started working on a new 🤯 feature for my social network 3 days ago

made a small part, but even that took a lot of time

then i still need to do the ui

this week i’ll try to make a big article on the next 13 questions about multithreading in java

I will try to make the first working demo of the social network in 4-5 months.

What is a DAG? Arborescense (Part 2)

What do Make build scripts, Git history, the IOTA blockchain, and AI in TensorFlow have in common? They all run on the same foundation — DAG.

Article - https://selfuniversity.vercel.app/en/blog/en/arborescense_dag

finally made the authentication aspect for getting tokens etc via context and auth request through an annotation (for the backend)

completely switched to postgresql, fixed seg fault and other bugs in the ui (gtk) and more or less finished the desktop application itself

today started improving the security of my transcendental bridge, and will start finalizing its database and implementing my ideas there

and yeah, i also spent a long time making a dag-replicated distributed ring buffer (for wal etc)

and started working on a basic editor ✅

made a couple of new sniffers (for tcp, icmp), took a few days to make them.
i already had a few, will try to add all of them

also almost finished the entire gui, from there you'll be able to run everything needed — translator, editor, bridge, etc.

also a couple of days ago came up with an incredible idea on how to make an architecture with practically no servers, where almost everything will be stored locally, where in theory you'll be able to change a whole lot (who to show a post to, who to hide it from, and infinitely many other things)
it might also be impossible to block this network

i sketched out a small plan, seems like it should work.

but i still need to somehow connect all the ideas into one scheme, where all processes are logically structured

honestly the idea is very interesting, i'll try to make it happen

so basically, i've more or less sketched out my architecture

started implementing it today, in theory everything should work

but we'll see what happens

many editor features will actually be implemented through this architecture, but for now i'll try to make a couple of features for local network

i'll have another powerful feature for filtering messages, words, and everything else, since almost everything is stored locally, there won't be any anonymity issues

finally more or less finished my in-memory db.

added pointer compression, a segregated fits allocator, thread local storage
improved security, fixed segfaults, added optimizations (cache line alignment, simd, etc.)

tested the db on a small benchmark — handles 1000+ threads, scales up to 1.6gb+ in the pool (and 5700 blocks), reuses memory correctly, 0 errors in valgrind, ran under load for about an hour with no crashes or leaks

came up with a rough way to make popular traffic analyzers not notice certain packets.
this will definitely be the foundation of the entire social network architecture

also birthed an idea for a subproject within the social network.
several algorithms, architectures, etc. will be implemented based on it — it could easily become a startup if polished to perfection

i've already sketched all of this out schematically (but only part of it)
maybe later i'll go into more detail about it

and finally — a perfect one‑year streak on leetcode 🎉