- From: Poul-Henning Kamp <phk@phk.freebsd.dk>
- Date: Sun, 17 Nov 2013 08:30:16 +0000
- To: Amos Jeffries <squid3@treenet.co.nz>
- cc: ietf-http-wg@w3.org
In message <528829EB.2010403@treenet.co.nz>, Amos Jeffries writes: >On 16/11/2013 9:06 a.m., Nicolas Mailhot wrote: >> >> Le Ven 15 novembre 2013 18:31, Roberto Peon a écrit : >> >> I'm not convinced at all using a new port is infeasible. Yes, it would be >> devilishly hard for a new corner-case protocol. But the web as a whole is >> something else entirely and a new way to access it won't be dismissed so >> easily. The same exact thing was said about IPv6, almost 20 years ago. This argument only applies if you s/new/much improved/ In my not at all humble opinion, the current draft is nowhere near that. Ratifying the current fraft on port 100 would be pointless. >Exactly the same arguments apply to staying with port-80 in the first >place. There will be initial pain, but things will straighten out >eventually. The two thing which have started to worry me about staying on port 80 is: A) HTTP/2.0 benefits will only materialize after a couple of RTT's because of the upgrade kabuki dance. B) HTTP/1.1 infrastructure will be compromised while people iron the bugs out of their HTTP/2.0 code. I'd say that our best shot is: 1) Make HTTP/2.0 simpler to implement than HTTP/1.0 (by eliminating and simplifying the semantics, and not adding new complexity.) 2) Make HTTP/2.0 faster than HTTP/1.0 (encoding + mux + pipeline + shaving cookies so all requests fit a packet) 3) Avoid time-wasting undecidable political issues (aka: encryption) 4) Define a new port *AND* an upgrade method. 5) Sit back and see if people like it. -- Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 phk@FreeBSD.ORG | TCP/IP since RFC 956 FreeBSD committer | BSD since 4.3-tahoe Never attribute to malice what can adequately be explained by incompetence.
Received on Sunday, 17 November 2013 08:30:41 UTC