replicatr is now running stable... logs are showing a few odd things that suggest i can probably make it faster, but it's not crashing anymore, so i think i've eliminated all the unsafe array accesses and race conditions
gonna revise all the badger database transactions tomorrow, shrink down the duration they spend open and put the RWMutex back in place
i'm quite pleased actually to figure this out... makes me think i might have some ideas about how to actually do immediate mode GUI with Go with max framerates... it's all about building a good DSL that wraps the essentials in easy to read scripty style code... without costing anything in interpretation or disambiguation
looking forward to making #layer2 a meme on #nostr
relays that are just dumb caches backed by big, fat distributed databases
i especially want to have a crack at making PnyxDB a #layer2 for nostr
200+ replica, weakly consistent low latency distributed database... seems to me like a perfect fit for #nostr, and then maybe a #layer3 becomes possible with crazy stupid big databases like the one that #arweave has got running (gonna be working on a project for them soon, related to this, and this is what is in my mind... was my first instinct - currently finishing up a #layer2 for Internet Computer Protocol and next up, #layer3 with arweave
none of these people involved in these aspirational shitcoin projects aiming at massive decentralised storage and compute really have a clue what tehy are doing, based on my understanding of DST theory, but i'll take their money and build the thing they don't realise they want built
cos that's how i roll
Showing page 1 of
1 pages
