Isn't it a result of wrong assumptions based on developer's experience with employer-provided high-end laptops and workstations? Like "oh, i'm gonna to add this feature and that tweak here and it will be good enough because any PC (around me, in the office) has 16 (ok, at least 8) Gb of RAM AND 100+ Mbps Internet connection"? And the other side has the point as well - everyone remembers famous Bill Gates quote "640KB should be enough for anyone". By the time the protocol gets widely adopted - nobody would care and majority of systems will have 8 Gb of RAM. What a waste!
On Fri, Mar 31, 2017 at 9:04 AM, <bytevolc...@safe-mail.net> wrote: > On Fri, 31 Mar 2017 12:14:34 +0200 > Reyk Floeter <r...@openbsd.org> wrote: > >> Isn't QUIC the hot new thing now? It is UDP, so Google can reinvent >> TCP and turn even more of their browser into an OS-replacement ;) > > Oh come on now, how else will Google be able to claim they are > inventing or innovating? What will they say at the meetings with their > shareholders if they can't reinvent the wheel? > > Perhaps Microsoft will join in and release IPv6 Service Pack 1. > >> Seriously, there are benefits of implementing HTTP/2, and it would be >> an interesting exercise to do so, but it is also adds many problems >> and some complexity. > > The benefits are there, but I feel it encourages lazy and disorganized > web development, leading to stupidly bloated and inefficient sites, and > requiring the latest stupidly bloated and inefficient browsers. > > Back in the dial-up days, I remember web pages without much bloaty > rubbish. They had to be fast, because networks were high-latency, > low-bandwidth beasts. > > More efficient protocols are great, but it's a bit like increasing CPU > speeds and RAM; fantastic, but not if the software starts becoming > bloated beyond your wildest dreams. > >> >> So: maybe. >> >> Reyk > -- Best regards, Yury.
