Simply by making use of "name=value" tags; moving them to the same level as a file or foldername; and using the filesystem directly as database.
TL;DR:
Fact is, there's only 2 possibilities:
- I've overlooked/underestimated something.
- Or someone else has overlooked/underestimated something that I can figure.
What I am suggesting is to use the existing FOSS-licensed technologies designed for storing and managing big data - and really letting go of files-in-folders with excel sheets and incompatible databases and embedded metadata to a data design making use of the Object Storage environment's capabilities.
Putting meta with the data. Where it belongs. And draw relationships right with the technically same Objects: Support (de)referencing any kind of "link".
If I'm correct, then I may introduce a change for anyone dealing with digital data, comparable to going beyond 8.3 uppercase-only ASCII characters in a filename (1995), to Object Oriented programming and responsive design: Using a common filesystem. 😄️
Such filesystems and applications already exist for over 10 years, so a first prototype can be created in less than 12 months.
"AHAlodeck by August" - so to say.
Take a deep-dive, reading also the additionally exciting Introduction.
"Great Scott, Thomas! 😱️😎️
I wrote these notes when I was travelling to the future and back to ... Our current "situation" with, eh... "digital files". 😕️ And metadata. DAM!
They solved it! In the future!
Listen here: I wrote down as much as I could, but it's not all. It was so simple.
There was a slogan, that everyone knew. It was: "Put your meta where your data is"
I may commit or edit later, whenever I have time.
It's so simple. Read on..."
These notes are regarding a technical draft for using Object Storage Filesystems beyond "just storing files", by moving metadata and database needs to the filesystem's "tagging" feature.
- What if you could put any metadata with your file?
- What if filename/foldername was "a tag option" and nothing more?
- What if you could create and handle database/catalog/DAM entries in your regular file manager/browser?
- What if that would work with any application or digital environment?
- What if that was all FOSS-licensed?
- What if that was well-supported by small and big companies?
Sounds pretty crazy, but I'd like to challenge building a working prototype.
- Provide an S3 compatible, local-only (but network-able) Object storage (MinIO?).
- Copy cultural heritage object files - and attach their metadata as-is.
- Copy existing pre-structured or embedded metadata values to the "name=value" space.
- Query that storage by its "Objects and their metadata" in MongoDB.
- Have a web-browser-based application to display and interact with these objects:
- regular file manager functionality.
- display relationships (as graph? node network?)
- plain viewer for basic formats (think IIIF)
The documents are currently merely a notepad for ideas, as well as may contain concrete technical implementation considerations or remarks that may be considered when implementing certain options.
If you like you are welcome to engage in this adventure:
You may see the whole digital world around you in a very different way. That may be the real, awesome-digital 21st century! 🌟️
It's a low-hanging-fruit, yet endless rabbit-holes and possibly a never-ending story. But a good one definitely.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
