hckrnws

There's probably good advantages to a well tuned more bespoke multiplexing protocol toolkit…

But it sure makes me crazy crazy sad that QUIC and HTTP3 are still rarefied technologies with little mainstream adoption. The hyperscalers all seem to have a library or two they've made… https://github.com/mmmarcos/awesome-quic?tab=readme-ov-file#...

…but there's still no mainstream http3 that works well with a http framework of any kind that one can just use. Hyper's h3 kicks along slowly. Node.js has been beleaguered by openssl's endless beating around the bush reinventing how they want to do quic, ignoring input. But that's finally starting to move. https://github.com/hyperium/h3 https://github.com/nodejs/node/issues/57281

Muxio was able to define itself & an inpl pretty fast, and performance & usability are probably both medium or high. That ease is nice. But I still want to see more social excitement and effort on quic. There's some pretty good performing early products like Apache Iggy for quic based messaging, which is nice. But quic feels so amazingly general purpose, is such a wonderful upgrade for all manners of application transports, that could replace so many bespoke solutions. https://github.com/apache/iggy