Re: {minutes} Geneva F2F day 1 16/09/2014

On Sat, Sep 20, 2014 at 6:41 PM, Silvia Pfeiffer <silviapfeiffer1@gmail.com>
wrote:

> On Sun, Sep 21, 2014 at 12:42 AM, Glenn Adams <glenn@skynav.com> wrote:
> >
> > On Sat, Sep 20, 2014 at 7:30 AM, Silvia Pfeiffer <
> silviapfeiffer1@gmail.com>
> > wrote:
> >>
> >> Most interesting read. Seems like there is not that much that doesn't
> >> map in either direction.
> >>
> >> A few additions / corrections about WebVTT that relate to this
> >> discussion and that were missed (though Simon caught most):
> >>
> >> 1.) Mapping WebVTT - 608/708
> >> This wasn't mentioned, but I wanted to make sure people are aware that
> >> there is already a spec that provides a basic mapping between WebVTT
> >> and 608 and 708 captions at
> >>
> https://dvcs.w3.org/hg/text-tracks/raw-file/default/608toVTT/608toVTT.html
> >> .
> >> It does indeed require some CSS.
> >>
> >> 2.) Repetition of cue settings:
> >> There's a plan to allow for some file metadata to set different
> >> default values for settings, so repetition is not necessary.
> >>
> >> 3.) CSS works by using ::cue and ::cue-region - the first for cues and
> >> the second for regions.
> >> These selectors work from a Web page (wherever CSS is specified there)
> >> and override default styling of WebVTT cues.
> >>
> >> 4.) cue overalpping:
> >> It's actually not true that cues cannot overlap in time. They can and
> >> are expected to. In fact, chapter cues are specifically defined as a
> >> particular kind of overlapping cues that are hierarchically
> >> structured, or in your words: are nested.
> >>
> >> 5.) cue ids:
> >> I think it might be possible to map numeric identifiers to XML
> >> identifiers - at least it's possible to use them like this for CSS:
> >> ::cue(#\31) { color: green; }
> >> This references a cue with id=1. That might work for mapping to TTML
> >> xml:id too.
> >
> >
> > it would be necessary to prefix a numeric identifer with at least one
> alpha
> > to use with xml:id
>
> Interesting. How does this work in the context of browsers?
>
> For comparison, I'm looking at SVG, though I'm wondering if that's
> applicable to TTML.
>
> Here's how they define ID in SVG:
> http://www.w3.org/TR/SVG/svgdtd.html#DTD.1.4
>
> <!ENTITY % SVG.id.attrib
>     "id ID #IMPLIED"
> >
>
> (so, it's "id" and not "xml:id")
>
> It works, e.g. in an example document like this, the element is
> identifyable:
>
> <svg xmlns='http://www.w3.org/2000/svg' id='1'>
>     <script>
>       alert(document.getElementById('1'));
>       alert(document.getElementById('1') == document.documentElement);
>     </script>
> </svg>
>

The syntax of the ID attribute value type is defined originally by SGML,
but in the XML context (SVG is an XML language) as the Name construct [1].

[1] http://www.w3.org/TR/REC-xml/#NT-Name

[4]   NameStartChar   ::=   ":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] |
[#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] | [#x37F-#x1FFF] |
[#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] |
[#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF][4a]   NameChar
   ::=   NameStartChar <http://www.w3.org/TR/REC-xml/#NT-NameStartChar> |
"-" | "." | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040][5]   Name
   ::=   NameStartChar <http://www.w3.org/TR/REC-xml/#NT-NameStartChar> (
NameChar <http://www.w3.org/TR/REC-xml/#NT-NameChar>)*

In XML Schema 2, the value space for ID typed attributes is defined
equivalently at [2].

[2] http://www.w3.org/TR/xmlschema-2/#ID


>
> It seems to me that SVG have moved away from xml:id


SVG has never moved away from the formal XML definition of ID. Also, it
never directly used xml:id, which is a different name for the otherwise
unqualified 'id' attribute, both of which use the same value syntax.


> , since the #myRect
> reference in the example below doesn't work
>
> <svg xmlns="http://www.w3.org/2000/svg"
> xmlns:xlink="http://www.w3.org/1999/xlink"
>      version="1.2" baseProfile="tiny" viewBox="0 0 480 360">
>
>   <desc>Forward reference of use element</desc>
>
>   <use xlink:href="#myRect" x="200" fill="green"/>
>   <circle cx="450" cy="50" r="50" fill="yellow"/>
>
>   <g fill="red">
>     <rect xml:id="myRect" width="100" height="100"/>
>   </g>
> </svg>
>
> but starts working if you add id="myRect" on the <rect> element.


Again, SVG did not adopt use of the attribute named xml:id, but uses the
unqualified 'id' attribute [3]. However, from the perspective of value
syntax, both are the same and neither admit '1' as a legal value.

[3] http://www.w3.org/TR/SVG/struct.html#IDAttribute
id = "name"Standard XML attribute for assigning a unique *name* to an
element. Refer to the Extensible Markup Language (XML) 1.0
<http://www.w3.org/TR/2008/REC-xml-20081126/> Recommendation [XML10
<http://www.w3.org/TR/SVG/refs.html#ref-XML10>].


> It
> continues to work even when replacing the "myRect" string with merely
> a number such as "1".
>

That is a violation of SVG syntax. It may be that some UAs don't enforce
that syntax as required by SVG.


>
> Is there any intention to move to that approach in TTML, too?


You are mixing up two things here:

(1) the names of attributes id vs xml:id, and
(2) the value syntax for these attributes.

SGML, HTML1 through HTML4, and XML variously make use of id. Some XML
languages additionally or alternatively use xml:id [4]. Both id and xml:id
have the same value syntax, neither of which accept '1' as a legal value.

[4] http://www.w3.org/TR/xml-id/

E.g., HTML4 specified [5]:

   - *ID* and *NAME* tokens must begin with a letter ([A-Za-z]) and may be
   followed by any number of letters, digits ([0-9]), hyphens ("-"),
   underscores ("_"), colons (":"), and periods (".").

[5] http://www.w3.org/TR/REC-html40/types.html#type-name

HTML5 made a non-backward compatible change by removing this syntax
restriction, allowing any single character value even digits provided they
are unique in a document.

In your tests, I suspect you are using the somewhat bastardized form of SVG
that is embedded in HTML5 and that supports this non-backward compatible
change regarding the ID value type.


> If so,
> then maybe the WebVTT identifiers should be mapped into an "id"
> attribute rather than the "xml:id" attribute?
>

We did have an issue [6] (now closed - withdrawn) to consider allowing id
in addition to xml:id, but even if that were to occur, we would not alter
the value syntax for id to match the syntactically weaker form introduced
by HTML5.

[6] http://www.w3.org/AudioVideo/TT/tracker/issues/205

XML validity is still considered a positive requirement for TTML.


>
> Just some thoughts...
>
> Cheers,
> Silvia.
>