stick around for tweets about scatterpad.com, my magnum opus. i also made streams.place a couple years ago.
joodaloop.com/riffs/collab...
joodaloop.com/riffs/collab...
minimal spreadsheet editor (made mostly by @ampcode.bsky.social) with me directing and choosing the tech stack:
- @solidjs.com
- @automerge.org
- partykit.dev
works without logging in, new spreadsheet at each URL, very similar to what @tldraw.com does for it's pages
minimal spreadsheet editor (made mostly by @ampcode.bsky.social) with me directing and choosing the tech stack:
- @solidjs.com
- @automerge.org
- partykit.dev
works without logging in, new spreadsheet at each URL, very similar to what @tldraw.com does for it's pages
minimal spreadsheet editor (made mostly by @ampcode.bsky.social) with me directing and choosing the tech stack:
- @solidjs.com
- @automerge.org
- partykit.dev
works without logging in, new spreadsheet at each URL, very similar to what @tldraw.com does for it's pages
- me, logging out of twitter again
- me, logging out of twitter again
i have faith in @ryansolid.bsky.social though, the time slicing stuff is great
i have faith in @ryansolid.bsky.social though, the time slicing stuff is great
it will not slot into existing backends like all the other sync engines are trying to, it will force a very particular way of writing server code
but it will be a good way
- snapshot materialisation
- subscription sets
are all it should need
it will not slot into existing backends like all the other sync engines are trying to, it will force a very particular way of writing server code
but it will be a good way
- the IndexedDB layer (probably borrow the KV-store implementation from the replicache.dev codebase)
- the in-memory query layer (would be nice to have some kind of IVM layer, based of d2ts)
- the IndexedDB layer (probably borrow the KV-store implementation from the replicache.dev codebase)
- the in-memory query layer (would be nice to have some kind of IVM layer, based of d2ts)
but with async local-first it’s the perfect low-cost fit, simply shuffling data along slowly
but with async local-first it’s the perfect low-cost fit, simply shuffling data along slowly
everyone i know is very passionate about it but i almost never want to open one app's data in another one
and when i do, exporting to a standardised format is what i really care about
everyone i know is very passionate about it but i almost never want to open one app's data in another one
and when i do, exporting to a standardised format is what i really care about
→ make multiple todo lists in scatterpad.com
→ every morning, pick tasks from across them to do that day
→ do them
→ make multiple todo lists in scatterpad.com
→ every morning, pick tasks from across them to do that day
→ do them
write path through event log, all mutations are semantic, idempotent and affect a single item
read path through database objects, using the event log as a way to only send back objects that have changed
write path through event log, all mutations are semantic, idempotent and affect a single item
read path through database objects, using the event log as a way to only send back objects that have changed