Re: Poll for preferred API alternative

On 09/06/2012 07:51 PM, Hutton, Andrew wrote:
> Hi Harald,
>
> I take your point but currently you are only talking about the SDP that
> Chrome produces today which is likely to change and may not be the same
> SDP that another browser vendors chooses to implement so examples of
> existing interworking problems with the SDP Chrome produces are not so
> useful.  All I am saying is that we still need to do some work to nail
> this down.

I agree, this must be nailed down (but I think the responsibility for 
this part is currently with the IETF rtcweb WG). And what browsers 
produce as descriptions must be similar enough to enable the app to just 
apply what the browser at the other end produced  - that being an offer, 
answer or something else - with no modifications and things should just 
work. Simple apps should never have to resort to manipulation of these 
descriptions.

This applies regardless if we stick to SDP or, for some reason, adopt 
some other format.

I think it is OK if the descriptions must be manipulated for interop 
(with non-browser end-points) cases; after all, the main target for this 
work is the browser-browser case.

Stefan

>
> Regards
>
> Andy
>
> *From:*Harald Alvestrand [mailto:harald@alvestrand.no]
> *Sent:* 06 September 2012 18:05
> *To:* public-webrtc@w3.org
> *Subject:* Re: Poll for preferred API alternative
>
> On 09/06/2012 06:30 PM, Hutton, Andrew wrote:
>
>     Hi All,
>
>     I think the discussion on SDP issues raises an important point as
>     some people might be of the impression that the PeerConnectionAPI is
>     almost complete but reality is that the SDP blob that comes out of
>     it is not specified and it is likely that it will end up looking
>     quite different to SDP that existing implementations are familiar
>     with. We are therefore quite some way from a standard API that will
>     be interoperable between different browsers and legacy
>     implementations without a lot of tweaking and knowledge of SDP in
>     the application.
>
> I've heard this theory advanced.
>
> I've not heard it advanced by the people who have written the
> applications that interwork between the present, early WebRTC
> implementations and existing, legacy SDP-using devices, such as sipml5.
> What comes out of Chrome is SDP, and conforms to the SDP RFCs. (If not -
> file bugs!)
>
> I'd like this line of argument supported by "when I fire Chrome browser
> SDP at <device>, it doesn't interwork because of <reason>". That's much
> more actionable than "I think".
>
>            Harald
>

Received on Thursday, 6 September 2012 18:17:20 UTC