Re: WebSocket2

On 2/10/2016 10:21 a.m., Van Catha wrote:
>> How something (registeration) which is just "Expert Review" can be
> considered
>> to be change of actual HTTP/2 specification ?
> 
> Well in the case that it will provide no resistance. The only possible
> value of adding a Setting
> I can see is to advertise what the headers currently advertise (is
> WebSocket2 available, can we compress, other possible other things).
> 
> Keep in mind different paths on the authority can use different compression
> methods.
> 
> The resistance I see is that it will have to be "Expert Reviewed" before
> being included, and a part of that 16bit space would have to
> be reserved for WebSocket2 extension settings. First, I doubt it would
> happen in a reasonable time frame and second

I believe the relevant expert(s) are reading this thread already. So if
a clear need for such extension is found and a definition document
created (eg the WebSocket2 RFC / Draft) it wont take much longer than
getting that spec defiend to use it in the first place. :-)

Implementation rollout will be what it will be. Regardless of any Draft
review.

> I just do not see the necessity.  We have custom headers in HTTP/1.1 for a
> long time and no one has had a problem with it.
> 

That is the more likely reason a negotiated extension is unlikely to
happen. If it is not needed at all, then it might be easier not to bother.

Amos

Received on Sunday, 2 October 2016 05:49:07 UTC