Re: [rtcweb] Requesting "SDP or not SDP" debate to be re-opened

[Silvia Pfeiffer <[e-mail]>]

Is it possible to continue using SDP but not do o/a ?

While I am all for having an easy API to manipulate SDP for the common use
cases (and this abstract away the mangling) I am not too fussed about the
use of SDP itself.

I wonder if the restricted negotiation that you propose will still need
something like SDP to communicate the parameters of the negotiation. So we
could leave the data structure intact, but have a different negotiation
algorithm between browsers. We could even offer in the SDP a choice of
which negotiation algo to use so as to make interfacing with the SIP works
easier.

I hope this makes sense.

Cheers,
Silvia.
 On 22 Jun 2013 03:37, "Robin Raymond" <[e-mail]> wrote:

>
> To me the issue is not easy of use but compatibility, as what is produced
> should increase compatibility not decrease it. That's one of the major
> drawbacks of SDP with offer/answer as I've described numerous times -- it
> increases incompatibility rather than decreases it.
>
> I too want to see an API where people don't need to understand a complex
> mechanism to be able to produce a suitable application. Unfortunately, the
> current SDP approach all too often requires learning SDP by JavaScript
> developers to perform normal edge case scenarios (e.g. putting media on/off
> hold). That's simply not acceptable in my mind.
>
> That's why I will drafting up an alternative approach which can have a
> reasonable non-SDP based API without offer/answer but I will provide a
> JavaScript shim on top that will provide the current "simple" approach,
> including being able perform offer/answer using SDP (as some people
> prefer). This should more than adequately satisfy the charter's limited 1.0
> goals and responsibilities and give the non-SDP people a basis for moving
> forward where they can feel comfortable will satisfy their needs as well.
>
> -Robin
>
>
>
>   Hutton, Andrew <[e-mail]>
>  21 June, 2013 12:52 PM
>  Now this is getting ridiculous.
>
>  You do know that one of the goals is to build an API that web developers
> without knowledge of IETF protocols can use to build apps. I was only using
> this as an example where this has been achieved.
>
>  I also think you know that there are many apps out there already which
> are much more than high school projects.
>
>  I believe we are simply making life difficult by trying to do too much
> at once  and we need to rethink what is really needed and deliver on what
> we have said we we will deliver please read the charter.
>
>  Andy
>
>
> _______________________________________________
> rtcweb mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtcweb
>
>