1. 2

  2. 1

    A sort of post-mortem of the HTTP/1.1 process. Not too taxing to read in one go, and there’s a lot to be learned still today, about HTTP and about protocol work in general.

    Would love it if somebody could find the A/V of the talk too, but the slides work pretty well in isolation.

    The predictions on further updates to HTTP are interesting to compare and contrast with what happened with HTTP/2. One company being in charge of a huge site, a huge infrastructure and a large share of the browser market changed a lot of the assumptions. But not all.

    When it comes to the many suggestions for what could be better with HTTP, I’d say most of them are today being solved by other protocols entirely, not by updates to HTTP. Or not solved at all because worse is better.

    TL;DR, from slide 57, basically the justification for the title:

    Summary and Conclusions

    Design by evolution

    • Leads to adequate design
    • But not optimal design

    HTTP is

    • Definitely not optimal
    • Probably adequate