Re: [hybi] Sub-Registry Protocols that are not exclusive to WebSockets

Hello Scott,

On 2016/04/29 04:15, Scott Morgan wrote:
> Hi All,
>
> I just wanted to point out that there is not always a hierarchy between
> protocols.

That's a valid point, at least in theory.

> For example the protocol I am working on ASBP (Asynchronous
> Services Bus Protocol) is intended to be layered on top of any of;
>
> WebSockets
>
> Http/2
>
> Http 1.1
>
> Http 1.0
>
> https://datatracker.ietf.org/doc/draft-adligo-hybi-asbp/?include_text=1
>
>
> In the ASBP case I would guess that I should register ASBP as a WebSocket
> sub protocol, as well as a sub protocol for all of the above protocols.

There's no registry for protocols over HTTP. My understanding is that 
this is essentially because the protocol/architecture (as explained e.g. 
by REST) of HTTP is essentially fixed, although of course it's always 
possible to find ways to bend or tweak it.

Also, I don't think it makes sense to have different mappings for HTTP 
1.1 and 1.0. The same most probably applies for 1.1 and 2.0, because 
they are supposed to be semantically identical.

On the other hand, as far as I understand, the WebSockets sub protocol 
registry was created because WebSockets is not much more than TCP for 
the Web (the main difference to TCP being the message concept). So it 
looked prudent to create a registry to reduce the possibility of 
confusion one layer higher up.

However, because WebSockets is usually used between two consenting 
parties (the browser which downloaded a Web page from a server, and that 
server), the Websockets Sub-Protocol Registry has been rather 
underutilized, and there are many usages of WebSockets without any 
explicit subprotocol.

Regards,   Martin.

> On the other hand perhaps the title 'Websockets Sub-Protocol Registry'
> should be changed to 'Protocols Supported Over WebSockets' in order to
> facilitate clarity.
>
>
> Cheers,
>
> Scott
>
>
>
> _______________________________________________
> hybi mailing list
> hybi@ietf.org
> https://www.ietf.org/mailman/listinfo/hybi
>