Re: [Wish] WG recharter

Hi Bernard,

When the WG was presented (and approved) at Dispatch, the goal was to reuse
"whatever JSEP supports", that's why we can't produce any media protocol in
this WG. Anyway, do you have any suggestions about how the charter text
should be modified so it covers your concerns?

Best regards
Sergio

On Thu, May 4, 2023 at 6:15 AM Bernard Aboba <bernard.aboba@gmail.com>
wrote:

> Not sure what “movement” you’re looking for.
>
> No one is going to create a separate spec for transport of CMAF in Data
> Channel, just like there is no “spec” for transport of CMAF in MoQ. There’s
> no spec, because it’s just single line in a document: “one CMAF chunk per
> message.”  Done.
>
> What would be nice is if the Charter included a document that would derive
> what the WHEP requirements are, based on the use cases. Right now, it seems
> like the requirements are “whatever JSEP supports”.   For at least some use
> cases (e.g. live sports) that’s not even “in the ballpark”.
>
> > On May 3, 2023, at 08:06, Sean Turner <sean@sn3rd.com> wrote:
> >
> > Hey so this has been sitting around for a bit and just wanted to check
> on whether there has been any movement I might not be aware of.
> >
> > Cheers,
> > spt
> >
> >> On Apr 17, 2023, at 15:18, Tim Panton <tim@pi.pe> wrote:
> >>
> >> I think there might well be need for a group to specify how to carry
> various things over DataChannels,
> >> (We’ve tunnelled webpages and websockets, lighting control signals,
> identity proofs, subtitles, certificates, files, firmware updates etc)
> >> but I don’t think WISH is the place to have that discussion.
> >>
> >> @bernard Perhaps we should sketch something out for dispatch...
> >>
> >> T.
> >>
> >>>> On 17 Apr 2023, at 19:06, Bernard Aboba <bernard.aboba@gmail.com>
> wrote:
> >>>
> >>> Juliusz said:
> >>>
> >>> "I think that's an important point.  In WHIP, Sergio has been careful
> to
> >>> stick to signalling, and defer all other protocol features to existing
> >>> standards.  That's what makes WHIP so eminently implementable."
> >>>
> >>> [BA]  Except that there *are* "eminently implementable" standards,
> such as CMAF for containerization and MSE API (for rendering).  To prove
> the point, CMAF is supported in early MoQ drafts and browser clients demos
> supporting MSE have implemented "rewind".  If WebTransport were swapped out
> for RTCDataChannel, those same demos would run in Chromium.
> >>>
> >>> So the protocols have already been defined and implemented, and
> running code already exists.  In that sense, the remaining functionality is
> largely in the realm of "signaling" (e.g. how to request CMAF over
> RTCDataChannel, or how to request "rewind").
> >>>
> >>>
> >>>
> >>> On Mon, Apr 17, 2023 at 10:33 AM Juliusz Chroboczek <jch@irif.fr>
> wrote:
> >>>> Regarding the specific DRM and rewind functionalities, there is not a
> standard
> >>>> way of implementing them that we can reuse here.
> >>>
> >>> I think that's an important point.  In WHIP, Sergio has been careful to
> >>> stick to signalling, and defer all other protocol features to existing
> >>> standards.  That's what makes WHIP so eminently implementable.
> >>>
> >>> -- Juliusz
> >>> --
> >>> Wish mailing list
> >>> Wish@ietf.org
> >>> https://www.ietf.org/mailman/listinfo/wish
> >>
> >
>