TimedText/WebVTT Wide Review

From W3C Wiki
Jump to: navigation, search

BugZilla and GitHub

Bugs that are related to the FPWD (13 November 2014) or other Wide Review requests are tagged in BugZilla with the WhiteBoard word 'widereview'

There is also a Github issues repository. These should be scanned to make sure that we have addressed all that needed to be addressed prior to CR.

Implementations and Authoring Guides

The rather large number of implementations indicates fairly extensive review.

There is a introduction to VTT Concepts, originally on the WebPlatform site.


There is a comprehensive test suite at Web Platform Tests for which the current results for selected platforms can be viewed online. The results do not cover all implementations of VTT.

First Wide Review Review Requests and Responses

Messages sent requesting review internally to the W3C, of the FPWD

Thierry requested cross-review from I18N, Accessibility, CSS and HTML. Bugs were initially filed and handled in BugZilla; unresolved ones were transferred to GitHub and further trackign happened there. The responses are tracked in the tables below; however, tracking was done in the bugs/issues (BugZilla and GitHub) rather than in meetings or by email exchange.

  • I18N: [Status: pending response from TTWG and acceptance by I18N]
num issue bugzilla disposition type status (CG) status (WG)
I-1 6.2.1 line heights 28269 works for me, accepted by author (use CSS) ed 2.4 3.4
I-2 3.2 scroll value and vertical writing 28268 proposed to define scrolling only for horizontal text in this version sub 2.4 3.3
I-3 3.1 Line wrapping for non-Latin scripts 28267 fixed, we defer to CSS, and use overflow-wrap from there; accepted by author sub 2.4 3.4
I-4 6.2.1 processing model handling of bidi 28266 fixed by including a bidi example; accepted by author sub 2.4 3.4
I-5 incomplete ruby implementation 28265 marked s a possible future extension of the ruby support; accepted by author sub 2.4 3.4
I-6 4.3.2 very few character escapes 28264 fixed, we simply refer now to HTML for all escapes sub 2.4 3.3
I-7 "valid" language tags 28263 fixed, we require validity by authoring tools only sub 2.4 3.3
I-8 4.3.2 emphasis spans 28262 it's possible to work around it just like in HTML by styling b, i etc.; change accepted by author sub 2.4 3.4
I-9 WebVTT percentage digits? 28261 fixed, responder satisfied sub 2.4 3.4
I-10 zero or more characters? 28260 fixed, responder satisfied sub 2.4 3.4
I-11 Normalization and string identity issues 28259 fixed, responder satisfied sub 2.4 3.4
I-12 more formal grammar? 28258 won't fix; we supply authoring guide as a separate document sub 2.7 3.3
I-13 start/end linked to left/right 28257 fixed, clarified start and end to be line-left and line-right sub 2.4 3.3
I-14 no non-ASCII examples 28256 fixed, responder satisfied sub 2.4 3.4
I-15 no provision for indicating overall content language(s) 28255 won't fix, overall language is provided by the environment sub 2.4 3.3
I-16 term "NOTE" may not be friendly to users in other languages 28254 won't fix, the VTT syntax is rooted in latin text; accepted by resonder ed 2.4 3.4
  • CSS: [Status: pending response from TTWG and acceptance by CSS]
    • Request for review: to Bert during Interaction call.
    • CSS WG review: April 2015
    • The following bugs were filed
num bugzilla disposition type status (CG) status (WG)
C-1 Can WebVTT be extended to ref an external stylesheet? fixed, we allow inline styles which can in turn @import external stylesheets sub 2.4 3.3
C-2 The example in section 1.1 is possibly a use case for a new property and keyword for CSS, tentatively called 'text-wrap: balance' fixed, we use text-wrap: balance. Note that this introduces a dependency on CSS Text 4 which is still in editor's draft. sub 2.4 3.3
C-3 "Section 5.5: Some people wonder if HTMLElement is the right type to use for all the nodes, or if it is better to use more specific subclasses." fixed, HTMLElement no longer mentioned, specific text introduced sub 2.4 3.3
C-4 You can remove the entire section about how to evaluate the selector once section 5.5 defines an accurate DOM mapping fixed sub 2.4 3.3
C-5 this section doesn't seem to apply the IDs of region objects. handled as part of the change to 28474 sub 2.4 3.3
C-6 "group of selectors" has no defined meaning. You probably want either "compound selector list" or "complex selector list". fixed sub 2.4 3.3
C-7 It looks like cue regions are basically blocks that get filled with cues. Why not just make those an element selectable by ::cue()? We can't change this without risking confusion between cue identifiers and region identifiers sub 2.7 3.3
C-8 Even though regions have an ID, it looks like you can't select by it. fixed, selector added sub 2.4 3.3
  • Accessibility:[Status: pending response from TTWG and acceptance by Accessibility]
num bugzilla disposition type status (CG) status (WG)
A-1 Users of Magnification no change? The layout algorithm is designed to be robust against font size changes ge 2.7 3.3
A-2 The spec should include feature explanations in plain language no change, we rely on external documents to provide an authoring guide. ge 2.7 3.3
A-3 Captions on the audio element fixed, explanations added ge 2.4 3.3
  • HTML:[Status:Done]
    • Request for review: to Mike during Interaction call.
    • HTML WG review: March 2015 (no comments)

Messages were sent requesting review internally to other standard bodies:

  • MPEG mp4 file format ad-hoc (list is private, alas; they have a packing into mp4). MPEG replied that they had defined the packing, and implemented reference software, and had no comments.
  • 3GPP SA4 (multmedia and codecs)
  • EBU timed text (November 13th 2014, sent to EBU-tt@list.ebu.ch)
  • DASH industry forum (email to the chairs, November 13th 2014)
  • SMPTE (November 13th 2014, mail sent to 24TB)

