Re: [hybi] comments on draft-ietf-hybi-websocket-requirements-00

[gabriel montenegro <[e-mail]>]

Hi Micheil,

Please see inline...



----- Original Message ----
> I think there's only one change, in what I've read of this thread (since the 
> message
from Gabriel, 9th July 2010) that I could argue against as a protocol 
> implementor
and active maintainer of a near spec-compliant server.

The 
> only thing that does stand out is the message size constriants, which 
> would
not be a good idea. If a client sends a message that is too large, then 
> it is up to
the server-side developer to decide if they want to disconnect 
> the socket for that
reason.

[gm] Please see my response to Greg's message on this topic. 
One thing that comes to mind is what if it's a legitimate client trying to figure out
what the largest message to the server is. finding this out by facing a series of
socket disconnects can be tedious.

The second point I find interesting is the 
> sub-protocol negotiation point, this is
something that I think would possibly 
> be a good idea, but may add unneeded
complexity.

[gm] Is this because of the unfortunate use of the word "negotiation"?

The original wording in the draft alrady called for sub-protocol support thus:

   REQ. 11:  The WebSocket Client MUST be able to request the server,
      during the handshake, to use a specific WebSocket sub-protocol.

the proposed rewording just called out the fact that the client proposal 
might involve more than one sub-protocol, but the server would only choose
one. This is exactly how Upgrade works in HTTP1.1 already.

thanks,

Gabriel