Question for Kin: Real time protocols are an increasingly important part of the web. To that end, what predictions can you make about the state of WebSocket APIs in three to five years, and how confident are you in those predictions?
I’d say that Websocket APIs have done well in gaining mindshare when it comes to delivering real time data and content. Adoption will continue, especially in financial API realms. It enjoys healthy adoption which people see and will emulate.
However, in many cases people will realize it was overkill for what they needed. That they didn’t need bi-directional streams, and that one direction streams would have been sufficient. Giving rise to support of other standards like Server-Sent Events (SSE), which operate within the HTTP protocol, making for easier client adoption, and more precisely delivering what people want.
This cannibalization of Websockets will continue through the evolution of the HTTP/2 spec, taking advantage of the improvements available for one-direction, and multi-direction streams, along with concurrent threads. With speed and performance gains introduced, along with leveraging binary exchanges, eventually websockets won’t be seen as attractive.
It isn’t going anywhere, but it won’t stay the coolest kid on the block, and fall out of favor in the next decade.