Re: [rtcweb] Data Channel Negotiation and reopening of decisions

On Feb 16, 2013, at 4:30 PM, Matthew Kaufman (SKYPE) wrote:

> You're right of course... I just can't keep straight which of the things are poor choices and which are reasonable choices that have been bastardized to the point of being poor choices.
> 
> If we're going to use SCTP, and it has perfectly good unidirectional channels, why can't we just expose those to users of the W3C WEBRTC API?
Making data channels unidirectional and even making reliability settings a per message property
(so exposing the SCTP API) would be very simple and make the additional protocol superfluous.

However, WebRTC and RTCWEB groups preferred to make the API similar to Websockets...
> 
> Apparently this "we built a simple protocol to allow us to set up bidirectional pairs of SCTP streams to create channels" isn't as simple as it first looked, so we should be revisiting the original assumption, not creating a complicated protocol just because we can.
At least for me, this is a very simple protocol. Even if we add some error handling. Very easy to understand
and from implementing it, very easy to implement.

Best regards
Michael
> 
> Matthew Kaufman
> 
> Sent from my iPad
> 
> On Feb 15, 2013, at 10:40 PM, "Randell Jesup" <randell-ietf@jesup.org> wrote:
> 
>> On 2/15/2013 4:56 PM, Matthew Kaufman (SKYPE) wrote:
>>> None if this would be a problem if we weren't making the mistake of using SCTP, which itself made the mistake of doing full-duplex channels instead of unidirectional channels with optional one-to-one or even many-to-one associations (as RTMFP has, as an example)
>> 
>> SCTP only has unidirectional streams.
>> 
>> If you're referring to the DataChannel protocol, there was a strong wish on the mailing list (and W3) for websockets-like bidirectional channels, so we built a simple protocol to allow us to set up bidirectional pairs of SCTP streams to create channels.
>> 
>> -- 
>> Randell Jesup
>> randell-ietf@jesup.org
>> 
>> _______________________________________________
>> rtcweb mailing list
>> rtcweb@ietf.org
>> https://www.ietf.org/mailman/listinfo/rtcweb
>> 
> _______________________________________________
> rtcweb mailing list
> rtcweb@ietf.org
> https://www.ietf.org/mailman/listinfo/rtcweb
> 