On Sat, Nov 23, 2013 at 9:41 PM, James M Snell <jasnell@gmail.com> wrote: > Yes, but the "they can just use 1.1" argument is being used consistently > to justify design choices for 2.0. That's just wrong. > I guess what I'm saying here is that yes, if you have an old device that really can't handle TLS (and given that TLS runs on some pretty modest devices, I'm skeptical), you can use HTTP/1.1. It's not a big deal. If you're so constrained that you can't handle TLS, then you probably aren't going to see much benefit from the other aspects of HTTP/2.0 either. But I don't want to hold up progress on HTTP/2.0 for everyone because of a very minority claim that legacy hardware performance is bad. Anyway, I think the summary I put forth above had a lot more meat to it than this minor, parenthetical comment about HTTP/1.1. It might be more fruitful to discuss that. Mike > On Nov 23, 2013 9:36 PM, "Tim Bray" <tbray@textuality.com> wrote: > >> Unconvinced. Life is full of trade-offs. Enhancing HTTP/2's privacy story >> is not without costs. When balancing costs and benefits, it's reasonable to >> consider that the costs may be mitigated by the continuing availability of >> HTTP/1.1. It's not axiomatic that HTTP/2 has to be the single best choice >> in all situations, in order to be a success. >> On Nov 23, 2013 8:40 PM, "Matthew Kerwin" <matthew@kerwin.net.au> wrote: >> >>> On 24 November 2013 11:12, Mike Belshe <mike@belshe.com> wrote: >>> >>>> >>>> (that can still use HTTP/1.1 if they want to!) >>>> >>>> >>> I have to chime in here. I keep seeing this point mentioned on the >>> list. This is not a valid counterpoint, or a justification for adding >>> something to HTTP/2, or an excuse for ignoring someone's use-case for HTTP. >>> If anybody chooses HTTP/1.1 over HTTP/2 for _any reason_ other than >>> laziness or stubborn change aversion, then HTTP/2 has failed in its primary >>> purpose. >>> >>> -- >>> Matthew Kerwin >>> http://matthew.kerwin.net.au/ >>> >>
Received on Sunday, 24 November 2013 07:23:11 UTC