Re: [rtcweb] (no subject)

I think there is a readability benefit to have all the stuff spelled out at
media-level. I don't see what we gain by moving things to session level;
the byte savings are not significant enough to be worth optimizing for.

However, if there is a clear interoperability advantage for one or the
other, let's just SHOULD that.

On Sun, May 18, 2014 at 7:15 AM, Christer Holmberg <
christer.holmberg@ericsson.com> wrote:

>  Hi,
>
>
> I don't think we need to say anything. Because, if I receive session-level
> attributes on the wire, I should not have to move them around in my JS app
> before passing the SDP to JSEP. And vice verse.
>
>
>
> This section is instructions to the browser author.
>
>
>
> Correct. My mistake.
>
>
>
> Then, assuming we want to say something, is there a reason why we can’t
> use MUST (whatever MUST of course needs to be decided :)? It will life
> easier for JS apps that, for whatever reason, need to process the SDP
> generated for the browser.
>
>
>
> Regards,
>
>
>
> Christer
>
>
>
>
>
> -Ekr
>
>
>
>
>
>  Also keep in mind that the session vs media does not only affect SDP
> attributes - you have the same issue for the c= line.
>
> Regards,
>
> Christer
>
>
> -----Original Message-----
> From: rtcweb [mailto:rtcweb-bounces@ietf.org] On Behalf Of Cullen Jennings
> Sent: 17 May 2014 21:28
> To: Eric Rescorla
> Cc: rtcweb@ietf.org
> Subject: Re: [rtcweb] (no subject)
>
>   as a purely pragmatic issue, some device seem to have SDP bugs that
> cause them to ignore some attributes at the session level. I don't really
> care what we do on this as mostly I think theses devices should fix their
> bugs but when people talk abut SDP interop issues, this is occasionally an
> issue that comes up.
>
> So I agree we need to have a SHOULD on saying where things go that could
> be at either a session level or a m= level. Session level makes more sense
> from a standards point of view but m= level might have slightly better
> interoperability.
>
> I don't care which we choose - perhaps some of the people that considerers
> themselves and experts on why SDP has interop problems could chime in on
> how important this is.
>
>
>
> On May 11, 2014, at 11:52 PM, Eric Rescorla <ekr@rtfm.com> wrote:
>
> > https://github.com/rtcweb-wg/jsep/issues/19
> >
> > JSEP S 5.2.1 reads:
> >
> >    Attributes that are common between all m= sections MAY be moved to
> >    session-level, if explicitly defined to be valid at session-level.
> >
> > We should probably encourage this. I propose we make this a SHOULD
> > rather than a MAY.
> >
> > -Ekr
> >
> > _______________________________________________
> > 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
>
>
>
> _______________________________________________
> rtcweb mailing list
> rtcweb@ietf.org
> https://www.ietf.org/mailman/listinfo/rtcweb
>
>