Re: [rtcweb] Adam Fineberg's choices was Re: Straw Poll on Video Codec Alternatives

[Tim Panton <[e-mail]>]

On 8 Jan 2014, at 14:40, cowwoc <[e-mail]> wrote:

> On 08/01/2014 6:16 AM, Tim Panton wrote:
>> On 7 Jan 2014, at 17:42, Adam Fineberg <[e-mail]> wrote:
>> 
>>>> 	• Browsers MUST support both H.264 and VP8, other entities MUST support at least one of H.264 and VP8
>>>> 		• Are you in favor of this option [Yes/No/Acceptable]: No
>>>> 		• Do you have any objections to this option, if so please summarize them: Shouldn't differentiate between types without clearer delineation.
>> I think that's entirely fair - here's a draft definition:
>> Browser:
>> 	An entity that supports multiple usages through dynamically loadable code (javascript) and can render arbitrary user interfaces
>> using HTML5. (e.g chrome)
>> Other entities:
>> 	A non-browser entity supports a single dedicated usecase and will lack the ability to dynamically load a user interface or code
>> to support differing use cases. (e.g. a doorbell)
>> 
>> It's like blank tape vs pre-recorded cassettes (if you are old enough) .
> 
> That is a problematic definition. Native applications running on mobile devices can technically do everything you mentioned in the "browser" definition. I thought the entire point of this option is to say that web browsers must implement X but native applications must implement Y.

True, but any mobile app that allows the user to select a (possibly user defined) HTML5+javascript interface needs to support both codecs - because it can't know in advance what the other end will support. - If I point my general-purpose-home-automation-console app at the boiler I may need to render an h264 flame image, when accessing with my doorbell the same app may render a VP8 stream of the delivery driver. - So the general purpose app needs the 'browser' dual codec MTI.

Conversely the boiler will probably ship with a dedicated this-make-of-boiler-only mobile app with a prebuilt UI which only speaks h264.
The device specific app needs the ' other entity' single codec MTI.

> 
> I was expecting "browser" to be defined as an "application that enable users to render arbitrary web pages" whereas other entities are "special-purpose applications that enable users to carry out a limited set of operations on a limited set of (web) pages".

I was avoiding the words 'pages','web-server','HTTP' from my definition as I feel that the transport and encapsulation isn't the issue. It's the dynamic - user selected - unpredictability of the created-at-runtime interface required by the other peer that matters here.

> 
> Gili
> _______________________________________________
> rtcweb mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtcweb

Tim Panton - Web/VoIP consultant and implementor
www.westhawk.co.uk