Samuel
I don’t have specialist knowledge here but I’d imagine you’d want some sort of metadata (description etc) and then just an array of blobRefs
2 5
alice
or i can put atproto on huggingface bsky.app/profile/alic...
alice@alice.mosphere.at· 27 Jun
a nice byproduct of my backfill is i'm now the proud owner of one of the largest dataset of bluesky posts, roughly all 2.6 billion of them, with all their metadata from the post lexicon living in 552GB zstd'd parquet files, which for now live on a simple hetzner storage box
1 1 7
Emily Hunt
That'd be really cool! I think the right venue for this would be the IETF or similar, though. It would take a LOT of input to make something generic enough to work. Also, @vmx.cx's matadisco records already contribute here too
1 3
Denis
How does hosting stuff on atproto work? Won’t it be centralized?
1
Samuel
you put it own your own server, atproto serves as the distribution layer
1 5
Al Merose
This is very similar to an idea I’ve had — I think the designs would be similar on atproto to torrents: github.com/zarr-develop... Furthermore: decentralized, replicated stores would be essential for many domains of science, especially climate science
A Torrent ZarrStore · zarr-developers zarr-python · Discussion #3036
Coming back from Cloud Native Geo, a major theme was how to build resilient data infrastructure (or ecosystem). A common idea that came up in many discussion groups was could we somehow store scien...
github.com