Explore
New notes
Show Notes
Replying to npub1yxprssc
Thank you. I'm still improving and studying better the negentropy protocol. I converted the code from C++ to Rust in basically a day and half (I'm not a C++ dev so some thing could be converted badly) and are there some things to fix/improve. Like, I noticed an unstable performance with reconcilie_with_ids method with a set of 50000 items: it usually took 600 ms the final reconciliation but often it took 2 or 4 secs (executing exactly the same code). Thanks, I'll replace the String with a vector. BTW, I'm planning to release also bindings for some languages (Kotlin, Swift and Python) using UniFFI, like I have done for rust-nostr libraries.
0
0
0
0
Replying to npub1u8lnhlw
From my understanding (I can worng), range-based set reconciliation (negenteopy) are good with small set, while BCH-based (minisketch) works better with large set and provide error correction. Maybe npub1yxprssc can correct me or add some more info.
0
0
0
0