- From: Martin Thomson <martin.thomson@gmail.com>
- Date: Wed, 19 Feb 2014 14:40:06 -0800
- To: Yutaka Hirano <yhirano@google.com>
- Cc: Roberto Peon <grmocg@gmail.com>, HTTP Working Group <ietf-http-wg@w3.org>
On 19 February 2014 06:34, Yutaka Hirano <yhirano@google.com> wrote: > Does your "a way to map the semantics efficiently" mean Plan C? Or you have > another idea? I don't have a concrete plan, but here's what I believe the semantics you want to preserve are. From there, I think that you will find the delta is minimal from what is already in the HTTP/2 spec: WebSockets produces an ordered sequence of messages. The messages are exchanged between a client and a resource that is identified by a URI of either ws: or wss: scheme. These messages are of arbitrary length. Messages can contain either binary data or UTF-8 text. You'll note that I don't mention frames at all here. >From that description, the only thing that HTTP/2 can't provide is the binary/text distinction. If you add compression, then that's another item to add. I think that what Roberto describes with HEADERS + DATA (and END_SEGMENT) is workable for this.
Received on Wednesday, 19 February 2014 22:40:34 UTC