Ask for tests for normative changes?

[foolip]

I've been sending PRs for a lot of specs like this one: https://github.com/WebAudio/web-audio-api/pull/1423

For context, see https://github.com/foolip/testing-in-standards/blob/master/policy.md.

Is MSE in a state spec-wise and test-wise where this would make sense? If it is I'm happy to send such a PR, but I'd like to check if there are major blockers that would have to be resolved first.

Thanks!

[jdsmith3000]

Normative changes to MSE aren't currently possible since the Media Working Group charter has expired. Should that change, your request to match normative changes with test case support seems reasonable.

[foolip]

Does that mean that https://w3c.github.io/media-source/ is essentially unmaintained? What happens if it's discovered that all implementations do one thing and the spec says another, so that the only reasonable course is to make a normative change to the spec? Would one have to fork the spec?

[jdsmith3000]

Changes to the spec would likely require incubation. Logging issues here is still appropriate.

[foolip]

OK, so I'll leave this PR sitting around for when the spec becomes active again.

[paulbrucecotton]

normative changes to MSE aren't currently possible since the Media Working Group charter has expired.

Actually the current HME WG charter is in affect until April 2018.

But as explained in the announcement email to the HME WG, the current charter was extended to permit ONLY maintenance on the then MSE Recommendation. Now that EME is a Recommendation the charter also permits maintenance on the EME Recommendation.

"Following a recent W3C discussion on the maintenance of W3C Recommendations, for now, the best approach is to do maintenance in the Working Group, and extending the Working Group to allow that. As such, the Working Group is only allowed to make corrections that do not add new features [4] to its Recommendation(s) under this charter extension."

A revised HME WG charter approved by the W3C Advisory Committee would be required to permit substantive changes to be made to either of the MSE or EME W3C Recommendations

/paulc HME WG Chair

[plehegar]

The WG does exist indeed to allow maintenance of the Recommendations of MSE and EME. This means the Group is allowed to do a new revision of the Recommendation as long as they don't include new features, so it does allow for substantive and editorial corrections. There is the case until April 2018 and there is a proposed change to the W3C Process to permit W3C to do so beyond that date even without a Working Group around. As an example, #199 can perfectly be done to revise the current Recommendation since it doesn't add a new feature imho.

For new features, those could be done with the incubation but we would need a new Working Group to put them under the W3C patent policy.

Regarding @foolip proposal, I strongly support it. This is orthogonal to the maintenance vs new features difference in fact. Whether the work is through a WG or through incubation, we should always get changes to specs along with tests.

[guest271314]

What is the significance of the "April 2018" date? What is "incubation"? It does not appear to be currently possible to access the accumulated media at a MediaSource object. Would exposing the underlying media resource as an ArrayBuffer https://github.com/w3c/media-source/issues/209 be considered a "feature"?

[foolip]

@guest271314, April 2018 is the end date of https://www.w3.org/2013/09/html-charter.html. And incubation is what we call things that happen in the WICG: https://www.w3.org/community/wicg/

[guest271314]

@foolip Have not yet composed a specification from scratch, yet; though do have some experience within the realm of technical writing relevant to intellectual property.

One approach that have been considering is a decodeVideoData() method (preferably standalone; without a <video> element being necessary and could be used in Worker), modeled on Web Audio API OfflineAudioContext. That is, the ability to pass an ArrayBuffer to the method which get the video data (video and audio), having a .start() method with parameters allowing "from", "to"; which would return an ArrayBuffer comprising only that portion of the video, capable of being rendered at an HTMLMediaElement (i.e., <video>). That would allow concatenating N specific portions of N videos into a single ArrayBuffer, or "VideoBuffer", capable of "seamless" playback as a single video.

Have coded the same (working) procedure with OfflineAudioContext.

How to start the process of specifying a decodeVideoData object?

[guest271314]

@foolip How to create a repository at https://www.w3.org/community/wicg/? Does WHATWG have a similar group/process?

[wolenetz]

Looks to me like a spec / README.md bugfix in scope for V2.