Re: Giving the Framing Layer a real name

[Mark Nottingham <[e-mail]>]

To me, a distinct name of a major, separable part of a protocol is only good spec hygiene; in HTTP we already have "representations" (nee "entities"), "resources" and so forth. They allow people to talk about different things that are happening with clarity; right now, people are using the term "HTTP/2.0" very, very sloppily, and that's a concern. "It means what I mean" is not a great basis for communication.

I don't think that changing the name will add any more complexity unless we allow it in based upon assumptions that it therefore *has* to be completely separable, and I'm explicitly not bringing that to the table.

Cheers,


On 28/02/2013, at 5:56 AM, James M Snell <[e-mail]> wrote:

> No one said that subprotocols would not still be possible on top of
> the same framing mechanism. Mark's question was about naming only, not
> design. There is still a good design separation between the HTTP
> semantics and the framing layer, and the framing layer would still be
> reusable beyond just carrying http semantics.
> 
> On Wed, Feb 27, 2013 at 10:51 AM, Roberto Peon <[e-mail]> wrote:
>> Then you can't do websockets, etc or whatever other protocol (maybe video?,
>> who knows) the web platform decides to do in the future on the same
>> socket/session.
>> 
>> That would be a poor tradeoff.. and for what gain?
>> What is the additional complexity of having the framing allow for non HTTP
>> semantics?
>> -=R
>> 
>> 
>> On Wed, Feb 27, 2013 at 10:13 AM, James M Snell <[e-mail]> wrote:
>>> 
>>> +1
>>> 
>>> On Feb 27, 2013 10:03 AM, "Martin Thomson" <[e-mail]>
>>> wrote:
>>>> 
>>>> On 27 February 2013 00:49, Mike Belshe <[e-mail]> wrote:
>>>>> My vote would be to nuke the layering and instead fold section 4 (the
>>>>> HTTP
>>>>> Layering) into the appropriate sections of the framing layer.  Trying
>>>>> to
>>>>> make these generic frames seems like a distraction, and it would be
>>>>> simpler
>>>>> for folks to read if these were just the basics of HTTP framing.
>>>> 
>>>> I said as much to mark in private: the framing layer is for HTTP.  A
>>>> name implies that it might stand alone.  That's not going to be true.
>>>> The work to make it true is not worthwhile either.
>>>> 
>> 

--
Mark Nottingham   http://www.mnot.net/