Re: [hybi] Extensibility mechanisms?

[Scott Ferguson <[e-mail]>]

Maciej Stachowiak wrote:
> You said it solved all the requirements, so I was interested to find out if this was really so. It looks like that is not the case with your current draft, although it does have some nice properties.
>   
You're right about that. I do apologize for the overreach.

The missing binary/framing/control/stream requirements are the issue I'm 
worried about, and wanted to show it was possible to satisfy with a 
simple protocol, creating a more formal writeup of the related framing 
discussions on the list. The handshake was tacked on only for completeness.

Basically a proof of existence, because those requirements were being 
written off as too difficult. (And because no one else was stepping up 
to do it and they're trivial.)

Also, some of the related discussions were not grounded in an actual 
protocol.  For example, Jamie Lockier's ideas to reduce round trips 
sound like they're impossible to reconcile with the cross-site scripting 
issues. In the concrete proposal, both you and Mike saw problems 
immediately, but I don't recall his ideas being dismissed in that 
thread. (Although it was a few months ago.)

-- Scott
> Regards,
> Maciej
>
>
>
>
>