We have messages of review from the following:

    • Kaltura who indicated that they have read the spec. and implement good chunks of it, but did not (at that time) have formal comments.
    • Discussion of 'green' This was an issue for the mapping of CEA-608/708, not the core VTT specification
  • Andreas of IRT raised these other points in his review:
    • Overall readability:
      • You have to read the rendering section (to understand authoring)
      • Often a part of a section is dependent of other parts
      • You most probably have to read the complete spec
      • it would be a big help if the information about a "feature" like positioning is all in one place
      • Graphical representations would help a lot to understand the abstract concepts.
      • Some concepts from the HTML spec are so vital to WebVTT that a short summary would be of help
    • Grammar
      • it would be a great help in addition to the prose there is a formal representation e.g. in Extended Backus–Naur Form (EBNF).
    • Alignment:
      • Maybe it would help if it is made more concrete how "vertical" and "horizontal" relate to the rendering pane of the video?
      • The text position depends on the text alignment
    • Overall response to these follows: We have tried to re-factor the specification for better readability, but it remains a formal specification for implementors rather than authors (as noted also to WAI), and as such we depend on the existence of (external) authoring guides such as is/was at WebPlatform Docs. We would not like to provide a formal grammar as we fear it would be impossible to align completely with the parsing algorithm. We have tried to improve the text positioning and alignment prose and make it less tangled.
    • Bugzilla bugs handling Andreas' reports
num bugzilla disposition type status (CG) status (WG)
M-1 28070 Title of 3.1 should be "WebVTT cues" fixed ed 2.4 3.3
M-2 28071 The concept of line position is unnecessarily used when writing direction is introduced fixed sub 2.4 3.3
M-3 28072 Consistent use of snap-to-lines flag fixed sub 2.4 3.3
M-4 28073 Rename "line position" to something more appropriate? fixed, better wording (line offset etc.) sub 2.4 3.3
M-5 28074 Link up mention of "region" in text position definition fixed sub 2.4 3.3
M-6 28075 Add a visual representation of region anchoring fixed, diagram added ed 2.4 3.3
M-7 28408 Consider reorganising the Rendering section section re-organized ed 2.4 3.3

Second Wide Review Review Requests and Responses

Thierry sent messages to Privacy, Security, i18n, CSS, Media and Entertainment, Media Text Tracks, and APA, on Sept 7th 2017. No issues were received from outside the WG; issues were received from two members of the WG, and tracked in GitHub.

The resulting issues are in three classes:

  • the open list is now a few minor editing issues that can be handled in the CR period;
  • the issue that is pending are in the process of reconciliation with the editor or commenter
  • issues that are resolved

The tags on the issues indicate the CG, WG, and Commenter status. There is no tracking table here; the tables above are a poor reflection of the dialogue, and for this round, tracking was done solely in the GitHub issues.

Substantive changes since (Second) Wide Review Review

The following table contains all the PRs and breaks them into classes; technical (substantive), editorial, and administrative.

num type title
388 sub Clarify the "WebVTT cue settings list" definition.
389 sub Add a "User agents that do not support CSS" conformance class.
390 ed Move definition of chapter title text
394 ed Add some explanatory text about the different types of VTT files.
395 sub Add default color and background color names for applicable classes.
396 ed Fix up algorithm markup and improve a bikeshed style.
398 ed Break up section 6 a bit more to make it more readable.
399 ed Clarify line cue setting specification.
400 sub Region clarification issue 376
407 adm Add code of conduct as proposed by plh
408 sub Turn long alt text into longdesc.
409 ed Add some introductory paragraphs to the Data model section.
410 sub Add a paragraph about the reasons why limiting styles.
411 sub Update line height for lines in regions.
417 sub Unsigned long issue 414
418 sub Class override
419 sub Timeranges clarification
420 sub Cue setting clarifications
421 sub Computed position alignment
422 sub Don't make edge margins depend on aesthetics.
423 sub Region lines cannot be negative now that they are unsigned long.

Guidelines for Status of comments

Processing by the CG for each comment

A first review raised comments about WebVTT First Public Working Draft dated 13 November 2014.

For each comment, please fill the "status CG" column:

  • 1 - Open: Not yet processed by the CG
  • 2 - Processing by the CG
 2.1- Pending proposal by CG,
 2.2- Pending CG resolution,
 2.3- Done - CG resolution,
 2.4- Done - CG resolution and spec update
 2.5- Done with Partial CG resolution,
 2.6- Done with Partial CG resolution and spec update
 2.7- Done but comment Rejected
  • For each comment, please fill the "type" column:

a- typo b- editorial c- substantive d- general comment e- undefined

Processing by the TTWG for each comment

Once the CG has processed all comments, the next steps are:

  • 3 - Processing by the TTWG
 3.1- CG resolution approved by WG
 3.2- Approved and Response drafted
 3.3- Response send to commenter
 3.4- Response agreed by commenter
 3.5- Response rejected by commenter  (need more discussion - back to step 2.5)
 3.6- Response partially agreed by commenter (need more discussion - back to step 2.5)

Next steps processed by the TTWG

  • 4- Publication of a new WebVTT WD to trigger a new and last wide review (on diffs only)and go back to step 1 above in the CG, to process new comments if any.

Final steps processed by the TTWG

  • 5 - Disposition of comments edited.
  • 6 - WebVTT CR exit criteria resolved
  • 7 - WebVTT CR version updated for TR
  • 8 - request CR publication with the director
  • 9 - if approved by the director, Publish CR
  • 10 - Draft and send call for implementation
  • 11 - Draft announcement for W3C home page.