Re: [hybi] I-D Action: draft-ietf-hybi-thewebsocketprotocol-09.txt

[Patrick McManus <[e-mail]>]

It is likely that firefox 6 will be shipped with a websockets -09
implementation with the API prefixed under Moz*.

Hopefully Firefox 7 will be shipped with -09 under the normal WebSocket
namespace - we're waiting for the binary issue with the API to be fully
sorted out. That depends on what happens on the JS side of the fence.

On Tue, 2011-06-21 at 15:15 +0200, Iñaki Baz Castillo wrote:
> 2011/6/21 Greg Wilkins <[e-mail]>:
> > Having the draft version in the handshake has proven a very workable
> > solution that prevents the heuristic version determination that was
> > previously necessary.  Is there any alternative solution?
> 
> Having experimental working implementations in order to improve a spec
> or detect/fix issues is great. Relying on such implementations is not
> so great, even less when you are talking about widely extended web
> browsers. In other words, I don't consider "polite" to experiment with
> the rest of the world just to realize of issues in a no mature
> specification (which is still a draft).
> 
> IMHO browser vendors could provide an experimental version with
> WebSocket support so developers and people in this WG can test it
> against the current specification. Ok, there it could make sense a
> *temporal* header like "X-WebSocket-TMP-Version". But such
> version/revision value should have no relationship with the final
> protocol version (i.e. "1.0") which will be submitted to the IETF.
> This is, don't mix experiments with the final spec.
> 
> Just my opinion. Regards.
> 
> 
>