- From: Roberto Peon <grmocg@gmail.com>
- Date: Thu, 12 Dec 2013 14:27:29 -0800
- To: Amos Jeffries <squid3@treenet.co.nz>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
- Message-ID: <CAP+FsNd-YR=TA96djAd+yMs_Xt_C5jUqMWmwyhoyDva=vHz1_g@mail.gmail.com>
Protocol negotiation is extremely useful. You can happily translate any time I say "ALPN" to mean "protocol negotiation using the registry used for ALPN" Don't worry so much :) -=R On Thu, Dec 12, 2013 at 2:11 PM, Amos Jeffries <squid3@treenet.co.nz> wrote: > On 2013-12-13 07:41, Roberto Peon wrote: > >> These things are all possible with ALPN. >> > > ALPN, ALPN, ALPN! I'm hearing a *lot* of dependency being placed on ALPN > capabilities. > Are we designing HTTP/2 as a protocol or as a set of ALPN extensions? > > I rather think we need to avoid locking so much HTTP/2 feature dependency > into an *optional* extension feature of a protocol which only covers a > sub-set of the HTTP use-cases. ALPN provides a performance optimization for > starting HTTPS (both 1.x and 2.0), very little else. > > If something does not work on HTTP/2-over-TCP without an intermediary > protocol layer, even inefficiently, then it does not meet requirements > implicit in section 2 of the HTTP/2 specification. > > " > 2. HTTP/2.0 Protocol Overview > > > HTTP/2.0 provides an optimized transport for HTTP semantics. > > An HTTP/2.0 connection is an application level protocol running on > top of a TCP connection ([TCP]). The client is the TCP connection > initiator. > " > ... *TCP* ... no mention of mandatory TLS there, or anywhere else in > section 2.*. > > Please stop assuming non-TCP layers underneath HTTP/2. > > Amos > > >
Received on Thursday, 12 December 2013 22:27:56 UTC