Re: [rtcweb] Filling in details on "trickle ICE"

[Eric Rescorla <[e-mail]>]

On Fri, Aug 24, 2012 at 9:12 AM, Martin Thomson
<[e-mail]> wrote:
> On 24 August 2012 08:59, Eric Rescorla <[e-mail]> wrote:
>> On Fri, Aug 24, 2012 at 8:55 AM, Martin Thomson
>> <[e-mail]> wrote:
>> I suppose it depends on whether you think it's important for
>> independently developed Web applications and/or softphones
>> to be able to talk to each other.
>
> Maybe I needed to be clearer:
>
>>> These are not necessarily interoperability issues.  You can solve all
>>> of these problems without creating another RFC.
>
> I'd be interested in hearing the interoperability reason for, for
> example, giving up on a particular component instead of waiting for
> additional candidates.  What effect does that have on
> interoperability.

I'm not following what you are saying here. Ignoring the question of
non-trickle (i.e., strict RFC 5245) endpoints, if you are a trickle endpoint
that gives up under conditions when RFC 5245 says you should give
up, you are likely to have problems. You don't think it's worth writing
that down somewhere more permanent/definitive than an email
to the mailing list?

-Ekr