The @qri_io project thinks of itself not just as heavy @IPFSbot, bu also libP2P.

They are dedicated to building a data commons.
"npm for data" @qri_io @b_fiive

Offline / local data, remote data, versions, etc.
" @qri_io accepts messy data, but not _invalid_ data". Usually CSV ingest, which "upgrades" to SQL queryable.

Schema inference, column names

Dataset document format is common - metadata, README, commit info
Git recommends files no larger than 100MB. @qri_io works on datasets that are much larger, IPFS is a good fit for versioning / content addressing, and block-level storage and sharing.
Biggest difference between data sets and code in version control, is that data sets get much much bigger -- GBs in size.

@qri_io came up with logbook -- append only log, informed by operational transforms / CRDTs. Kind of like git reflog, but tuned for datasets
Built in structured data diff'ing support, including in @qri_io desktop
Wrapping up now with some discussion. Excited to see where @FISSIONcodes and @qri_io can work together on shared open source specs and code.

See sprint boards: https://github.com/orgs/qri-io/projects/31

And RFCs: https://github.com/qri-io/rfcs 
Here's a video of @qri_io Desktop

Electron app, local IPFS node
Because we're already getting pings on this, yes the main presentation was recorded, and we'll post it later today with a link in this thread.

Thanks @ricogardaphe and @b_fiive for joining us today!
You can follow @FISSIONcodes.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: