- From: Elad Alon via GitHub <sysbot+gh@w3.org>
- Date: Wed, 15 Nov 2023 13:33:57 +0000
- To: public-webrtc-logs@w3.org
I completely agreed with @tovepet. > > [@eladalon1983] It requires more code from the app - even if the app doesn't want to use this feature. (Foreshadowing.) > > [@jan-ivar] Please explain. See my code vs. yours in [this comment](https://github.com/w3c/mediacapture-screen-share/issues/255#issuecomment-1805518491). > > \[@eladalon1983\] Can you give a concrete example of an application that needs to make the decision to source-inject or not on a dynamic basis? > > \[@jan-ivar\] Can you give a concrete example of why an application must make this decision to source-inject or not upfront? I don't think that's a reasonable way to respond to my question, or to represent my position. You represent my position as "some applications must make this decision upfront." I have never said that, nor could I have say it; clearly, any decision an application could make upfront, it could still *deliver* to the user-agent at a later time-point with no great hardship. Rather, I claimed the application is able to make the decision upfront - which greatly simplifies the spec, the UA-implementation and the app's own code. So I reiterate my question - what do we lose by not going with this simpler model? What application needs the extra functionality? > [@youennf] We should dig into the actual proposal more to further evaluate Elad's concerns. Agreed. Let's dig into what issue is solved by allowing applications to dynamically decide whether source-injection happens, rather than deciding ahead of time. We should only go with more complicated solutions if they serve an extended purpose. > [@youennf] For instance, the spec should state that the old tracks migrates (and thus only start receiving new frames) after the event was fired. This begs the question of when configurationchange is fired for instance. With the non-injection-model, we (1) stop delivering frames, (2) invoke the callback and (3) formally stop the old tracks. I don't think it's in-scope to specify what the source-injection model does. I think that's a follow-up. -- GitHub Notification of comment by eladalon1983 Please view or discuss this issue at https://github.com/w3c/mediacapture-screen-share/issues/255#issuecomment-1812546088 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Wednesday, 15 November 2023 13:33:59 UTC