4 ms·

Nitpick - It's HTTP/2 not HTTP/2.0 We've all learned from the failure of SNI and IPv6 to gain widespread adoption. (Thank you windows xp and Android 2.2) HTTP/

by cflat 12y ago

Nitpick - It's HTTP/2 not HTTP/2.0 We've all learned from the failure of SNI and IPv6 to gain widespread adoption. (Thank you windows xp and Android 2.2) HTTP/2 has been designed with the absolute priority of graceful backward compatibility. This creates limits and barriers on what you can do. Transparent and graceful backward compatibility will be essential for adoption. I agree, HTTP/2 is Better - not perfect. But better is still better.

nly 12y ago
Not that I believe SNI and IPv6 are failures, but HTTP/2 faces exactly the same failure case as IPv6 (the lack of adoption due to HTTP/1.x being 'good enough').
- dragonwriter 12y ago
  > Not that I believe SNI and IPv6 are failures, but HTTP/2 faces exactly the same failure case as IPv6 (the lack of adoption due to HTTP/1.x being 'good enough'). HTTP/2 isn't really like IPv6 in that fewer people need to act to adopt it -- if the browser vendors do (which they are already) and the content providers do (which some of the biggest are already), then its used. Its specifically designed to be compatible with existing intermediate layers (particularly when used with TLS on https connections) so that as long as the endpoints opt in, no one else needs to get involved -- and with one of the biggest content providers also being a browser vendor who is also one of the biggest HTTP/2 proponents... IPv6 requires support at more different levels (client/server/ISP infrastructure software & routers, ISPs actually deciding to use it when their hardware/software supports it, application software and both client and server ends, etc.) which makes adoption more complex.
  - ethbro 12y ago
    > > Not that I believe SNI and IPv6 are failures, but HTTP/2 faces exactly the same failure case as IPv6 (the lack of adoption due to HTTP/1.x being 'good enough'). > HTTP/2 isn't really like IPv6 in that fewer people need to act to adopt it -- if the browser vendors do (which they are already) and the content providers do (which some of the biggest are already), then its used. I, for one, welcome the HTTP/1.x+2 future of 5-10 years from now. (Obligatory http://xkcd.com/927/ http://xkcd.com/927/ )

by cflat 12y ago

nly 12y ago

Not that I believe SNI and IPv6 are failures, but HTTP/2 faces exactly the same failure case as IPv6 (the lack of adoption due to HTTP/1.x being 'good enough').

dragonwriter 12y ago
> Not that I believe SNI and IPv6 are failures, but HTTP/2 faces exactly the same failure case as IPv6 (the lack of adoption due to HTTP/1.x being 'good enough'). HTTP/2 isn't really like IPv6 in that fewer people need to act to adopt it -- if the browser vendors do (which they are already) and the content providers do (which some of the biggest are already), then its used. Its specifically designed to be compatible with existing intermediate layers (particularly when used with TLS on https connections) so that as long as the endpoints opt in, no one else needs to get involved -- and with one of the biggest content providers also being a browser vendor who is also one of the biggest HTTP/2 proponents... IPv6 requires support at more different levels (client/server/ISP infrastructure software & routers, ISPs actually deciding to use it when their hardware/software supports it, application software and both client and server ends, etc.) which makes adoption more complex.
- ethbro 12y ago
  > > Not that I believe SNI and IPv6 are failures, but HTTP/2 faces exactly the same failure case as IPv6 (the lack of adoption due to HTTP/1.x being 'good enough'). > HTTP/2 isn't really like IPv6 in that fewer people need to act to adopt it -- if the browser vendors do (which they are already) and the content providers do (which some of the biggest are already), then its used. I, for one, welcome the HTTP/1.x+2 future of 5-10 years from now. (Obligatory http://xkcd.com/927/ http://xkcd.com/927/ )