W3C home > Mailing lists > Public > public-webrtc@w3.org > January 2012

Re: CHANGE: Remove numeric constants from WebRTC API

From: Harald Alvestrand <harald@alvestrand.no>
Date: Thu, 12 Jan 2012 23:07:13 +0100
Message-ID: <4F0F5991.1000601@alvestrand.no>
To: Justin Uberti <juberti@google.com>
CC: Cullen Jennings <fluffy@cisco.com>, "public-webrtc@w3.org" <public-webrtc@w3.org>
On 01/12/2012 10:02 PM, Justin Uberti wrote:
> +1
>
> On Thu, Jan 12, 2012 at 11:10 AM, Cullen Jennings <fluffy@cisco.com 
> <mailto:fluffy@cisco.com>> wrote:
>
>
>     I think we should do whatever Hickson is going to do for HTML5 spec.
>

What I can find in the HTML5 spec is a term called "enumerated 
attribute", which is used in quite a number of places:
http://www.whatwg.org/specs/web-apps/current-work/#enumerated-attribute

The number of occurences of "enumerated attribute" and of "const" seems 
to be on the same order of magnitude - ~30 of each.

>
>     On Jan 10, 2012, at 2:10 AM, Harald Alvestrand wrote:
>
>     > Hello group, and happy 2012!
>     >
>     > One of the last discussions of 2011 was the discussion triggered
>     by Tab Atkins about numeric constants in the API: he strongly
>     recommended against them, and recommended that we instead use
>     strings (possibly in the special form of WebIDL enumerations,
>     which are mapped to JavaScript strings by the language binding.
>     >
>     > It seems to me clear from the discussion that:
>     >
>     > - The WebIDL community is strongly in favour of such a change
>     > - There has been weak or no real opposition to the idea in the
>     WebRTC WG
>     > - If such a change is to be made, it is better to do it as soon
>     as possible, so that implementations don't have to change after
>     significant deployment of code has been made.
>     >
>     > But it's a bit of work, and I don't want to ask the editors to
>     do it if there's the possibility that the WG could decide not to
>     do it anyway.
>     >
>     > I'm therefore calling for the WG's consensus to do this change.
>     > It affects at least the following:
>     >
>     > - MediaStreamTrack: readyState (LIVE, MUTE and ENDED)
>     > - PeerConnection: readyState, iceState and sdpState
>     >
>     > In all four cases, the current "const unsigned short"
>     declarations woudl be replaced with strings; as I understand the
>     enum spec from WebIDL [1], this:
>     >
>     >    const unsigned short LIVE = 0;
>     >    const unsigned short MUTED = 1;
>     >    const unsigned short ENDED = 2;
>     >    readonly attribute unsigned short readyState;
>     >
>     > would become this:
>     >
>     >    enum ReadyStateType { "live", "muted", "ended" };
>     >    readonly attribute ReadyStateType readyState;
>     >
>     > Please comment on this - the chairs will attempt to draw a
>     conclusion based on mail arrived on Tuesday, Jan 17, at 9AM GMT (7
>     days from now).
>     >
>     >                   Harald, for the chairs
>     >
>     > [1] http://dev.w3.org/2006/webapi/WebIDL/ - latest draft 5 Jan 2012
>     >
>
>
>
Received on Thursday, 12 January 2012 22:07:51 UTC

This archive was generated by hypermail 2.3.1 : Monday, 23 October 2017 15:19:26 UTC