manage my eBooks, I naturally use Calibre. This program stores all its books in an internal folder structure that cannot be modified, but it does have an export function that allows me to determine the structure of the export paths – meaning where the author and book title should appear in the path. There, I can also specify a genre or book series. For genres, I use the typical categories known from the book trade.
The book trade categories are structured in multiple levels, such as non-fiction, chemistry, organic chemistry. Unfortunately, I can’t enter complete paths in the genre field. So, I wrote a tool that takes the deepest category – organic chemistry, for example – and, after exporting, recreates the full path, including non-fiction and chemistry above it. It reads an Org file where I’ve stored my adapted book trade structure and adjusts the paths accordingly.
I then copy this folder structure to my eBook reader. The goal is to ensure that even eBook readers that don’t natively support stable folder creation can handle a hierarchical structure. Additionally, I can use my program for visualizing eBooks in such a way that a new shelf is started for each main category.
Documents
My documents take up about 1.5 GB. Here, I use PDF because it can be opened almost anywhere. The structure is ~/myfolderstructure/ISODATE_Type_Sender_Content.pdf. The ISO date at the beginning helps me sort everything chronologically.
The documents folder has the wildest structure. Essentially, I store scanned incoming mail there – these are the PDF documents that reside there. But I also store outgoing mail there, which is why you’ll also find text or Open Document formats.
This tree structure has grown organically over the past ten years. I use it in several places, both for directories and for folders in my mailbox. And it has largely proven itself. If you look closely, you can still see remnants of attempts to better organize everything. But wherever I work on the files, I’ve since reversed those changes.
As a tree, it looks like this:
. ├── Beruf │ ├── Bewerbung │ │ └── lebenslauf.odt │ ├── 2017-10-17_Abfindung.pdf │ ├── 2017-10-18_Arbeitsbescheinigung.pdf │ ├── 2025-08-17_Arbeitsvertrag.pdf │ ├── 2025-12-31_Zwischenzeugnis.pdf │ ├── 2026-04-23_Laufzettel.pdf ├── Familie │ ├── Familienstand │ │ ├── Hochzeit │ │ │ ├── 2003-11-15_Heiratsurkunde.pdf │ │ │ ├── 2025-09-12_Aufgebot.pdf │ │ │ └── 2026-05-29_Heiratsurkunde.pdf │ │ ├── Todesfall │ │ │ ├── Beerdigung │ │ │ ├── 2016-08-11_Sterbeurkunde.pdf │ │ │ ├── 2016-08-12_Angebot_Bestatter.pdf │ │ │ ├── 2016-08-18_Einäscherungsurkunde.pdf │ │ │ ├── 2016-08-25_Baumvertrag_a.pdf │ │ │ ├── 2016-08-25_Baumvertrag.pdf │ │ │ ├── 2016-08-25_RE_Hankemeier_Bestattung.pdf │ │ │ ├── 2016-08-26_Beisetzungsurkunde.pdf │ │ │ ├── 2016-09-30_ErbscheinMitVorblatt.pdf │ │ │ └── 2016-09-30_Erbschein.pdf
I will definitely write a separate article about this structure. It will explain in detail why I chose each folder the way it is today. This will also bring me closer to the topic of plaintext than is the case in this article.
Conclusion
For me, herding means taking responsibility for my digital collections. It’s not about just filling up storage space, but about consciously handling what I preserve. This gives me a good feeling and even saves me time in the end.