What do people think about the use of Slack for this group? What would people prefer ideally (some other existing tool/product, paid slack, something entirely new specifically for this type of group, ...?)

Hah: • 2018-12 at the dawn of the Slack (You have to read upwards from there; we weren't disciplined about using threads back then.) • 2019-05 Steve Krouse's note • 2020-04 survey with a question on the subject • 2024-06 (100+ comments!) conclusion • 2026-05 trying out Colibri on ATProto

Thank you for the pointers! reading.

I forgot we concluded we should build our own one. I was thinking with the atproto replication that it would give us some flexability to try out different skins over a Slack model without having to commit to a single solution initially. I do think an atproto dataplane is very consistent with most of our values. the (extremely vibed) technical documentation for the atproto data is here which I think it useful for trying a view out (and you can validate against the working Colibri viewer).

hello from ATProto (not slack) :p

I don't think anybody looked at the mock up. I made it work better now. It syncs the full archive (~50Mb) into IndexDB and offers text search, as well as access to the wiki and YouTube videos in one place. Its just an idea I am not wed to it. Its not very bold conceptually, maybe people can think of how we could be bolder. tomlarkworthy.github.io/lopebooks/notebooks/Feeling_of_Computing.html

It seems to have a lot of useful capabilities and it seems easier to browse than slack. It might be my computer, but it felt slow (as in things were updating visibly on the page over several seconds so I wasn't sure when it was settled). The annotation functionality is very interesting to me - what's the vision for that?

yeah thats the sync coming down, you can only download 100 records at a time on ATProto, and there is 100k of them, and throughput is rate limited. There are 2 classes of problems • 1. glitchy sync: I was trying to do something and the UI re-rendered breaking my UI action • 2. unavoidable sync: historical content was appearing in channels on first use of the app as it streamed in If its 2 then that is expected. Everything is cached so its a 1-off cost until you are up to date. It backfills latest first so once the channels have a page's worth of historical data you shouldn't notice the sync anymore even though it might only be 10% the way through the backfill. The benefit is near instant substring querying over what you have so far.

yeah it was #2. My observation is the sync seemed to keep starting from 0 on refresh until it "finished" once

yeah but then its off-protocol and a special bespoke server to babysit. I want to minimize that kind of thing, its not decentralized. I like that anyone can make this pure clientside, the backend is brought to the front end with a database inside the browser (IndexDB). ATProto is the data store. You can run that app as a local file and hack on it personally.

ooh nice, super handy! I was imagining it would enable granular replies within the slack/atproto discussion threads

well thats the kind of bolder conceptual thinking I like! yeah you could definitely have spatial and fine grained replies

and collections i.e. lists of posts that exist in a different hierarchy

the brilliant thing about ATProto is that this is possible, you can create your own schema (lexicon) and have your own domain model of how to point at other records even in 3rd party lexicons. The app would then be just rendering those data annotations. Its a good idea. Quite busy ATM for the next few weeks but its a good idea I like a lot. I also thought it be cool if the annotations were multiplayer in this particular app.

Busy as well the next few weeks but happy to riff on these things. I'm a bit out of date on ATProto and such for a use case like this so I'll need to research a little. In my current multiplayer annotation system I have my own source of truth and do plan to federate the different representations to nostr, activitypub, atproto, etc. - my main gap is understanding what validation and authorization you can do on ATProto or if you're just supposed to handle data quality and moderation etc in your viewer if ATProto is the source of truth