Like maybe you can take your files with you, but you can't take your meaning.
efficient frontier might be: What file formats have the most semantics?
Conversation
My sense is that the connection between semantics and interoperability is a lot more ambiguous, and there is no 1:1 connection.
2
3
I like 's insight here, that interoperability is emergent, and emerges around product-market fits. These are heavily contingent, and I think can span a range of tight, loose, open, closed semantics.
Quote Tweet
“Interop is a process not a thing.” @stefanomaz
Show this thread
2
1
Absolutely. BRB adding “emergent semantic slop” to my quarterly OKRs to test if I get fired 🙃
1
2
There is a gradual line between "Files with proper metadata" and "a database". A database has the potential to be much more seamless but also much more locked down and inflexible. Files are the QWERTY keyboards of data, they are just good enough
2
2
5
This is interesting. What if the file was a SQLite database?
2
1
Each file would have varying schema. Or: calling a database a "file" does not make it not a database
Quote Tweet
Replying to @ben_mathes @gordonbrander and @stefanomaz
There is a gradual line between "Files with proper metadata" and "a database". A database has the potential to be much more seamless but also much more locked down and inflexible. Files are the QWERTY keyboards of data, they are just good enough
1
2
however, maybe you are saying something unintuitive or subntle that I'm missing. Make the case?
1
I'm mostly thinking this: twitter.com/andy_matuschak
and also how SQLite is considered a long-term archival format for data by the US Library of Congress. sqlite.org/locrsf.html
Quote Tweet
Maybe someday automerge’s serialization format will be as durable and universal as SQLite’s, but in the meantime, it’s certainly not amenable to casual concurrent local access in some user script.
Show this thread
1
1
I hear you, but it sounds like orthogonal concerns. SQLite is perfectly fine as an archival standard but that's separate from the desire for the data to also be grouped, share and manipulated by a wide range of people and processes.
1
2
Point well made, and well taken.
I've read your post so I know where you're coming from. I suppose I'm just playing around with this idea that we need not worry so much about designing file formats — which is difficult! — if we just use SQLite + a plain txt schema + versioning.
(And lump that all together in a zip archive-like thing? Oh dear).
1
Hey, I'm all for results over dogma. If you can describe (roughly) how this could work for a wide range of use cases (e.g. audio, video, spreadsheets, rich text) I'm all for it.
3