What’s wrong with how things are?

I do professional “Import/Export” business.

• Of (digital) data.
• Like all the time.
• And it’s getting less funny.
• And extremely expensive

Why?

Good question.

Generic handling of “Digital Meta/Data”?

1. file manager?
2. spreadsheet?
3. database?
4. catalog?
5. schema?
6. code?
7. and?

Wait. What?

No more files and folders?

That’s the one thing that we all really understood and manage since 50 years now. Wow. And you want to take that from us?!

Better: Extended Files and Folders

Imagine you wake up, being responsible for one simple job:

Consolidate all digital collections - and their management systems in a way that runs smoothly, efficient - and flows. For more than 10 years without need for an update or change.

And all you get is a default file-manager. And all other applications that work with “regular files”.

Relax, “There is no spoon.”

• There are no files.
• There are no folders.
• Just blocks of “data” on a storage.
• And hopefully enough metadata to index and find them…
• And enough virtual space to be saved.
• your filename and path are just metadata fields
• your filename and path are just metadata fields
• your filename and path are just metadata fields

We’ve gotten used to a Digital Playmobil world.

Bleu or Rogue?

I’ve just given you a 4th dimension:

• Imagine you could “checkbox” zero-or-more folder-paths.
• To “put a file” in multiple folders (context)
• Imagine you could have many filenames/paths - on one file.

Related Metadata

• your filename and path are just metadata fields
• your filename and path are just metadata fields

… (+) add more key/value pairs as you like.

What would that UI-feel like?

• Think browser bookmark handling.
• Think emails.
• Think any messenger.
• search/query/filter on anything.
• right-click-edit-metadata on anything.

Bookmark Manager

• A bookmark is “a link with metadata”.

Now just “save” it.

Did you need to:

• Give it a filename?
• Select a folder?
• Add tags?

Where’s the catch?

• Do you have enough “space” to save it?
• Do you have permission to do so?
• Does your application give you means to do so?
• Do you know enough about it - so you can find it later?
• …

Annotation & Context:

• Timestamp of creation is saved by default.
• It has a “title” information.
• It has a “link” information.

Imagine this being your new “LEGO” piece of data/information.

What if we want more?

• More data fields (key/value)?
• Relationships?
• Got a database?
• Got a spreadsheet?
• Got a format+tool to ‘embed’ it?
• Got a “schema”?
• Got migration/import/export plans & resources?

Want more than 8 characters?

Before 1995, it was “8.3 letters” ASCII-only. UPPERCASE. Do you (still) care?

21st Century Computing

• simply “save” your data?
• simply “tag/annotate” your data sufficiently to find/use it later?
• right-click-edit-metadata on any data-object in any format, at any time?
• keep meta+data reliably together?
• have versioning, multi-tiering, geo-distribution, network-scaling, failover handling, etc. all included by default?

Sounds like: The Cloud?!

The tech-stack designed and used for “Big Data” was to overcome existing files-in-folders limitations.

• Since the early 2000s.
• A lot has happened since then.

GLAM data /is/ Big Data now.

• Because our “normal” now was their “big” in 2003.
• Data is handled very differently, but noone told us…
• We’re all dealing with growing numbers of “Related Data Objects”.
• And mixed “collections”.
• On mixed devices/environments.

From Playmobil to LEGO.

Now add…

• Arbitrary key/value fields.
• Links between Objects (Relationships).
• UIs to deal with these basic building blocks.

Big Data “features”

• Schema evolution (Iceberg)
• Object storage (Swift)
• File versioning (Parquet)
• …

Collection handling?

• Ingest?
• Access?
• Merging?

Memory Institutions

• Need digital storage (memory)
• Generic “big data” stack scales from small-to-global
• Allows to become blueprint for a memory-network.
• Could serve the public with better-than-cloud services.

And all of this…

• faster
• easier
• cheaper
• with less-updates/migrations necessary

Than how we do it today.

Wait. What?

No more files and folders?

Sorry for being over-dramatic. I’ve just recently discovered that I am allowed to admit I am a: hyper-sensitive-indigo-rainbow-warrior-energy something.

And therefore we need to be clear: I do NOT like working with UUIDs.

UUIDs? Love the idea, hate the looks.