- From: Eric Rescorla <ekr@rtfm.com>
- Date: Thu, 2 Apr 2015 06:39:24 -0700
- To: Göran Eriksson AP <goran.ap.eriksson@ericsson.com>
- Cc: "Cullen Jennings (fluffy)" <fluffy@cisco.com>, public-webrtc <public-webrtc@w3.org>
- Message-ID: <CABcZeBP4xsoP09Gsq9VfqVV_pus79dz-DCjJShwL-4BOYkBxnw@mail.gmail.com>
On Thu, Apr 2, 2015 at 2:12 AM, Göran Eriksson AP < goran.ap.eriksson@ericsson.com> wrote: > > > On 02/04/15 00:55, "Cullen Jennings (fluffy)" <fluffy@cisco.com> wrote: > > > > >> On Apr 1, 2015, at 4:29 PM, Eric Rescorla <ekr@rtfm.com> wrote: > >> > >> > >> > >> On Wed, Apr 1, 2015 at 2:29 PM, Göran Eriksson AP > >><goran.ap.eriksson@ericsson.com> wrote: > >> Hi, > >> > >> Just a few questions to ensure I don¹t misunderstand: > >> > >> > > >> >s the name implies, WebRTC 1.0: Real-time Communication Between > >>Browsers > >> >is to be considered as a first version of APIs for real-time > >> >communication. The working group will, once WebRTC 1.0: Real-time > >> >Communication Between Browsers reaches Candidate Recommendation, > >>consider > >> >proposals for backward-compatible object-oriented extensions to this > >>API. > >> > >> I assume ³backward-compatible² include the possibility of the 1.0 API > >> being supported using a js-shim on the evolved object-oriented/inspired > >> low-level API's? > >> > >> The word ³extension²; does that mean new functionality ³only² could be > >> object-oriented or does it also allow for existing functionality in 1.0 > >>to > >> be supported with low-level oo- API's, replacing 1.0 approach API's, > >>were > >> the WG to consider that motivated and desirable? > >> > >> > >> > >> I interpret this as requiring that implementations written to the 1.0 > >>API > >> function with the 1.1 API. If implementations want to internally do > >>just OO > >> APIs and have a JSL, that's their business. > >> > >> Cullen, is that what you meant? > >> > >> -Ekr > >> > >> > > > >Yes - exactly. > > > >I think of it as if I have a website that works with version X of the > >browser that only has the 1.0 API, that same website keeps working when a > >user uses version X+1 of the browser that has the 1.1 API. How the > >browser makes that happen and how they decide to split up their > >implementation is no worry of mine. I realize some browsers use various > >JS polyfills to make stuff happen. > > Words are important,:-)! Yes, that browser UA’s use JS to implement API’s > are not uncommon. Another way of supporting backward compatibility with > JS-shim/polyfill/etc would be to have it shipped with the web app meaning > the web site would have to change, but I understand that this option would > be blocked for the WG. I would not consider this second option to be backward compatible. And supporting backward compatibility to Web > 1.0-enabled site (though not that many) is relevant. > > However, then the meaning of “extension” is even more important since that > also could be interpreted as new low-level APIs *only* for new > functionality and not to do what 1.0 does, e.g. in a slightly different > way. > No, I don't think that's correct. There can be more than one way to do things. > It gives me the impression that this puts quite some constraints on the > evolution of WebRTC API’s and constraints written in stone (charter)- a > WebRTC 2.0 API would have to include the 1.0-API’s also for sites designed > for 2.0. > Well, I've been hearing this called "1.1", but yes, that is my interpretation. -Ekr Is that a correct interpretation? > > > Best Regards > Göran > > > > > > > > > > > > > >
Received on Thursday, 2 April 2015 13:40:33 UTC